Home

AIX Logical Volume Manager from A to Z: Troubleshooting and

1. i print u3 WARNING WARNING WARNING WARNING WARNING WARNING WARNING print u3 print u3 Desc PHYSICAL VOLUME DECLARED MISSING PLEASE SEE ERRPT print u3 print u3 Error Label 9 print u3 Sequence number 1 print u3 Error ID 2 print u3 Error Class 3 print u3 Error Type 4 print u3 Resource Name 6 print u3 Resource Type 7 print u3 Resource Class 8 print u3 i print u3 007 mail s PHYSICAL VOLUME DECLARED MISSING root lt lt EOF WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING Desc PHYSICAL VOLUME DECLARED MISSING PLEASE SFE ERRPT Error Label 9 Sequence number 1 Error ID 2 Error Class 3 Error Type 4 Resource Name 6 Resource Type 7 Resource Class 8 186 AIX Logical Volume Manager from A to Z Troubleshooting and Commands The following commands should be run to activate the error notification chmod 755 usr lib ras pvmiss notify odmadd tmp pvmiss add Note that SSA provides its own tools such as ssa_ela and ssa_healthcheck to notify the administrator of problems 2 8 Special considerations for rootvg The following section discusses special considerations involved when working on a rootvg volume group 2 8 1 rootvg problem determination in maintenance mode If a machine will not boot and an
2. S lsvg p lowg lowvg PV_NAME PV STATE TOTAL PPs FREE PPs FREE DISTRIBUTION hdisk6 active 1075 1075 215 4215 215 275 215 hdisk7 active 1075 1072 215 212 215 215 215 lsvg l lowg lowvg LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT lowlv1 0 0 0 closed syned N A lowlv2 jfs 2 2 1 closed syncd lowfs2 lowlog jfslog 1 1 Ae closed syned N A J We can now safely rebuild the logical volumes from maps At this point we may choose to make our data in lowfs2 available Alternatively we may prefer to remount the file systems after LVM manipulation is complete Recall that early in this procedure we created lowlv1 map and lowlv2 map To repair the unmirrored lowlv1 we first remove the remnant zero size lowlv1 seen in the 1svg 1 with mlv We can run mk1v as we did in the final repair stage for Corruption example 2 PVID corruption on page 75 rmlv lowlv1 Warning all data contained on logical volume lowlv1 will be destroyed mlv Do you wish to continue y es n o y mlv Logical volume lowlv1 is removed mklv y lowlv1 m lowlvl map lowvg 3 lowlv1 An unfortunate side effect of the mk1v is that the LVCB on lowlv1 has been wiped We have to rebuild the entry for lowfs1 in etc filesystems from the tmp filesystems bak that was made earlier We then run chfs a log dev lowlog lowfs1 to update the LVCB The lowfs1 file system will then mount successfully Any lost logical volume polic
3. if tHaG t if 1 Cc leanp we have entered clear CRIAG c if FUG amp 1 m m ti y coment aut me handle printf extended lines slash substr 0 length 0 if slah multi y else m ti i if 1 we have left cleanup CFG print 0 if 1P amp 0 print set xv add debug line LP substr 0 length 0 1 emp Ivmtrace file orig gt tmp lvmtrace file if n SIFIAG then exec gt SIFITE 2 gt amp 1 fi gh xv tmp lvntrace file pams r m rf tnp 1trace retum Src 394 AIX Logical Volume Manager from A to Z Troubleshooting and Commands E 2 dspmsg_index This script adds a line after each dspmsg in a high level command giving the associated string literal for example 1 s Volume Group deleted since it contains no physical volumes and displays the script to stdout Some people may find this makes LVM scripts more readable This output cannot be used in place of the high level commands Syntax dspmsg lt filename gt Flags none 1 usr bin ksh awk print 0 if 0 dspmsg system print n print n lvmmsg 5 gt print yg E 3 chpvid chpvid sets the pvid on a disk to a given value This was useful in building the examples for this book and may be of use as a last resort in recovery situ
4. A 16 The mkcd command The following summarizes the options for the mkcd command mkcd Creates a multi volume CD or CDs from a mksysb or savevg backup image Usage mkcd d cd_device S m mksysb_imagel M mksysb_target s savevg_image v savevg_vol_group Ccd_fs_dir cd_image_dir V cdfs_vol_group G B p pkg_source_dir R S i image data u bosinst data e P l package_list b bundle_file z custom_file D 238 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Prevents mkcd from adding boot images non bootable CD to the CD Use this flag if creating a mksysb CD that you will not boot Before installing the non bootable mksysb CD you must boot a same level V R M product CD The mkcd command defaults to creating a bootable CD for the machine type of the source system See the Notes section b bundle_file Gives the full pathname of the file containing a list of filesets to be installed after the mksysb is restored This file is copied to usr sys inst data user_bundles bundle_file in the CD file system and also copied to RAM in case the CD is unmounted The file would be listed as BUNDLES usr sys inst data user_bundles bundle_file in the bosinst data file C cd_fs_dir Specifies the file system used to create the CD file system structure which must have at least 640 MB of available disk space If you do not specify the c fla
5. type generic rep nls index 0 a x usr bin tcbck we do nothing here as tcb is tumed off imfs lx victimlog victimlog is a jfs log not a jfs file system so this does nothing read LVNAME LVID LOCKED LOCKED oy At this point we have the logical volume and volume group information still remaining in the ODM We also have the entries in dev for the volume group and the logical volume devices Removing these is essentially what exportvg does This is accomplished simply by the putlvodm G in the following code Removing a volume group also handles the dependent logical volumes and all the associated devices It is not uncommon for so much functionality to be concentrated in a single line of the high level commands These scripts act as drivers to the powerful low level commands Once you have identified these key low level sections in the high level scripts the scripts become much easier to understand and debug One can look at the surrounding checking code and by identifying what check is not being passed find the problem area to recover remove all the logical volume information and the VGNAME from the ODM putlvodm G SVGNAME gt dev null 2 gt amp 1 putlvodm G victimvg 1 gt dev null 2 gt amp 1 If putlvodm failed output warning and continue if ne 0 then dspmsg s 1 cmdlvm cat 894 lvmmsg 894 n exportvg SVGNAME gt amp 2 fi
6. What other subjects would you like to see IBM Redbooks address Please rate your overall O Very Good OGood O Average O Poor satisfaction Please identify yourself as O Customer O Business Partner O Solution Developer belonging to one of the following groups O IBM Lotus or Tivoli Employee O None of the above Your email address The data you provide here may be used to provide you with information from IBM or our business partners about our products services or activities O Please do not use the information collected here for future marketing or promotional contacts or other communications beyond the scope of this transaction Questions about IBM s privacy policy The following link explains how we protect your personal information http www ibm com privacy yourprivacy Copyright IBM Corp 2000 415 AIX Logical Volume Manager from A to Z Troubleshooting and Commands SG24 5433 00 Wip Printed in the U S A SG24 5433 00
7. c Copies Sets the number of physical partitions allocated for each logical partition The Copies variable can be set to a value from 1 to 3 the default is 1 d Schedule Sets the scheduling policy when more than one logical partition is written The Schedule variable can be one of the following p Establishes a parallel scheduling policy This is the default for scheduling policy S Establishes a sequential scheduling policy 244 AIX Logical Volume Manager from A to Z Troubleshooting and Commands e Range Sets the inter physical volume allocation policy the number of physical volumes to extend across using the volumes that provide the best allocation The Range value is limited by the UpperBound variable set with the u flag and can be one of the following x Allocates across the maximum number of physical volumes m Allocates logical partitions across the minimum number of physical volumes This is the default range G Groupid Specifies group ID for the logical volume special file Reads the PhysicalVolume parameter from standard input Use the i flag only when PhysicalVolume is entered through standard input Sets the logical volume label The default label is None The maximum size of the label file is 127 characters Note If the logical volume is going to be used as a journaled file system JFS then the JFS will use this field to store the mount point of the file system on that logical vol
8. if tVAL is specified it should be large enough to fit MAX TOTPPS if n S tVAL then if StVAL lt SMINTVAL then dspmsg s 1 cmdlvm cat 1158 lvmmsg 1158 n chvg StVAL Sname SMINTVAL gt amp 2 errhandler return fi else errhandler dspmsg s 1 cmdlvm cat 732 lvmmsg 732 n chvg name gt amp 2 EXIT _CODE 2 J 64 AIX Logical Volume Manager from A to Z Troubleshooting and Commands If we look at the code where convert1016 was called we can see that this was not in fact fatal EXIT_CODE of 2 is often used by the high level scripts to report a partial success chvg will continue attempting to fulfill whatever other requests were made of it in this case the c flag to make the volume group concurrent capable After a high level script has run through its tasks it will typically reset the EXIT_CODE to zero to indicate success if appropriate and run exit Since the scripts initially set a korn shell trap with a line such as trap cleanup 0 1 2 15 shortly after the script begins the execution of its main function not explicitly defined as a main in korn shell instead indicated by a comment line cleanup will now be executed Note Since we are trapping signals 0 dummy 1 SIGHUP 2 SIGINT and 15 SIGTERM it is best to terminate an LVM command by sending it a SIGTERM kill 15 process ID if it cannot be run through to completion It is certainly undesirable to send a SI
9. 0 ne 0 J We run the above check to offer a warning message as if something has gone wrong the putlvodm may still have partially succeeded and we don t Chapter 2 Problem determination and recovery 119 want to roll back from this situation An error at this stage is better investigated manually rather than via scripting Displaying a message of the form Warning cannot remove volume group 2 s from device configuration database alerts the administrator to investigate and perform problem determination If the put1vodm succeeded nothing really remains of the volume group from AIX s point of view at this moment We know there can t have been a reference to it in the kernel as we couldn t have varied it off a pre requisite for exportvg that we saw above if there had been Now all that remains is to tidy up the trusted computing base and remove the map file in etc vg remove VG entry frm TCB database if TB is Qn D if z IC B OFF a x usr bin tdock then tdock d dev SVGNAME gt dev null 2 gt amp 1 fi z POL uniguetype attribute ICB SIATE deflt tcb disabled values width type m generic reo nls index 0 a x usr bin tdock J remove mapped file bin m f etc va vo echo SVGID tr a z A Z gt gt dev mill 2 gt 82 tr a z A Z echo 00017d37 b7ee896 bin m f etc vg v9Q000170376B7EEB96 1 gt dev nill 2 gt 1 EXI
10. Recovering from Disk Drive Problems They are also discussed in Chapter 1 LVM commands on page 1 From the point of view of software problem determination the investigator is usually faced with the problem of how to clear up the ODM and possibly the VGDAs after these procedures have been incorrectly performed This can be treated as a normal VGDA or ODM corruption situation The hardware issue that deserves special attention because of its tendency to generate long and hard to debug problems is when a disk has had a stroke The disk still gives the appearance of working but recurrent LVM problems or data corruption occur This a comparatively rare situation but it is worth mentioning If a disk does not certify when diagnostics are run then this should be suspected and hardware support should become involved The procedure to run diagnostics on a disk is as follows the diagnostic menus are entered with the diag command Chapter 2 Problem determination and recovery 181 FUNCTION SELECTION 801002 Move cursor to selection then press Enter Diagnostic Routines This selection will test the machine hardware Wrap plugs and other advanced functions will not be used Advanced Diagnostics Routines This selection will test the machine hardware Wrap plugs and other advanced functions will be used Task Selection Diagnostics Advanced Diagnostics Service Aids etc This selection will list the tasks supported
11. g VGID The volume group ID p PVID The physical volume ID r PVState The remove or return state of the physical volume The values are as follows 1 PV temporarily REMOVED from VG 2 PV RETURNED to VG Note that returning a PV to the VG requires execution of lvaryonvg to perform recovery and thereby activate the returned Physical Volume a AllocState The Allocate Noallocate state of the Physical Volume The values are as follows 4 Partitions in PV cannot be allocated for LV NOALLOCPYV 8 Partitions in PV can be allocated for LV ALLOCPV Appendix B Intermediate level commands 295 B 12 The Icreately command The following summarizes the options for the 1createlv command Icreatelv An object file command that creates an empty logical volume that belongs to a specified volume group Usage Icreately N LVname g VGID n MinorNum M MirrorPolicy s MaxLPts p Permissions r Relocation v WriteVerify w MWConsistency N LVname The logical volume name g VGID The volume group ID n MinorNum The minor number assigned to the logical volume s MaxLPts The maximum number of logical partitions allowed for this logical volume an integer between 1 and LVM_MAXLPS 65535 Note this is not the number of logical partitions allocated M MirrorPolicy The mirror policy The value is as follows 1 Sequential mirroring 2 Parallel mirroring p Permission Set
12. importvg y workvg hdisk5 PV Status hdisk5 00017d372f1de834 PVACTIVE hdisk6 0000000000000000 INVPVID 00017d372 67da78 NAMIDMTCH varyonvg Volume group workvg is varied on workvg PV Status hdisk5 00017d372f1de834 PVACTIVE hdisk6 0000000000000000 INVPVID 00017d372 67da78 NAMIDMTCH varyonvg Volume group workvg is varied on mount workfs mount 0506 334 workfs is not a known file system grep p workfs etc filesystems Ke J We are now in an even worse position than before AIX no longer even knows a file system called workfs should exist The mount information log device and so forth have been removed from etc filesystems In this case we have only lost one file system entry However if our volume group had contained many small file systems lets say we are using a database product that doesn t use raw logical volumes the impact of manually re adding them would be more severe Now that the point is made the question becomes How to recover from this situation We begin the information gathering process again First we check the state of our VGDAs with 1queryva Chapter 2 Problem determination and recovery 79 1queryvg to check syntax 0516 162 lqueryvg VG identifier or PV name not entered Usage lqueryvg g VGid p PVname NsFncDaLPAvt 1queryvg g getlvodm v workvg At p hdisk5 Conc Autovar Varied on Co Conc Autovar Varied on Co Max LVs 256 PP Size 22 Free PPs 21
13. m Allocates logical partitions across the minimum number of physical volumes This is the default range k Synchronizes data in the new partitions Appendix A High level LVM commands 251 m MapFile Specifies the exact physical partitions to allocate Partitions are used in the order given in the MapFile parameter Used partitions in the MapFile parameter are not legal since the new logical volume cannot occupy the same physical space as a previously allocated logical volume All physical partitions belonging to a copy are allocated before allocating for the next copy of the logical volume The MapFile parameter format is PVname PPnum1 PPnum2 In this example PVname is a physical volume name for example hdiskO as specified by the system It is one record per physical partition or a range of consecutive physical partitions PPnum is the physical partition number PVname Name of the physical volume as specified by the system PPnum Physical partition number s Strict Determines the strict allocation policy Copies of a logical partition can be allocated to share or not to share the same physical volume The Strict parameter is represented by one of the following y Sets a strict allocation policy so that copies for a logical partition cannot share the same physical volume This is the default for allocation policy n Does not set a strict allocation policy so that copies for a logical partition can
14. t L LVID The LVID of the logical volume that is to be queried p PVName The name of the physical volume containing the VGDA N Returns the logical volume name 300 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Returns the VGID n Returns the maximum number of logical partitions M Returns the Logical Volume mirror write policy 1 Sequential mirroring 2 Parallel mirroring S Returns the current state of the logical volume C Returns the current size in logical partitions S Returns the physical partition size P Returns the logical volume permission attribute 1 Read write 2 Read only R Returns the bad block reallocation attribute 1 Bad block reallocation 2 No bad block reallocation V Returns the write verify state 1 Write verify is on 2 Write verify is off O Returns the open closed state 1 Open 2 Closed W Returns the mirror write consistency state 1 Mirror write consistency is on 2 Mirror write consistency is off a Returns all the static attributes of the logical volume LVName VGID MaxLP MirrorPolicy MWConsistency LVState LVSize PPSize PermissionFlag ReallocateFlag WriteVerify OpenClose StripeExp StripeWidth NumCopies BkMirrorCopy without headings d Returns the dynamic attribute of the logical volume The logical partition map Returns each copies of the mirror in the partition ma
15. y d BlockNumber Searches for references to a specified disk block Whenever the fsck command encounters a file that contains a specified block it displays the inode number and all path names that refer to it f Performs a fast check Under normal circumstances the only file systems likely to be affected by halting the system without shutting down properly are those that are mounted when the system stops The f flag prompts the fsck command not to check file systems that were unmounted successfully The fsck command determines this by inspecting the s_fmod flag in the file system superblock This flag is set whenever a file system is mounted and cleared when it is unmounted successfully If a file system is unmounted successfully it is unlikely to have any problems Because most file systems are unmounted successfully not checking those file systems can reduce the checking time i InodeNumber Searches for references to a specified i node Whenever the fsck command encounters a directory reference to a specified i node it displays the full path name of the reference Appendix D Other related commands 341 Assumes a no response to all questions asked by the fsck command and does not open the specified file system for writing o Options Passes comma separated options to the fsck command These options are assumed to be file system implementation specific except that the following are currently supp
16. An alternative method to resolve this second situation is importvg y fullvg hdisk9 94 AIX Logical Volume Manager from A to Z Troubleshooting and Commands lsvg o 0516 304 Unable to find device id 00017d374c013elfin the Device Configuration Database vgid 00017d374c013el1f rootvg Isvg rootvg importvg y fullvg hdisk9 fullvg lsvg o rootvg fullvg Li J However this assumes we know which lost volume group is giving the phantom entry This could be obtained by searching disks which are not imported for the phantom VGID using low level commands for instance lqueryvg g 00017d374c013e1f these are discussed in The low level commands on page 107 Using importvg is preferable since it gives the benefit of checking that it is built into the high level scripts 2 6 2 2 etc Some commands also have a dependency on free file system space in the file system as volume group lock files are created in etc vg At these levels the absence of etc vg also gave rise to 0516 028 internal mapfile errors If etc is full we may also experience problems adding or removing ODM objects in etc objrepos An example of this is as follows varyoffvg fullvg 0516 362 putlvodm Unknown Object Data Manager error 0 0516 942 varyoffvg Unable to vary off volume group fullvg It is prudent system management to keep some space free in all rootvg system file systems 2 6 2 3 dev A typical symp
17. If the disk certifies correctly you will see the following la CERTIFY MEDIA TASK 802548 hdisk4 in location 30 58 00 9 0 The certify operation has completed successfully Disk Drive Capacity 4512 MB Data Errors Recovered 0 Data Errors Not Recovered 0 Equipment Check Errors Recovered 0 Equipment Check Errors Not Recovered 0 To continue press Enter x F3 Cancel Esc 0 Exit Enter 184 AIX Logical Volume Manager from A to Z Troubleshooting and Commands If the certify fails hardware problem determination should now be performed Sometimes people get confused about the difference between certify and format Certify will check the disk format will erase all information from the disk This is only way to be certain that confidential information is erased from a disk You cannot dd over the raw disk because some information may still be on the disk as relocated over bad blocks Note that if you format over the disk this also means you will lose the bad block relocation tables It is also important to realize that certify is an attempt to fail the disk This attempt may not succeed You can only use this to prove that a disk is bad not that a disk is good If more extensive testing of a disk is required a usr lib ras rdwrtest rdwrtest should be run Note this will permanently erase any information in the logical volume being tested The accompanying file RE
18. LPs PPs Distribution Mount Name of the logical volume to which the physical partitions are allocated The number of logical partitions within the logical volume that are contained on this physical volume The number of physical partitions within the logical volume that are contained on this physical volume The number of physical partitions belonging to the logical volume that are allocated within each of the following sections of the physical volume Outer edge outer middle center inner middle and inner edge of the physical volume Point File system mount point for the logical volume if applicable Appendix A High level LVM commands 225 M Lists the following fields for each logical volume on the physical volume PVname PPnum LVname LPnum Copynun PPstate Where PVname Name of the physical volume as specified by the system PPnum Physical partition number LVname Name of the logical volume to which the physical partitions are allocated Logical volume names must be system wide unique names and can range from one to 64 characters LPnum Logical partition number Logical partition numbers can range from one to 64 000 Copynum_ Mirror number PPstate Only the physical partitions on the physical volume that are not current are shown as stale n Accesses information from the variable descriptor area specified by the DescriptorPhysicalVolume variable The information may not b
19. Similar to B flag except that space is reserved in the VGSA VGDA for up to 1024 physical volumes Should be used if the volume group is likely to be expanded to include more than 128 physical volumes Reads the PhysicalVolume parameter from standard input m MaxPvSize Specifies the maximum size of the physical volume s When this flag is used the calculation of the number of physical partitions are based on the size of the physical volume If this flag is not specified and the t flag not used then 1016 physical partitions are assumed n Specifies that the volume group is not automatically varied on at system startup s Size Sets the physical partition size for the volume group where the size is in megabytes and must be a power of 2 from 1 to 1024 the default is 4 t factor Changes the limit of the number of physical partitions per physical volume where the number of physical volumes is the factor times 1016 The range is from 1 to 16 for normal VGDA systems and 1 to 64 for big VGDA systems Increasing this factor will reduce the number of physical volumes allowed to make up the volume group see Chapter 1on components in AIX Logical Volume Manager from A to Z Introduction and Concepts SG24 5432 V MajorNumber Specifies the major number that will be used by the volume group Specifies that the volume group will be automatically varied on at system start Similar to c this flag has no affec
20. The final issue with mirroring discussed here is that attempts to remove mirrored copies of a logical volume may sometimes fail with the error message number 0516 912 particularly if a systems crash occurred while a first attempt was being made We have already seen a procedure to target specific areas of a logical volume with lquerylv r then lreducelv followed by ldeletelv This is relevant here Chapter 2 Problem determination andrecovery 177 178 LVID lqueryvg Lp hdisk8 grep stalelv cut c 1 187 lquerylv L LVID r grep getlvodm p hdisk8 gt stalemap lreducelv 1 SLVID s we l stalemap stalemap 2 7 2 2 LVCBs We have already seen an example of rebuilding LVCBs with chfs and putlvcb in 2 5 Corruption example 2 PVID corruption on page 75 Remember that whenever we run a mklv when restoring from maps the LVM will repopulate the LVCB with the default values Historically the main cause of LVCB corruption has generally been databases which are writing to raw logical volume being unaware that the first 512 bytes are reserved for the logical volume control block Obviously if a database has done this it is very unwise to simply overwrite a good LVCB into the beginning of the logical volume It is very likely this will cause the database to become corrupt This issue should be taken up with the database vendor Fortunately none of the latest versions of the popular databases have this proble
21. The following summarizes the options for the 1queryvgs command Appendix B Intermediate level commands 303 Iqueryvgs An object file command that queries the IDs of all volume groups in the system It is assumed that there is only one system in the environment Usage Iqueryvgs m ModulelD N G A t m ModulelD Kernel Module ID of the logical volume device drive This is returned when the LVDD is loaded N Returns the number of volume groups created so far in the system G Returns the VGIDs and major numbers of the volume groups in the system A Returns the number of volume groups and each VGID and major number t Includes tags in the output of the above options B 24 The Ireducelv command The following summarizes the options for the 1reducelv command lreducelv An object file command that reduces the number of allocated logical partitions in a logical volume Usage Ireducelv I LVID s Size filename I LVID The LVID of the logical volume that will be reduced s Size The number of logical partitions to deallocate from the logical volume Must be less than the current number of logical partitions in the logical volume filename The file that contains the map of the partitions that will be deallocated The file has an entry for each of the partitions which is a triplet containing Physical Volume ID Physical Partition Number Logical Partition Nu
22. corrvg 0516 304 lsvg Unable to find device id 0000000000000000 in the Device Configuration Database 0516 022 lsvg Illegal parameter or structure value lsvg 1 corr corr LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT 0516 1147 lsvg Warning logical volume corrlv may be partially mirrored corrlv jis 3 0 1 closed synced N A lsvg p corr 0516 022 lsvg Illegal parameter or structure value 0516 304 1lsvg Unable to find device id 0000000000000000 in the Device Configuration Database corr Pv ID PV STATE TOTAL PPS FREE PPs FREE DISTRIBUTION 0000000000000000 2272777 272777 77 lslv 1 corrlv 0516 022 lslv Illegal parameter or structure value 0516 304 1lslv Unable to find device id 0000000000000000 in the Device Configuration Database corrlv N A PVID COPIES IN BAND DISTRIBUTION 0000000000000000 003 000 000 pi dd if dev zero seek 128 count 2106 of dev hdisk6 2106 0 records in 2106 0 records out lsvg M corrvg corrvg hdisk6 1 215 hdisk6 216 corrlv 1 hdisk6 217 corrlv 2 hdisk6 218 corrlv 3 hdisk6 219 1075 J Another possible symptom of partition map corruption is nonsensical map output lsvg M corrvg corrvg hdisk6 1 4 hdisk6 5 RESERVED 1919492161 stale 0516 304 lsvg Unable to find device id 00017d378ac4e080 83 in the Device Configuration Database This is shown here for recognition purposes AIX Logical
23. usr bin etc usr sbin usr uco EXIT CODE 1 Initialize exit code This will be reset to 0 before EXIT CODE 1 C exiting ally if exportvg completes successfully J Chapter 2 Problem determination and recovery 111 We have just begun running commands We set up the PATH and initialize the EXIT_CODE This EXIT_CODE logic is common to many of the high level commands a N Trap on exit interrupt break to clean up trap cleanup 0 1 2 15 trap cleanup 0 1 2 15 PROGNAME basename 0 basename tmp lvmtrace18118 exportvg PROGNAME exportvg ODMDIR etc obj repos ODMDIR etc obj repos export ODMDIR _ export ODMDIR Note that scpmprr is hard coded into the LVM scripts We now begin the parsing process This procedure is present in all of the scripts First any flags used are parsed and then the command line arguments are parsed In exportvg the parsing of flags is somewhat trivial since unusually among the high level commands exportvg has no possible flags Thus there is no actual code to parse for flags For purposes of explanation a quick diversion is made into the importvg script Flag parsing is a loop where internal variables of the form letter FLAG are set up based on the flags with which the high level command was called An associated letter VAL may also be set if a value was passed with the flag These flags may come from the user input or even from
24. 00 root itsosrv1 home dug scripts 449 lspv n hdiskl hdisk4 0516 022 lspv Illegal parameter or structure value PHYSICAL VOLUME hdisk4 VOLUME GROUP rootvg PV IDENTIFIER 00091974d381d097 VG IDENTIFIER 00017d37e1762ac7 PV STATE P P P P STALE PARTITIONS P P P P ALLOCATABLE P2 P PP SIZE P P P P P LOGICAL VOLUMES TOTAL PPs AREENA VG DESCRIPTORS PERRE EF FREE PPs USED PPs 7 FREE DISTRIBUTION USED DISTRIBUTION Ft J We see this because the VGDA on hdisk1 does not know anything about hdisk4 This can be a useful test for corruption If it is suspected that the VGID in the ODM has become out of sync the v flag allows us to try using other VGID values to look at the VGDA on the disk 2 6 3 2 Isvg 1svg will show if the ODM knows about a volume group 1svg o will tell us which volume groups have been varied on 1svg also has the n flag to target a specific descriptor area 1svg M can be used to dump the complete set of maps for a volume group L can be used for investigation without explicitly breaking the ODM lock 2 6 3 3 Islv Again L ignores an ODM lock 1slv m will show us the map files for a particular logical volume Chapter 2 Problem determination and recovery 97 Note It is often useful in recovery situations to have the map files for the logical volumes on the disks Sometimes it is not possible to determine the maps after corruption has occ
25. 0516 306 getlvodm Unable to find othervg in the Device Configuration Database 0516 942 varyoffvg Unable to vary off volume group othervg varyoffvg lowvg varyonvg lowvg PV Status hdisk6 00017d377c3abace INVPVID hdisk7 00017d377b71d55c PVACTIVE varyonvg Volume group lowvg is varied on usr sbin syncvg 16 18604 Segmentation fault coredump _ usr sbin syncvg 18 18606 Segmentation fault coredump J Now we no longer have access to our data and when we try to pull maps out we just get core dumps If we look at the VGDAs with oa we can see the following dd skip 128 if dev hdisk6 count 10 2 gt dev null od cx grep 11000 0011000 o o o o o 002 o 001 o o 002 o o o o dd skip 128 if dev hdisk7 count 10 2 gt dev null od cx grep 11000 0011000 2 o o o o o 002 o 001 o o 002 o o o o The corruption the first question mark character in each line has propagated across from the VGDA on hdisk6 to hdisk7 All the VGDAs are now corrupt and we are in a deadlock situation This propagation occurred as soon as we ran the first 1reducelv The VGDA wheel the mechanism for keeping all the VGDAs in sync was spun and the corruption spread across all the disks Once this has happened our only option would be to completely destroy the volume group and remake it and its logical volumes from maps This is unlikely to happen when working with high level commands as the error checkin
26. 1 1 6 Bring a volume group into the system The importvg command makes the previously exported volume group known to the system Only one physical volume is needed to identify the volume group any remaining physical volumes those belonging to the same volume group are found by the importvg command and are included in the import An imported volume group is automatically varied on unless the volume group is Concurrent Capable You must use the varyonvg command to activate Concurrent Capable volume groups before you access them When a volume group with file systems is imported the etc filesystems file is updated with values for the new logical volumes and mount points After importing the volume group and activating it with the varyonvg command you must run the fsck command before the file systems can be mounted The importvg command changes the name of a logical volume if the name already exists in the system It prints a message and the new name to standard error and updates the etc filesystems file to include the new logical volume name 10 AIX Logical Volume Manager from A to Z Troubleshooting and Commands a Note You may import an AIX Version 3 2 created volume group into an AIX Version 4 system and you may import an AIX Version 4 volume group into an AIX Version 3 2 system provided striping has not been applied Once striping is put onto a disk its importation into version 3 2 is prevented When you issue the importvg
27. For example the 2 flag is the same as the f dev rmi2 file Note Because of limitations on header block space in the tar command user numbers UIDs and group identification numbers GIDs larger than 65 535 will be corrupted when restored to certain systems The size constraint affects only the ownership and permissions causing no damage to the data Corruption of the ownership occurs on the following systems e Those that do not use uname and gname fields to check ownership e Those that do not have the same user and group IDs as the archiving system Appendix D Other related commands 389 m Note 1 The tar command is not enabled for files greater than 2 GB in size due to limitations imposed by XPG 4 and POSIX 2 standards 2 tar does not preserve the sparse nature of any file that is sparsely allocated Any file that was originally sparse before the restoration will have all space allocated within the file system for the size of the file The tar command manipulates archives by writing files to or retrieving files from an archive storage medium The files used by the tar command are represented by the File parameter If the File parameter refers to a directory then that directory and recursively all files and directories within it are referenced as well The tar command looks for archives on the default device usually tape unless you specify another device with the Archive
28. Other related commands 323 Updates the etc dumpdates file with the raw device name of the file system and the time date and level of the backup You must specify the u flag if you are making incremental backups The u flag applies only to backups by i node Causes the backup command to display additional information about the backup When using the v flag the size of the file as it exists on the archive is displayed in bytes Additionally a total of these file sizes is displayed when all files have been processed Directories are listed with a size of 0 Symbolic links are listed with the size of the symbolic link Hard links are listed with the size of the file which is how hard links are archived Block and character devices if they were backed up are listed with a size of 0 When the v flag is not specified the backup command displays only the names of the files being archived This option is used only when backing up by file name Currently disabled If the w flag is specified no other flags are applied W Displays for each file system in the etc dumpdates file the most recent backup date and level If the w option is specified no other flags are applied Level Specifies the backup level 0 to 9 The default level is 9 The backup command creates copies of your files on a backup medium such as a magnetic tape or diskette The copies are in one of the two backup e Specific f
29. PV Status hdisk6 00091974de731316 PVACTIVE hdisk9 00017d37297a96c8 PVMISSING 0516 056 varyonvg The volume group is not varied on because a physical volume is marked missing Run diagnostics varyonvg f workvg In such a situation we can regain access to our data by running varyonvg f Once the volume group is back online we can reducevg away the failed disks and extendvg new disks back in as required 2 6 2 Checking for free file system space The LVM commands can hit problems if we are running out of space in rootvg There are particular problems associated with full tmp or the root file system subdirectories etc and dev 2 6 2 1 tmp Most high level LVM commands require having free space in tmp This is used for storing various temporary files which can be saved for later viewing with trclvm t If this space is exhausted commands will begin to malfunction in various ways as shown in the following C N mkvg y fullvg hdisk9 0516 014 lcreatevg The physical volume appears to belong to another volume group 00017d3743cf2892 0516 631 mkvg Warning all data belonging to physical volume hdisk9 will be destroyed mkvg Do you wish to continue y es n o y 0516 029 lcreatevg The Physical Volume is a member of a currently varied on Volume Group and this cannot be overridden 00017d3743cf2892 0516 862 mkvg Unable to create volume group d tmp Filesystem 512 blocks Free Used Tused Iused Moun
30. Replacing a drive in a mirrored configuration 193 194 lgpv hdisko hdisk1 hdisk2 hdisk3 hdisk4 lsvg testvg VOLUME GROUP VG STATE VG PERMISSION MAX LVs LVs OPEN LVs TOTAL PVs STALE PVs ACTIVE PVs Ne 0004163128f3de5a 000416314bd724be 0004163192b1d7 5 000416314bd749 8 000416314bdada38 testvg active read write 256 0 0 3 0 3 MAX PPs per PV 1016 rootvg testvg testvg None testvg VG IDENTIFIER PP SIZE TOTAL PPs FREE PPs USED PPs QUORUM VG DESCRIPTORS STALE PPs AUTO ON MAX PVs 00041631d7475818 8 megabyte s 1611 12888 megabytes 1611 12888 megabytes 0 0 megabytes 2 3 0 yes 32 S The next step is to create a mirrored and striped logical volume in this volume group Our logical volume testlv has a stripe width of 3 and has two copies The following screen shows the characteristics and position of testlv AIX Logical Volume Manager from A to Z Troubleshooting and Commands lslv testlv LOGICAL VOLUME testlv VOLUME GROUP testvg LV IDENTIFIER 00041631d7475818 1 PERMISSION read write VG STATE active complete V STATE closed syncd TYPE jfs WRITE VERIFY off MAX LPs 512 PP SIZE 8 megabyte s COPIES 2 SCHED POLICY parallel
31. Sun Sep 12 13 00 37 1999 J Note that the fs field in the LVCB has been erased for 1vo2 While the check and free entries are not mandatory problems will occur if the 1og field is blank We must also rebuild the label field as it contains mount point information Clearly these will not be a concern for non JFS file systems or raw logical volumes Not fixing these will cause problems most painfully in highly available environments We can remake these with chfs Chapter 2 Problem determination and recovery 85 fo chfs a log dev loglv00 workfs getlvcb AT 1v02 ATX LVCB intrapolicy m copies 1 interpolicy m lvid 00017d37389db50b 2 lvname 1v02 label workfs machine id 17D374C00 number lps 4 relocatable y strict y stripe width 0 stripe size in exponent 0 type jfs upperbound 32 fs log dev loglv00 account false time created Tue Sep 14 08 52 18 1999 time modified Tue Sep 14 10 06 26 1999 L J Note that the chfs also re created the following ODM entry N odmget CuAt grep workfs Cuat name lv02 attribute label value workfs type R generic DU rep s nls index 640 SSL p We have now restored our data and also fixed the relevant LVM and ODM control data Although we could have done this by running putlvcb and performing an odmadad it is safest to stick to as high a level of command as possible Low
32. T flag When using the x r m or t flags the restore command automatically determines the archive format Note The restore command actively sparses files that are being restored If a file has block aligned and sized areas that are NULL populated then restore does not cause physical space for those file system blocks to be allocated The size in bytes of the file remain the same but the actual space taken within the file system is only for the non NULL areas Individual files can be restored from either file name or file system archives by using the x flag and specifying the file name The file name must be specified as it exists on the archive Files can be restored interactively from file system archives using the i flag The names of the files on an archive can be written to standard output using the t flag Users must have write access to the file system device or have Restore authorization in order to extract the contents of the archive The diskette device dev rfdO is the default media for the restore command To restore from standard input specify a dash with the flag You can also specify a range of devices such as dev rmt0 2 374 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note 1 If you are restoring from a multiple volume archive the restore command reads the volume mounted prompts you for the next volume and waits for your response After inserting
33. and n flags during execution Restrictions 1 The volume group must not be in an active state on the system executing the 1 flag 2 The volume group s disks must be unlocked on all systems that have the volume group varied on and operational Volume groups and their disks may be unlocked remain active and used via the varyonvg b u command 3 The physical volume name provided must be of a good and known state The disk named may not be in the missing or removed state 4 If an active node has both added AND deleted logical volumes on the volume group the L flag may produce inconsistent results The L flag should be used after each addition or deletion rather than being deferred until after a sequence of changes 5 If a logical volume name clash is detected the command will fail Unlike the basic importvg actions clashing logical volume names will not be renamed Provides a fast version of importvg that checks the Volume Group Descriptor Areas of only the disks that are members of the same volume group As a result if a user exercises this flag they must ensure that all physical volumes in the volume group are in a good and known state If this flag is used on a volume group where a disk may be in missing or removed state the command may fail or the results may be inconsistent Causes the volume not to be varied at the completion of the volume group import into the system Appendix A High level
34. dd bs 1 seek echo ib 8 n200000 11000 be of dev hdisk6 lsvg M corr Segmentation fault coredump lquerylv L VGID 1 r 00017d3787995016 216 1 00017d3787995016 217 2 00017d3787995016 218 3 lquerylv L VGID 1 rlawk print hdisk6 2 hdisk6 216 hdisk6 217 hdisk 6 218 Ld J Even if the first VGDA is completely removed and the second VGDA is corrupted lquerylv will still work and LVM will go to the second VGDA T dd if dev zero seek 128 count 2106 of dev hdisk6 2106 0 records in 2106 0 records out lsvg M corrvg corrvg hdisk6 1 215 hdisk6 216 corrlv 1 hdisk6 217 corrlv 2 hdisk6 218 corrlv 3 hdisk6 219 1075 A However if you have a second VGDA that is good and a first VGDA that is bad LVM will go the first VGDA and return incorrect information In the following example we corrupt the partition maps on the VGDA This can be resolved by removing the first VGDA after taking a backup copy LVM will then acquire maps from the second VGDA This procedure is very dangerous You should work with an IBM AlX support center before doing this since you need to know that the second VGDA is good The numbers used in all these examples were for a non big VG on a specific level of AIX There is no guarantee they will stay the same Chapter 2 Problem determination and recovery 157 158 a dd if dev hdl bs 256 seek 306 count 127 of dev hdisk6 lsvg M corrvg
35. eNinCopies gt 1 then Islv m SIV lank NR gt 2 amp 4 l printf s s n 5 4 gt SIV cqp2 if aNincopies 2 than lslv m V ak NR gt 2 amp amp 6 printf s s n 7 6 gt SV 03 fi fi V J 396 AIX Logical Volume Manager from A to Z Troubleshooting and Commands E 5 findlvm findlvm checks the basic ODM classes relevant to the LVM for a particular string Syntax findlvm lt string gt Flags none 1 usr bin ksh for class in CuAt CuDv CuDep CuDvDr PdAt PdDv do odmget class grep ip 1 done E 6 maker This short script creates the example volume groups logical volumes and file systems used for Corruption example 3 Low level VGDA corruption on page 146 Syntax maker lt 1st hdisk gt lt 2nd hdisk gt Flags none usr bin ksh gt ISK1 1 HDISK2 2 umount lowfs1 umount lowfs2 varyoffvg lowg exportvg lowvg echo maker ignore non fatal errors before this point mkvg t2 y lowvg SHDISK1 SHDISK2 mklv y lowlvl lowvg 3 mklv c 2 y lowlv2 lowvg 2 mklv t jfslog y lowlog lowvg 1 crfs d dev lowlvl a logname dev lowlog m lowfsl1 v jfs crfs d dev lowlv2 a logname dev lowlog m lowfs2 v jfs echo y logform dev lowlog mount lowfs1 mount lowfs2 echo mydatal gt lowfsl datal echo mydata2 gt lowfs2 data2 J m Appendix E Scripts used during this r
36. each logical partition when allocating Sequential or parallel scheduling policy Number of logical partitions currently in the logical volume Number of physical partitions currently in the logical volume AIX Logical Volume Manager from A to Z Troubleshooting and Commands Stale partitions Number of physical partitions in the logical volume that are not current BB policy Bad block relocation policy Inter policy Inter physical allocation policy Relocatable Indicates whether the partitions can be relocated if a reorganization of partition allocation takes place Intra policy Intra physical allocation policy Upper bound If the logical volume is super strict upper bound is the maximum number of disks in a mirror copy Mount point File system mount point for the logical volume if applicable Label Specifies the label field for the logical volume Each LP copy on separate PV Current state of allocation strict nonstrict or superstrict Striping width The number of physical volumes being striped across 1 Strip size The number of bytes per stripe 1 Only displayed if the logical volume is striped 1 2 13 List a logical volumes detailed mapping The 1slv command will also display information about the logical volume s physical partition usage by region and physical volume or the actual logical to physical partition mapping As seen in the previous section the physical volume map or the logical part
37. gt amp 2 exit fi PARALLELFLAG P 2 shift shift HOLD passive node writes H HFLAG H shift esac done end while there is a command line option fella ves LFLAG 1 shift CASE y eee re t t 4 N We have come through flag parsing and now check for incompatibilities among the different flags we have set This checking is also a major component of other scripts such as mklv 174 AIX Logical Volume Manager from A to Z Troubleshooting and Commands if z SCASE then dspmsg s 1 cmdlvm cat 938 lvmmsg 938 gt n SPROG gt amp 2 dspmsg s 1 cmdlvm cat 1145 lvmmsg 1145 gt n PROG gt amp 2 exit fi zy if n a n SLFLAG o n PFLAG then dspmsg s 1 cmdlvm cat 937 lvmmsg 937 gt n SPROG gt amp 2 dspmsg s 1 cmdlvm cat 1145 lvmmsg 1145 gt n PROG gt amp 2 exit fi n a n l1 o n if n SLFLAG a n SVFLAG o n SPFLAG then dspmsg s 1 cmdlvm cat 937 lvmmsg 937 n PROG gt amp 2 dspmsg s 1 cmdlvm cat 1145 lvmmsg 1145 n PROG gt amp 2 exit fi n l a n o n if n PFLAG a n SLFLAG o n VFLAG then dspmsg s 1 cmdlvm cat 937 lvmmsg 937 n PROG gt amp 2 dspmsg s 1 cmdlvm cat 1145 lvmmsg 1145 n PROG gt amp 2 exit fi n a n l1 o n if n SHFLAG o n SPARA
38. since when the volume group was created by mklv hdisk6 would have been used first and therefore has two VGDAs hdisk9 was brought in next by extendvg and therefore has only one VGDA Not only is our data intact it is also still available This would be the case whether we had quorum set or not since two out of three VGDAs are still around However if the machine is rebooted or other actions happen to vary off and then attempt to vary on the volume group our data will no longer be available On reboot initial investigation in response to user complaints about missing data shows 90 AIX Logical Volume Manager from A to Z Troubleshooting and Commands mount workfs mount 0506 324 Cannot mount dev lv02 on workfs There is a request to a device or address that does not exist svg rootvg workvg software lsvg o software rootvg Li J Checking the error report will show an LvM_MISSPVADDED error missing physical volume before the reboot therefore allowing us to fix this situation before it becomes a problem We may also have had opmsc errors if we had error log analysis running Assuming these are SSA disks we would also have seen SSA_LINK OPEN The green SSA status lights will also be flashing on either side of the pulled disk hdisk9 s own lights will of course be dead After the reboot we will see the Lvm_MISSPVADDED error again link open errors appear again for SSA since the erro
39. y workfs Checking dev rlv02 workf 0 Check Log processing for dev rlv02 1 Check Blocks and Sizes 2 wm B w 6 Check Pathnames Check Connectivity Check Reference Counts Check Inode Map Check Block Map 9 files 1120 blocks 31648 free mount workfs workts lost found We have now regained access to our data It is important that we do not stop here Although the file system is back up the system is not in the state it was before corruption occurs If we compare the LVCBs of a known good file system for example home on hd1 and the repaired file system we see AIX Logical Volume Manager from A to Z Troubleshooting and Commands 4 getlvcb AT 1v02 ATX LVCB intrapolicy m copies 1 interpolicy m lvid 00017d37389db50b 2 lvname 1v02 label None machine id 17D374C00 number lps 4 relocatable y strict y stripe width 0 stripe size in exponent 0 type jfs upperbound 32 fs time created Tue Sep 14 08 52 18 1999 time modified Tue Sep 14 08 52 18 1999 getlvcb AT hd1 AIX LVCB intrapolicy c copies 1 interpolicy m lvid 00017d37e1762ac7 8 lvname hdl label home machine id 17D374C00 number lps 4 relocatable y strict y stripe width 0 stripe size in exponent 0 type jfs upperbound 32 fs log dev hd8 mount true check true vol home free false time created Fri Aug 27 21 49 20 1999 time modified
40. 1 Simple ODM corruption 66 2 4 Gathering information about the problem 0 5 75 2 5 Corruption example 2 PVID corruption 00 05 75 2 6 Inspection commands 00 0 es 87 2 6 1 Checking the errorlog 0 eee eee 88 2 6 2 Checking for free file system space 000 05 93 2 6 3 The high level commands 00 eee eee eee 96 2 6 4 Checking fileset levels 2 2 0 0 0 e eee eee ee 98 2 6 5 Checking device availability 20 002 e ee eee 98 2 6 6 Checking the ODM 2 eee ee 98 2 6 7 The low level commands 000 cece eee eee 107 2 6 8 korn shell debug 0 a e a ee 109 2 6 9 Examining raw data on physical volumes and logical volumes 121 2 6 10 Watching system calls 0 2 0 0 0 ee 127 2 6 11 Examining internal kernel memory structures with crash 127 2 7 Rebuilding and repai sser ss 0 0 2 ee ee 128 2 7 1 ODM corruption 0 0 0 ee 129 2 7 2 LVM control data corruption 0 00000 146 2 1 9 JES PlODIOMS chess esis bean alee ahi amp pope ate Main Wee 179 2 7 4 Hardware failures 0 2 0 es 181 2 7 5 Setting up notification of LYM_MISSPVADDED errors 185 2 8 Special considerations for rootvg 0 eee ee ees 187 2 8 1 rootvg problem determination in maintenance mode 187 2 8 2 Reducing the size of hd6 2 ees 187 Chapte
41. 134217728 512 byte blocks New File System size is 8192 lsvg l targetvg targetvg LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT loglv00 jfslog 2 2 1 closed syncd N A 1v02 jfs 1 1 1 closed syned targetfs chlv n targetlog loglv00 0516 712 chlv The chlv succeeded however chfs must now be run on every filesystem which references the old log name loglv00 chlv n targetlv 1v02 chfs a log dev targetlog targetfs lsvg l targetvg targetvg LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT targetlog jfslog 2 2 1 closed syncd N A targetlv jfs T 1 1 closed syned targetfs mount targetfs echo data gt targetfs data Li J Let s assume all the ODM information required for this volume group has been destroyed We can replicate this kind of situation by carefully running odmdelete always check what you are going to delete by running an odmget on the criteria first fe N odmdelete o CuAt q name LIKE target 0518 307 odmdelete 10 objects deleted odmdelete o CuDvDr q value3 LIKE target 0518 307 odmdelete 3 objects deleted odmdelete o CuDvDr q valuel targetvg 0518 307 odmdelete 1 objects deleted odmdelete o CuDv q name LIKE target 0518 307 odmdelete 3 objects deleted odmdelete o CuDep q name targetvg _ 0518 307 odmdelete 2 objects deleted J Now we look to see what we have done 130 AIX Logical Volume Manager fro
42. 276 syncvg 29 277 sysdumpdev 381 system management mode 7 T tar 384 timestamp 101 trace 127 trelvm 63 trusted computing base 118 120 U umount 391 unmirrorvg 277 updatelv 279 updatevg 279 V varyoffvg 6 280 varyonvg 3 5 93 280 VGDA 60 75 80 90 VGDA corruption 146 VGID 117 VGSA _ 60 75 volume group 2 changing name 13 concurrent access 4 concurrent capable 10 20 defined 16 413 exported 10 identifier 20 63 majornumber 4 maximum number of logical volumes 19 name 4 11 number of open logical volumes 19 number of physical volumes 19 permission 19 state 19 40 total number of physical partitions 20 Ww write verify 25 33 414 AIX Logical Volume Manager from A to Z Troubleshooting and Commands IBM Redbooks review Your feedback is valued by the Redbook authors In particular we are interested in situations where a Redbook made the difference in a task or problem you encountered Using one of the following methods please review the Redbook addressing value subject matter structure depth and quality as appropriate e Use the online Contact us review redbook form found at http www redbooks ibm com e Fax this form to USA International Access Code 1 914 432 8264 e Send your comments in an Internet note to redbook us ibm com Document Number Redbook Title SG24 5433 00 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Review
43. 415 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Preface LVM Logical Volume Manager What is its role in the AIX operating System How does it perform this function Its role is to control disk resources by mapping data between a more simple and flexible logical view of storage space and the actual physical disks How it performs this function is a topic vast enough to fill two books The first volume AIX Logical Volume Manager from A to Z Introduction and Concepts G24 5432 describes the basic components and defines physical volumes volume groups and logical volumes This second volume focuses on the practical aspects of the Logical Volume Manager and describes in detail the commands including some undocumented commands and options The last section of this volume is on troubleshooting and provides hints on how to handle certain problems how to investigate them and what commands are needed to run and solve these problems Most importantly this volume discusses what commands not to run in order to avoid ruining your entire system This book is aimed at every IT specialist who wants to know more about the core element of AIX which is the Logical Volume Manager The team that wrote this redbook This redbook was produced by a team of specialists from around the world working at the International Technical Support Organization Austin Center Laurent Vanel is an AIX and RS 6000 specialist at the I
44. A LP PP1 PV1 PP2 PV2 PP3 PV3 0001 0216 hdisk6 0002 0217 hdisk6 0003 0218 hdisk6 0004 0219 hdisk6 0005 0220 hdisk6 0006 0221 hdisk6 0007 0222 hdisk6 0008 0223 hdisk6 0009 0224 hdisk6 0010 0225 hdisk6 p If we have a record of the precise commands used and the order in which they were run since the volume group was first created we may be able to re create the so called ghost maps of the logical volumes In practice this generally means having complete copies of the smit script or smit log from the time of volume group creation to the time the corruption occurs This is not an easy procedure and there can be no guarantees of success here 162 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Also note that the behavior of the allocp algorithm changed with AIX 4 2 and its behavior is not guaranteed to be identical across different levels or PTF of LVM In a more complex example we now create a new volume group allocpvg using SMIT containing hdisk6 4296 MB and hdisk7 4296 MB Since these disks are large we use a PP size of 8 We then use SMIT to create allocplv1 default allocation policy and so on containing 10 logical partitions We go on to create allocplv2 with 15 logical partitions again using all the SMIT specified defaults Now we create allocplv3 five partitions in size and with maximum inter policy set Finally we remove allocplv1 and create allocplv4 standar
45. A to Z Troubleshooting and Commands do not already exist you must use the a flag to cause the specified directory to be created D 5 The crfs command The following summarizes the options for the crfs command crfs Adds a file system Usage crfs v VfsType g VolumeGroup d Device LogPartitions m MountPoint n NodeName uMountGroup A yes no p ro rw a Attribute Value t yes no a Attribute Value Specifies a virtual file system dependent attribute value pair To specify more than one attribute value pair provide multiple a Attribute Value parameters The following attribute value pairs are specific to the Journaled File System JFS a ag 8 16 32 64 Specifies the allocation group size in megabytes An allocation group is a grouping of inodes and disk blocks similar to BSD cylinder groups The default ag value is 8 This attribute only applies to AIX Version 4 2 or later a bf true false Specifies a large file enabled file system See Understanding Large File Enabled File Systems for more information If you do not need a large file enabled file system set this option to false this is the default Specifying bf true requires a fragment size of 4096 and compress no This attribute only applies to AIX Version 4 2 or later a compress no LZ Specifies data compression If you do not want data to be compressed set this option to no The de
46. Drive hdiskl Available 10 60 00 9 0 16 Bit IVD SCSI Disk Drive hdisk2 Defined 10 60 00 10 0 16 Bit LVD SCSI Disk Drive hdisk3 Available 30 58 00 8 0 16 Bit SCSI Disk Drive hdisk4 Available 30 58 00 9 0 16 Bit LVD SCSI Disk Drive hdisk5 Available 10 60 00 10 0 16 Bit LVD SCSI Disk Drive lspv hdisko 0004163128f3de5a rootvg hdisk1 000416314bd724bc testvg hdisk3 000416314bd749f8 testvg hdisk4 000416314bdada38 testvg hdisk5 0004163192b1d297 None Ne hidsk5 is not part of any volume group yet It is time to use the magic command replacepv replacepv hdisk2 hdisk5 After issuing this simple command the results are as follows Chapter 3 Replacing a drive in a mirrored configuration 197 lspv hdisko 0004163128f3de5a rootvg hdisk1 000416314bd724bc testvg hdisk3 000416314bd749f8 testvg hdisk4 000416314bdada38 testvg hdisk5 0004163192b1d297 testvg lslv m testlv testlv home mirror LP EPL PV1 PP2 PV2 PP3 PV3 0001 0056 hdisk3 0028 hdisk5 0002 0028 hdisk1 0028 hdisk4 0003 0057 hdisk3 0029 hdisk5 0004 0029 hdisk1 0029 hdisk4 0005 0058 hdisk3 0030 hdisk5 0006 0030 hdisk1 0030 hdisk4 0007 0059 hdisk3 0031 hdisk5 0008 0031 hdisk1 0031 hdisk4 0009 0060 hdisk3 0032 hdisk5 0010 0032 hdisk1 0032 hdisk4 0011 0061 hdisk3 0033 hdisk5 0012 0033 hdisk1 0033 hdisk4 0013 0062 hdisk3 0034 hdisk5 0014 0034 hdisk1 0034 hdisk4 0015 0063 hdisk3 0035 hdisk5 0016 0035 hdisk1 0035 hdisk4 0017 0064 hdisk3 0036 hdi
47. File System NFS files If a remote file is specified the fileplace command returns an error message However the placement of the remote file can be displayed if the fileplace command is run directly on the file server 2 The fileplace command reads the file s list of blocks directly from the logical volume on disk If the file is newly created extended or truncated the file system information may not yet be on the disk when the fileplace command is run Use the sync command to flush the file information to the logical volume D 13 The fsck command The following summarizes the options for the fsck command fsck Checks file system consistency and interactively repairs the file system Usage fsck n p y dBlockNumber f ii NodeNumber o Options tFile V VFSName FileSystem1 FileSystem s dBlockNumber Searches for references to a specified disk block Whenever the fsck command encounters a file that contains a specified block it displays the i node number and all path names that refer to it 348 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Performs a fast check Under normal circumstances the only file systems likely to be affected by halting the system without shutting down properly are those that are mounted when the system stops The flag prompts the fsck command not to check file systems that were unmounted successfully The fsck command d
48. LANDesk MMX Pentium and ProShare are trademarks of Intel Corporation in the United States and or other countries UNIX is a registered trademark in the United States and other countries licensed exclusively through The Open Group SET SET Secure Electronic Transaction and the SET Logo are trademarks owned by SET Secure Electronic Transaction LLC Other company product and service names may be trademarks or service marks of others Appendix F Special notices 403 404 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Appendix G Related publications The publications listed in this section are considered particularly suitable fora more detailed discussion of the topics covered in this redbook G 1 IBM Redbooks For information on ordering these publications see How to get IBM Redbooks on page 407 e AIX Logical Volume Manager from A to Z Introduction and Concepts G24 5432 e Understanding IBM RS 6000 Performance and Sizing SG24 4810 e RS 6000 SP System Management Guide SG24 5628 GPFS A Parallel File System SG24 5165 e Getting Started with ADSM A Practical Implementation Guide SG24 5416 e Database Performance on AIX in DB2 UDB and Oracle Environments SG24 5511 G 2 IBM Redbooks collections Redbooks are also available on the following CD ROMs Click the CD ROMs button at http www redbooks ibm com for information about all the CD ROMs offered updates and formats CD ROM
49. LVM commands 217 Restores the ownership group ID and permissions of the logical volume special device files These values will be restored only if they were set using the u c and r flags of mklv and chlv commands This flag is applicable only for big VG format volume groups only V MajorNumber Specifies the major number of the imported volume group When used with the c flag sets the Concurrent Capable volume group to be autovaried on in concurrent mode When used without the c flag it does nothing Only use the c flag with HACMP It has no effect on volume groups and systems not using the HACMP product This flag only applies to AIX Version 4 2 or later In order for this auto varyon into concurrency of the volume group to take effect you must enter the following line into the etc inittab file re_clvmv 2 wait usr sbin clvm_cfg 2 gt amp 1 Attention This entry must be added after the entry used to initiate srcmstr The volume group name can only contain the following characters A through Z a through z O through 9 or _ the underscore the minus sign or the period All other characters are considered invalid Commands called getlvodm getvgname redefinevg varyonvg synclvodm and varyoffvg 218 You may import an AIX Version 3 2 created volume group into an AIX Version 4 system and you may import an AIX Version 4 volume group into an AIX Version 3 2 system provided striping
50. LVM problem is suspected the machine can be booted into maintenance mood and normal problem determination techniques as described throughout this chapter can be pursued The procedure for booting a machine into maintenance is found in the relevant hardware documentation for that model type However if file systems are not mounted when a maintenance shell is started some LVM commands will not be available unless the dev hd2 logical volume is mounted under mnt and the required commands copied into usr sbin This boot into maintenance is also required if a task that cannot be performed on an open logical volume is desired and that logical volume is a system logical volume in rootvg For instance changing the MWCC value of a logical volume is not possible while that logical volume is mounted Once the system is booted into maintenance mode lchangelv is not available in the maintenance mode dev ramo file system dev hd2 must be mounted and the 1changelv binary copied over to usr sbin We can obtain the appropriate LVID by running lqueryvg this is in dev ramo anyway and then run 1changelv 1 lt LVID gt w lt values gt In this command value is 1 to turn MWCC on or 2 to turn MWCC off 2 8 2 Reducing the size of hd6 A common problem with rootvg is how to reduce the size of the paging space of hd6 The following procedure is also interesting for its manipulation of the Chapter 2 Problem determination and recovery 187
51. Logical volume TYPE POSITION on physical volume RANGE of physical volumes MAXIMUM NUMBER of PHYSICAL VOLUMES to use for allocation Number of COPIES of each logical partition Mirror Write Consistency Allocate each logical partition copy on a SEPARATE physical volume RELOCATE the logical volume during reorganization Logical volume LABEL MAXIMUM NUMBER of LOGICAL PARTITIONS Enable BAD BLOCK relocation SCHEDULING POLICY for reading writing logical partition copies Enable WRITE VERIFY File containing ALLOCATION MAP Stripe Size Add a Logical Volume Press Enter AFTER making all desired changes Entry Fields asgard vg middle minimum yes yes yes O 512 yes parallel no Not Striped BOTTOM F1l Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image C F9 Shell F10 Fxit Enter Do 1 2 2 Copy a logical volume to a new logical volume 26 A new logical volume will be created with the same attributes as the source logical volume The user can chose to define the name for the new logical volume or let the system assign the name with the user just specifying the prefix AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note If you are copying a striped logical volume and the destination logical volume does not exist an identical copy including the striped block size and striping width of the source logical volume
52. PVID corruption on page 75 The aim is to acquaint the reader with some stylistic features of the LVM scripts which will aid in problem determination In this example we have a simple volume group victimvg made up of one physical volume hdisk10 containing logical volume victimlv on which is created a file system named victimfs The volume group is varied off and the exportvg will succeed Chapter 2 Problem determination and recovery 109 110 To get inside the functions we use trclvm The following section may be clearer if the script itself is viewed at the same time with an editor or pager so that it can be searched to follow the flow of execution The small awk script in Appendix E 2 dspmsg_index on page 395 can be used to display the text of a high level command with the 1vmmsg lines expanded to show English text Some people may find this makes the scripts more readable Va nome dug trclvm t exportvg usr bin ksh IBM PROLOG BEGIN TAG This is an automatically generated prolog introductory comments are parsed by the shell These are omitted here hash functions into system table hash getlvodm lqueryvgs putlvodm lvrelminor lvrelmajor dgpmsg ez alias t getlvodm lqueryvgs putlvodm lvrelminor lvrelmajor dspmsg D The above code resolves and saves the paths of the low level commands the script calls so the shell doesn t have to resolve the path of each command every time it
53. PagingSpace parameter specifies the paging space whose characteristics are to be shown Appendix D Other related commands 359 For NFS paging spaces the physical volume name and volume group name will be replaced by the host name of the NFS server and the path name of the file that is used for paging D 22 The mkfs command The following summarizes the options for the mkfs command mkfs Makes a file system Usage mkfs b Boot Label i i Nodes o Options p Prototype s Size v VolumeLabel V VFSName b Boot Names the program to be installed in block 0 of the new file system i i Nodes Specifies the initial number of i nodes on the file system This flag is ignored when creating a journaled file system Label Specifies the file system label for the new file system o Options Specifies a comma separated list of virtual file system implementation specific options o ag 8116132164 Specifies the allocation group size in megabytes An allocation group is a grouping of inodes and disk blocks similar to BSD cylinder groups The default ag value is 8 This option only applies to AIX Version 4 2 or later o bf true false Specifies a large file enabled file system If you do not need a large file enabled file system set this option to false this is the default Specifying bf true requires a fragment size of 4096 and compress no This option only applies to AIX Versi
54. The Ilqueryvg command The following summarizes the options for the 1queryvg command 302 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Iqueryvg An object file command that queries the attributes of a volume group Usage Iqueryvg g VGID p PVname N s F n c D a 1 P A v t g VGID The VGID of the volume group to query p PVname The name of the physical volume that contains the VGID N Returns the maximum number of logical volumes allowed in the volume group S Returns the physical partition size F Returns the number of free physical partitions in the volume group n Returns the current number of logical volumes in the volume group C Returns the current number of physical volumes in the volume group D Returns the total number of VGDAs for the volume group a Returns all the static attributes for the volume group l Returns each LVID logical volume name and state for each logical volume in the volume group P Returns each PVID number of VGDAs and state for each physical volume in the volume group A Returns all attributes for the volume group static attributes logical volume details and physical volume details V Used only with the p PvName flag to return the VGID from that physical volume Includes the tags with the output for the above options B 23 The Iqueryvgs command
55. The cpio o command The cpio o command reads file path names from standard input and copies these files to standard output along with path names and status information Avoid giving the cpio command path names made up of many uniquely linked files as it may not have enough memory to keep track of them and would lose linking information The cpio i command The cpio i command reads from standard input an archive file created by the cpio o command and copies from it the files with names that match the Pattern parameter These files are copied into the current directory tree You can list more than one Pattern parameter by using the file name notation described in the ksh command Note that in this application the special characters asterisk question mark and brackets and ellipses match the slash in path names in addition to their use as described in the ksh command The default for the Pattern parameter is an asterisk selecting all files in the Input In an expression such as a z the minus sign means through according to the current collating sequence A collating sequence can define equivalence classes for use in character ranges The cpio p command The cpio p command reads file path names from standard input and copies these files into the directory named by the Directory parameter The specified directory must already exist If these path names include directory names that AIX Logical Volume Manager from
56. This command will continue retries until lock is free If lock is inadvertent and needs to be removed execute chvg u mirrvg CH mklv y locklv mirrvg 1 0516 1201 putlvodm Warning Volume group mirrvg is locked This command will continue retries until lock is free If lock is inadvertent and needs to be removed execute chvg u mirrvg lt 2 y This lock appears in the ODM as a stanza similar to the following 106 AIX Logical Volume Manager from A to Z Troubleshooting and Commands CuAt name mirrvg attribute lock value 14428 type pR generic rep J nls _ index 0 The value field holds the process identifier of the process that owns the lock The lock should be broken with chvg u lt vgname gt A couple of other techniques exist to break the lock putlvodm k getlvodm v lt vgname gt and odmdelete o CuAt q name lt vgname gt and attribute lock However it is recommended that high level commands are used whenever possible varyonvg will also break an ODM lock however if the etc vg relevant map file is in use the vary on will fail varyonvg lostmirrvg 0516 004 varyonvg The mapped file is currently being used by another process The ODM lock should not be manually destroyed unless it is a stale lock left over by a process that has died If it is removed merely to run multiple commands simultaneously LVM corruption is likely to occur 2 6 7 The low level
57. Title Collection Kit Number System 390 Redbooks Collection SK2T 2177 Networking and Systems Management Redbooks Collection SK2T 6022 Transaction Processing and Data Management Redbooks Collection SK2T 8038 Lotus Redbooks Collection SK2T 8039 Tivoli Redbooks Collection SK2T 8044 AS 400 Redbooks Collection SK2T 2849 Netfinity Hardware and Software Redbooks Collection SK2T 8046 RS 6000 Redbooks Collection BkMgr SK2T 8040 RS 6000 Redbooks Collection PDF Format SK2T 8043 Application Development Redbooks Collection SK2T 8037 IBM Enterprise Storage and Systems Management Solutions SK3T 3694 Copyright IBM Corp 2000 405 G 3 Other resources These publications are also relevant as further information sources AIX Version 4 3 Technical Reference Base Operating System and Extensions Volume 2 SC23 4160 AIX Versions 4 1 and 4 2 Technical Reference Volume 2 Base Operating System and Extensions SC23 2615 AIX Versions 3 2 and 4 Performance and Tuning Guide SC23 2365 AIX Version 4 3 Technical Reference Kernel and Subsystems Technical Reference Volume 1 SC23 4163 AIX Version 4 3 Problem Solving Guide and Reference SC23 4123 HACMP Version 4 3 AIX Installation Guide SC23 4278 PSSP for AIX Version 3 1 Managing Shared Disks SA22 7349 HACMP Version 4 3 AIX Programming Locking Applications SC23 4281 OEM PCI Adapter Placement Guide SA23 2504 IBM Recoverable Virtual Shared Disk User s Guide and Reference GC23 3849 AIX V
58. Volume Manager from A to Z Troubleshooting and Commands The 0516 304 error message seen previously can also occur if we have a partial mirror or a false mirror This is discussed next Sixteen zeros errors and other corrupt PVIDs in the VGDA As we saw in the VGDA corruption with two VGDAs on one disk on page 156 a corrupted non mirrored logical volume may give 0516 304 errors in response to inspection commands In the partition map corruption shown previously is not the only reason this error number may be reported As the error message states the LVM is looking in the ODM for a PVID having the nonsense value 0000000000000000 This message will be returned any time there is a 0000000000000000 value taken from the VGDA and referenced against the ODM Such a value can occur due either to corruption within the logical volume partition map or if a PVID of 0000000000000000 is present as a physical volume definition at the start of a partition map You can check for a PVID of 0000000000000000 being present in the partition map by running lqueryvg Ptp In contrast lquerpv h dev hdisk 80 10 will give the real PVID on the hdisk This is shown below lsvg M corrvg corrvg 0516 304 lsvg Unable to find device id 0000000000000000 in the Device Configuration Database 0000000000000000 1 215 0000000000000000 216 corrlv 1 0000000000000000 217 corrlv 2 0000000000000000 218 corrlv 3 0000000000000000 219 1075 lqueryvg Ptp
59. We simulate the failing drive by removing hdisk2 from the system and reboot This screen shows the resulting state on the system As you can see the status of the disk is now defined but it is always present in the logical volume map A lspv hdisko 0004163128f3de5a rootvg hdisk1 000416314bd724bec testvg hdisk3 000416314bd749 8 testvg hdisk4 000416314bdada38 testvg lsdev Ccdisk hdiskO Available 10 60 00 8 0 16 Bit SCSI Disk Drive hdisk1 Available 10 60 00 9 0 16 Bit IVD SCSI Disk Drive hdisk2 Defined 10 60 00 10 0 16 Bit LVD SCSI Disk Drive hdisk3 Available 30 58 00 8 0 16 Bit SCSI Disk Drive hdisk4 Available 30 58 00 9 0 16 Bit IVD SCSI Disk Drive lslv m testlv testlv home mirror LP PP1 PV1 PP2 PV2 PP3 PV3 0001 0056 hdisk3 0028 hdisk2 0002 0028 hdisk1 0028 hdisk4 0003 0057 hdisk3 0029 hdisk2 0004 0029 hdisk1 0029 hdisk4 0005 0058 hdisk3 0030 hdisk2 0006 0030 hdisk1 0030 hdisk4 lt lt lines removed gt gt 0017 0064 hdisk3 0036 hdisk2 0018 0036 hdisk1 0036 hdisk4 0019 0065 hdisk3 0037 hdisk2 XN S We add a new disk to the system We even try to confuse the system by inserting the new disk at the same position than the previous one but that didn t trick the system at all The new disk is named hdisk5 and the status of hdisk2 is always defined 196 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Va lsdev Ccdisk hdiskO Available 10 60 00 8 0 16 Bit SCSI Disk
60. a multiple of the physical block size for that tape device When using the B or C options to extract an archive from tape the blocking factor should not be larger than the size of the archive as it exists on the tape The B flag and the c flag are mutually exclusive If you list both the cpio command uses the last one it encounters in the flag list Reads and writes header information in ASCII character form If a cpio archive was created using the c flag it must be extracted with the c flag C Value Performs block input and output using the Value parameter times 512 bytes to a record For instance a c2 flag changes the block input and output sizes to 1024 bytes toa record Creates directories as needed Copies all files except those matching the Pattern parameter 332 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Links files rather than copying them whenever possible This flag can only be used with the cpio p command Retains previous file modification time This flag does not work when copying directories Retains previous file modification time even when directories are copied Renames files interactively If you do not want to change the file name enter a single period or press the Enter key In the latter case the cpio command does not copy the file Swaps bytes This flag is used only with the cpio i command Note If there is an odd number of b
61. a minimal amount of information in the ODM we already said LVM was confused Chapter 2 Problem determination and recovery 131 synclvodm targetvg 0516 306 synclvodm Unable to find volume group targetvg in the Device Configuration Database 0516 502 synclvodm Unable to access volume group targetvg Performing problem determination logically we check the ODM The short script in Appendix E 5 findlvm on page 397 checks the ODM classes for a particular string Running findlvm target at this point returns nothing We know that no references to our volume group or its dependent logical volumes exist in the LVM To regain this basic ODM information we need to run redefinevg We can pass redefinevg the volume group name which of course we already know and either a VGID or a physical volume name It is a good idea to confirm what volume group we are running the redefinevg on beforehand We can do this by looking at the key VGID for the volume group with lqueryvg lqueryvg g 0001743770963601 At Max LVs PP Size Free PPs LV count PV count Total VGDAs Conc Allowed MAX PPs per MAX PVs Conc Autovar Varied on Co 256 22 1072 Logical 00017d3770963601 1 targetlog 1 00017d3770963601 2 targetlv 1 Physical 00017d375243d402 2 1 Total PPs 1075 Le We may now run the redefinevg i redefinevg i 00017d375243d402 targetvg Alternatively in a telephone support situati
62. an attempt to fix the problem is made If either of these are unsuccessful more investigation is performed or another hypothesis is drawn Once a cause is found it can then be repaired or rebuilt It is vital that the reader realize that merely fixing a problem for example the object database manager ODM corruption only addresses a symptom The LVM may be back up but the reason for its failure should also be considered or it may simply reoccur Problem determination is not complete until the cause has been determined as far as is possible bearing in mind there are often practical limits to how much problem determination may be performed after repairs have been made and a system is back online As an introduction to looking inside the LVM consider that most high level LVM commands exist as user readable korn shell scripts This allows the troubleshooter to identify each command s affect on the LVM structures of the volume group descriptor area VGDA the volume group status area VGSA and so on as well as on the AIX s ODM database 2 1 1 Breaking up high level commands 60 For example the mkvg command may be broken down an initial simplification as shown in Figure 6 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Parsing a Checking a lcreatevg a extendvg Figure 6 Flow of execution for mkvg Here we can see that mkvg
63. applies to AIX Version 4 2 or later Specifies to automatically expand the tmp file system if necessary This flag only applies to AIX Version 4 2 or later Commands called bosboot mkinstape mkszfile and backup m Note The savevg command will not generate a bootable tape if the volume group is the root volume group Although the tape is not bootable the first three images on the tape are dummy replacements for the images normally found on a bootable tape The actual system backup is the fourth image For more information about images on bootable tapes please see Troubleshooting an Installation from a System Backup General Information Regarding mksysb System Backups in the AIX Version 4 3 Installation Guide SC23 4112 The savevg command finds and backs up all files belonging to a specified volume group A volume group must be varied on and the file systems must be mounted The savevg command uses the data file created by the mkvgdata command This file can be one of the following image data Contains information about the root volume group rootvg The savevg command uses this file to create a backup image that can be used by Network Installation Management NIM Appendix A High level LVM commands 273 to reinstall the volume group to the current system or to a new system tmp vgdata vgname vgname data Contains information about a user volume group The vgname variable reflects th
64. be allocated to share or not to share the same physical volume The Strict parameter is represented by one of the following y Sets a strict allocation policy so that copies for a logical partition cannot share the same physical volume This is the default for allocation policy Does not set a strict allocation policy so that copies for a logical partition can share the same physical volume Sets a super strict allocation policy so that the partitions allocated for one mirror cannot share a physical volume with the partitions from another mirror S StripeSize Specifies the number of bytes per striped Must be a power of two between 4 K and 128 K for example 4 K 8 K 16 K 32 K 64 K or 128 K Note The d e m and s flags are not valid when creating a striped logical volume using the s flag 246 AIX Logical Volume Manager from A to Z Troubleshooting and Commands t Type Sets the logical volume type The standard types are jfs file systems jfslog journal file system logs and paging paging spaces but a user can define other logical volume types with this flag You cannot create a striped logical volume of type boot The default is jfs If a log is manually created for a file system the user must run the logform command to clean out the new jfslog before the log can be used For example to format the logical volume logdev enter logform dev logdev where dev logdev is the absolute path t
65. big VG structures in some recovery situations The reader should be advised that in some situations recovery is simply not possible or not possible as a customer procedure User data may either be gone one may lack sufficient control information to perform reconstruction or one may be in a deadlock situation where all approaches to repair are damaged It must also be stressed that once recovery is started if a recovery attempt fails the LVM control structures may be left in a worse state than before Reasons for this are shown in this chapter It is recommended therefore that as much control information as possible should be backed up beforehand At the least the ODM should be backed up and possibly backup copies of the VGDAs should be made using dd as well Note This chapter assumes the reader is comfortable with the LVM concepts Competency in korn shell scripting is required to make full use of the material in this chapter although it is not a pre requisite It would also be helpful for the reader to be familiar with the format of C language header files Copyright IBM Corp 2000 59 2 1 A methodology for problem determination Identifying the source of a problem in LVM or otherwise is often a circular process Initial investigation is performed An hypothesis is made based on the facts discovered Action is then taken to test the validity of the hypothesis Further data to test the theory is gathered or
66. command that adds an object to a created object class Usage odmadd InputFile s The odmadd command takes as input one or more InputFile files and adds objects to object classes with data found in the stanza files Each InputFile file is an ASCII file containing the data that describes the objects to be added to object classes If no file is specified input is taken from stdin standard input The classes to be added to are specified in the ASCII input file The file is in the following general format class1name descriptor1name descriptor1 value descriptor2name descriptor2value descriptor3name descriptor3value class2name descriptor4name descriptor4value descriptor5name descriptor5value C 2 The odmchange command The following summarizes the options for the odmchange command Copyright IBM Corp 2000 315 odmchange An object file command that changes the contents of a selected object in the specified object class Usage odmchange o ObjectClass q Criteria InputFile o ObjectClass Specifies the object class to modify q Criteria Specifies the criteria used to select objects from the object class Pattern matching can be done using comparison operators for example or pattern matches for example like More information on qualifying criteria see Understanding ODM Object Searches in AIX Version 4 3 General Programming Concepts Writing and Debugging Programs SC
67. command that is used to generate an allocation map that is required when a logical volume is created extended reduced or removed s Size t u VolumeGroup Usage allocp a Position c Copies e Range i LVID K k Strict S StripSize a Position Sets the intraphysical volume allocation policy the position of the logical partitions on the physical volume The Position variable is one of the following m Allocates logical partitions in the outer middle section of each physical volume This is the default position c Allocates logical partitions in the center section of each physical volume e Allocates logical partitions in the outer edge section of each physical volume ie Allocates logical partitions in the inner edge section of each physical volume im Allocates logical partitions in the inner middle section of each physical volume c Copies 1 2 0r3 Copyright IBM Corp 2000 285 e Range Sets the interphysical volume allocation policy the number of physical volumes to extend across using the volumes that provide the best allocation The value of the Range variable is limited by the Upperbound variable set with the u flag and is one of the following x Allocates logical partitions across the maximum number of physical volumes m Allocates logical partitions across the minimum number of physical volumes i LVid K 2 k Strict Determines the st
68. commands Strictly speaking these commands are not intended for customer use As already stated significant error checking is provided above this layer in the high level scripts and the high level commands also have the intelligence to drive the low level commands in such a way as to avoid their occasional peculiarities That said the low level commands provide a useful way to target with certainty a specific VADA or ODM object when it cannot be established where high level commands are obtaining their information from They can often provide more information than the corresponding high level commands as well They can also be used to replicate the actions of a high level script and therefore walk through it until a point of failure with opportunities to check the state of the LVM while problem re creation is taking place This must be done with caution Chapter 2 Problem determination and recovery 107 Focussing on the low level inquiry commands we will discuss getlvcb getlvodm lquerylv lquerypv and lqueryvg egetlvcb This command allows checking for LVCB corruption We have seen it used in Corruption example 2 PVID corruption on page 75 where it was noted that corrupt LVCBs may cause loss of mount points and log device information for file systems The implications of a corrupt LVCB are discussed further in LVCBs on page 178 eget lvodm This command allows target ing the ODM to find lists of free physical v
69. copy to place the physical partitions belonging to a copy as contiguously as possible and then to place the physical partitions in the desired region specified by the a flag Also by default each copy of a logical partition is placed on a separate physical volume m Note 1 When extending a striped logical volume the number of partitions must be in an even multiple of the striping width 2 When extending a striped logical volume only the striping width disks striped across is used If there is not enough partitions on the physical volumes used for this striped logical volume the extend of the logical volume fails 3 It is recommended that a logical volume using a large number of partitions more than 800 MB be extended gradually in sections 4 Changes made to the logical volume are not reflected in the file systems To change file system characteristics use the chfs command A 7 The extendvg command The following summarizes the options for the extendvg command extendvg Adds physical volumes to a volume group Appendix A High level LVM commands 215 Usage extendvg f VolumeGroup PhysicalVolume s f Forces the physical volume to be added to the specified volume group unless it is a member of another volume group in the Device Configuration Database or of a volume group that is active Commands called getlvodm linstallpv and putlvodm The physical volume is checked to
70. create a boot logical volume of the same name as many parts of AIX have this hardcoded We can simply run rmlv hd6 followed by mklv y hd6 t paging rootvg lt desired size in LPs gt We should now change the swapon line in sbin rc boot back to swapon dev hde We should then check if hd6 will be automatically swapped on at boot time with 1sps a If there is a nin the auto column we must run chps a y hd6 Next we rerun the bosboot If we created a temporary paging space it should now be turned off with chps n and rebooted It can then be deleted with miv The final piece of tidying up is to reset the dump device pointer with sysdumpdev P p dev hd6 assuming we altered the primary dump device 188 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Chapter 3 Replacing a drive in a mirrored configuration This chapter describes how to replace a failing hard drive in a mirrored configuration Mirroring is used to increase the data availability but you should always have a backup ready just in case something really wrong happens In a non mirrored configuration if a hard drive dies your only solution is pretty much the backup In a mirrored configuration this failure should be transparent to users and you should be able to replace the disk without loosing any data Based on your version of AIX this procedure can be quite simple with AIX Version 4 3 3 and the replacepv command or more complex with a sequence
71. create the boot images for all three platforms This type of backup also requires a user supplied and previously created mksysb image The non bootable volume group backup contains only the CD image of a volume group If this backup contains rootvg backup then you must boot from a product CD before restoring the mksysb image or use alt_disk install to install it If the backup volume group is a non rootvg volume group then use restvg to restore the image m Note The functionality required to create Rock Ridge format CD images and to write the CD image to the CD R device is not part of the mkcd command You must supply additional code to mkcd to do these tasks The code will be called via shell scripts and then linked to usr sbin mkrr_fs for creating the Rock Ridge format image and usr sbin burn_cd for writing to the CD R device Both links are called from the mkcd command Some sample shell scripts are included for different vendor specific routines You can find these scripts in usr samples oem_cdwriters If you do not give any file systems or directories as command parameters mkcd creates the necessary file systems and removes them when the command finishes executing File systems you supply are checked for adequate space and write access If mkcd creates file systems in the backup volume group they are excluded from the backup If you need to create multi volume CDs because the volume group image does not fit
72. error checks The fsdb command enables you to examine alter and debug a file system specified by the FileSystem parameter The command provides access to file system objects such as blocks i nodes or directories You can use the fsdb command to examine and patch damaged file systems Key components of a file system can be referenced symbolically This feature simplifies the procedures for correcting control block entries and for descending the file system tree To examine a file system specify it by a block device name a raw device name or a mounted file system name In the last case the fsdb command determines the associated file system name by reading the etc filesystems file Mounted file systems cannot be modified The subcommands for the fsbd command allow you to access view or change the information in a file system Any number you enter in the sub command is considered decimal by default unless you prefix it with either o to indicate an octal number or ox to indicate a hexadecimal number All addresses are printed in hexadecimal Because the fsdb command reads and writes one block at a time it works with raw as well as block I O D 15 The imfs command The following summarizes the options for the imfs command 354 AIX Logical Volume Manager from A to Z Troubleshooting and Commands imfs Removes file system data from etc filesystems Usage imfs x PVName VGName I LVName X R
73. f Device s SeekBackup File s restore To list files archived by file system Usage t T B h q v y b Number f Device s SeekBackup File s B Specifies that the archive should be read from standard input Normally the restore command examines the actual medium to determine the backup format When using a pipe this examination cannot occur As a result the archive is assumed to be in file system format and the device is assumed to be standard input f 366 AIX Logical Volume Manager from A to Z Troubleshooting and Commands b Number For backups done by name specifies the number of 512 byte blocks for backups done by i node specifies the number of 1024 byte blocks to read in a single output operation When the restore command reads from tape devices the default is 100 for backups by name and 32 for backups by i node The read size is the number of blocks multiplied by the block size The default read size for the restore command reading from tape devices is 51200 100 512 for backups by name and 32768 32 1024 for backups by i node The read size must be an even multiple of the tape s physical block size If the read size is not an even multiple of the tape s physical block size and it is in fixed block mode nonzero the restore command tries to determine a valid value for Number If successful the restore command changes Number to the new value writes a message abo
74. f dudlv or smit rmlv D Remove a Logical Volume Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields LOGICAL VOLUME name dud1lv F1l Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image _ F9 Shell1 F1O Exit Enter Do p Repeat step 2 for each non mirrored logical volume 190 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 3 1 3 Step 3 We can remove the failed disk from the volume group reducevg asgard_vg hdiskn or refer to 1 1 4 Remove a physical volume from a volume group on page 8 3 1 4 Step 4 Now remove the definition of the failed physical volume from the ODM rmdev l hdiskn d or refer to smit rmvdsk1 ra N Remove a Disk Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields DISK Name hdiskn KEEP definition in database no F1l Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image C F9 Shell F10 Exit Enter Do J 3 1 5 Step 5 Shut down the system remove the failed physical volume and add the new disk in the same location After the physical volume has been replaced reboot the machine If the failed physical volume is hot swappable the system doesn t need to be shut down If the failed physical volume is part of a SSA system then there are special commands to handle the SSA disk If the
75. first parses its command line arguments It then passes these on to the low level command 1createvg to build a VGDA containing the first physical volume It then invokes the extendvg command to bring additional physical volumes into the volume group once per physical volume It is apparent from inspection of the script that a number of other actions are also performed These include e Checking if the target physical volume devices are available or if they already participate in pre existing volume groups Generation of major and minor numbers for the volume group s entry in the dev directory and the ODM s CuDvDr Chapter 2 Problem determination and recovery 61 Z SVFLAG then Get major number MAJOR lvgenmajor VGNAME test_return GOTMAJOR 1 else Check major number MAJOR lvchkmajor VVAL VGNAME test_return GOTMAJOR 1 fi Get minor number should be zero MINOR lvgenminor MAJOR VGNAME test_retur GOTMINOR 1 Ne S e Updating the ODM to contain information for the new volume group Add new volume group to odm putlvodm CFLAG C VAL XFLAG X VAL v VGNAME o AUTO ON q 0 VGID if 0 If putlvodm failed output warning and continue then dspmsg s 1 cmdlvm cat 1024 lvmmsg 1024 n mkvg gt amp 2 fi It is not the intention here to set out every action performed by the mkvg script Rather the reader should be aware that it is
76. group Mount automatically at system restart Permissions Mount Options Start Disk Accounting Change the automount option Set permissions read write or read only Set the mount options nosuid or nodev Turn on or off accounting for this file system m Note The other fields cannot be changed and are just for information smit chjfs C e N Change Show Characteristics of a Journaled File System Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields File system name usr NEW mount point usr SIZE of file system in 512 byte blocks 933888 Mount GROUP bootfs Mount AUTOMATICALLY at system restart yes PERMISSIONS read write Mount OPTIONS Start Disk Accounting no Fragment Size bytes 4096 Number of bytes per inode 4096 Compression algorithm no Large File Enabled false Allocation Group Size MBytes 8 F1l Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image F9 Shel1 F10 Exit Enter Do Ne A Chapter 1 LVM commands 53 1 4 3 Defrag a file system 54 The amount of contiguous free space in a file system can be increased by reorganizing allocations to be contiguous rather than scattered across the disk For the command line options see defragfs in Appendix D 6 The defragfs command on page 338 or use the SMIT fastpath smit dejfs SMIT provides the following fields File system name Mount point of the file
77. have NBPI values from 512 to 128 K with corresponding maximum file system sizes D 3 The chps command The following summarizes the options for the chps command chps Changes attributes of a paging space Usage chps s LogicalPartitions a y n PagingSpace 330 AIX Logical Volume Manager from A to Z Troubleshooting and Commands a Specifies to use a paging space at the next system restart y Specifies that the paging space is active at subsequent system restarts n Specifies that the paging space is inactive at subsequent system restarts s LogicalPartitions Specifies the number of logical partitions to add The chps command changes attributes of a specific paging space The PagingSpace parameter specifies the name of the paging space to be changed If the paging space is increased in size the additional partitions will be available for use immediately even though they may not be used straight away To change the size of a Network File System NFS paging space the size of the file that resides on the server must first be changed and then the swapon command used to notify the client of the change in size of the paging space Note The primary paging space is hardcoded in the boot record Therefore the primary paging space will always be activated when the system is restarted The chps command is unable to deactivate the primary paging space D 4 T
78. hdisk6 Physical 0000000000000000 2 0 lquerypv h dev hdisk6 80 10 00000080 00017D37 8BB9F828 00000000 00000000 ou 7 A EAEE D If the 16 zeros are actually present in the VGDA but no partitions are mapped in they can be removed by specifying p 0000000000000000 to 1deletepv at all levels of AIX The alternative is to try reducevg which allows you to specify a pvid specifically to handle this kind of corruption situation 16 zeroes may be seen as a special case of any nonsensical PVID value existing on a disk Note that if any copies of logical volumes are in partition maps associated with this PVID and a reducevg Or 1deletepv is run those maps will be lost You would probably want to map the volume group export it and remake it in such a situation Chapter 2 Problem determination and recovery 159 Apart from the dd corruption shown in the last section this could also be caused by a false or partial mirror at AIX 4 2 1 This is shown as follows N 421 lslv m corrlv corrlv N A LP PP1 PV1 PP2 PV2 PP3 PV3 0001 0155 hdisk1 0490 hdisko 0002 0156 hdisk1 0003 0157 hdisk1 421 lsvg l rootvg rootvg LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT hd6 paging 32 32 i open syncd N A hd5 boot 1 1 1 closed syned N A hd8 jfslog 1 1 1 open syncd N A hd4 jfs 16 16 1 open syncd hd2 jfs 424 424 1 open syncd usr hd9var jfs 7 7 1 open syncd var hd3 jfs 44 44 1 open syncd tmp hd1 jfs 37
79. head 1 allocplv3 data dd skip 1 if dev allocplv4 head 1 allocplv4 data In some situations it may be desirable to examine the internal tables allocp uses to generate its maps This may be done by exporting the environment variable DBGCMD_ALLOCP 1 This will produce debug output in the file tmp DEBUGCMD Scraping If we cannot resurrect the ghost maps an alternative is to scrape the surface of the disks od cx dev hdiskn looking for LVCBs with This can be used to indicate the byte offset of the first logical partition from the beginning of the disk However this is in fact a very inefficient way to do the search and doesn t really give us enough information to begin finding an LP PP map As an alternative consider that the offset for the first physical partition in sectors is contained within the VGDA For a small VGDA this will be at block 4352 For a BigVG this is located at block 17408 This value will also change for different a values specified with the mkvg used to create the volume group Without discussing where we can find this by following pointers within the VGDA structure this information is proprietary the following alternative test can be used to check the PP1 offset for a particular number of disks in a volume group Note This test should not be done on the actual disk being recovered as it will place an LVCB down on the physical volume if PP1 contains data that will be corrupted Use a
80. in by the user only by other high level commands There is a high level of similarity between the flag variable names and corresponding roles among the scripts For instance sBrLac requests big VG conversion or creation and swrtac specifies to mirror write consistency However the reader cannot rely on a flag having a constant meaning across all scripts scrLac refers variously to concurrent capability number of mirrors clear boot record and exit in a case of name conflict Returning to exportvg we have still to parse the command line arguments Chapter 2 Problem determination and recovery 113 A N Parse command line arguments set getopt getopt lt Seb 22 5 if 0 Determine if there is a syntax error then dspmsg s 1 cmdlvm cat 606 lvmmsg 606 gt n exportvg gt amp 2 dspmsg s 1 cmdlvm cat 760 lvmmsg 760 gt n SPROGNAME gt amp 2 exit fi 0 0 shift F shift J The only real parsing check to be made is if the script has been passed the name of the volume group to be exported This will be the case if the shell 1 parameter variable is non null if n 1 if vgname argument on command line then VGNAME 1 Determine the major number of the root device file ls 1 dev IPL_rootvg sed s g gt tmp rootdevices read skipfld skipfld skipfld skipfld ROOTMAJOR skipfld lt tmp rootdevicess VGID getlvodm v SVGNAME test_returm chec
81. individually For example if several scratch file systems always need to be mounted together when performing certain tests they can each be placed in the test mount group They can then all be mounted with a single command such as the mount t test command The chfs command changes the attributes of a file system The new mount point automatic mounts permissions and file system size can be set or changed The FileSystem parameter specifies the name of the file system expressed as a mount point Some file system attributes are set at the time the file system is created and cannot be changed For the Journaled File System JFS such attributes Appendix D Other related commands 329 include the fragment size block size number of bytes per i node compression and the minimum file system size Table 1 Fragment size and nbpi restrictions nbpi Fragment size in bytes Maximum size in 512 byte blocks 512 512 1024 2048 4096 16777216 1024 512 1024 2048 4096 33554432 2048 512 1024 2048 4096 67108864 8192 512 1024 2048 4096 134217728 16384 512 268435456 16384 1024 2048 4096 536870912 32768 512 268435456 32768 1024 536870912 32768 2048 4096 1073741824 65536 131072 512 268435456 65536 131072 1024 536870912 65536 131072 2048 1073741824 65536 131072 4096 2147483648 AIX Version 4 1 is limited to NBPI values from 512 to 16384 In AIX Version 4 3 you can
82. is in file system format a list of files found on the archive is written to standard output The name of each file is preceded by the i node number of the file as it exists on the archive The file names displayed are relative to the root directory of the file system that was backed up If the File parameter is not specified all the files on the archive are listed If the File parameter is used then just that file is listed If the File parameter refers to a directory all the files contained in that directory are listed If the archive is in file name format information contained in the volume header is written to standard output This flag can be used to determine if the archive is in file name or file system format Appendix D Other related commands 371 Displays additional information when restoring If the archive is in file name format and either the xor T flag is specified the size of the file as it exists on the archive is displayed in bytes Directory block or character device files are archived with a size of 0 Symbolic links are listed with the size of the symbolic link Hard links are listed with the size of the file which is how they are archived Once the archive is read a total of these sizes is displayed If the archive is in file system format directory and non directory archive members are distinguished X VolumeNumber Begins restoring from the specified volume of a multiple volume file na
83. ksh Again we skip the prologue and initial comments there is alsoa cleanup function which in addition to the steps we saw in exportvg resets the umask and runs savebase if we are in run time as qpposed to maintenance mode The main body starts in the same way exportvg did PATH usr bin etc usr sbin usr ucb usr bin X11 soin PATH PATH usr bin ete usr sbin usr ucb usr bin X11 sbin usr bin ete usr sbin usr ucb home dug bin usr bin X11 sbin home dug home dug scripts VGEXTSIED VGEXISIED this is a variable used by the cleanup routine in the same mamner as exportvg s locked variable EXIT CODE 1 EXIT COE 1 OLD UYASK uresk save old umask value umask OLD UVASK 022 umask 117 set umask for rw rw umask 117 COVDIR etc dojrepos COMDIR etc dojrepos export AMR export ADR TG OFF odiget q attrilbute 10B STATE and deflt tcb disabled PdAt Odmget q attribute 10B STATE and deflt tcb disabled PdAr TCR OFFE P t uniquetype attribute I B_SIATE deflt t disabled values width type generic rep nls index 0 Trap cn exit interrupt break to clean up trap clearup 012 15 trap cleamp 01215 PROG basename 0 Just get the basmam of the commend basena trp lvmtvacel4910 redefinevg PROG redefinevg Ne 136 AIX Logical Volu
84. mapping of logical partitions to physical partitions When a user changes a logical volume from within the editor the editor checks that the modification is permissible in the context of the current editor state If the change is allowed the editor state is modified to reflect it so that the user can preview the results from within the editor However the actual system logical and physical volumes are not altered until the user elects to commit the changes upon exiting the editor The logical volume editor permits both precise allocation of physical volumes to logical volumes within a volume group as well as the definition of logical volume attributes However this editor does not currently support the creation and modification of higher level abstractions built on these logical volumes such as file systems and paging spaces After arranging the logical volumes within a volume group a user must still create or modify file systems paging spaces and journaling logs for example This can be done from within SMIT or by using commands such as mkfs chfs swapon and chps The editor also AIX Logical Volume Manager from A to Z Troubleshooting and Commands does not support creation or extension of volume groups which must be defined prior to entering the editor Most editing operations other than extension are not permitted on logical volumes that are in active use These consist of logical volumes with mounted file systems ac
85. migration Another command that performs this temporary mirroring is reorgvg hence it may also be affected you can see this check in the get_stales function If a disk is in this state it can be converted to a working state by running chvg t followed by an appropriate factor size for the largest physical volume in the volume group Stale partition errors As the reader should know by now situations such as lsvg M stalevg D stalevg hdisk8 1 108 hdisk8 109 stalelv 1 1 hdisk8 110 stalelv 2 1 hdisk8 111 stalelv 3 1 hdisk8 112 537 hdisk9 1 108 hdisk9 109 stalelv 1 2 stale hdisk9 110 stalelv 2 2 stale hdisk9 111 stalelv 3 2 stale hdisk9 112 537 ee J where the mirror on hdisk9 of stalelv is reported as stale may not be true corruption problems All that is indicated is that no synchronization action has been performed yet for instance a mklvcopy has been run without the k flag To synchronize the partitions all that is necessary is to run a syncvg While the remedy is simple it is useful to consider what is happening under the covers when syncvg runs To this end a trclvm syncvg S run Chapter 2 Problem determination and recovery 171 trclvm 1 stalelv bin ksh IBM PROLOG BEGIN TAG This is an automatically generated prolog The reader will be pleased to know that this command turns out to be a very simple script Once again we omit the prolog and description details of extern
86. node map 352 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Orphaned files and directories those that cannot be reached are if you allow it reconnected by placing them in the lost found subdirectory in the root directory of the file system The name assigned is the i node number If you do not allow the fsck command to reattach an orphaned file it requests permission to destroy the file In addition to its messages the fsck command records the outcome of its checks and repairs through its exit value This exit value can be any sum of the following conditions O All checked file systems are now okay 2 The fsck command was interrupted before it could complete checks or repairs 4 The fsck command changed the file system the user must restart the system immediately 8 The file system contains unrepaired damage When the system is booted from a disk the boot process explicitly runs the fsck command specified with the and p flags on the usr var and tmp file systems If the fsck command is unsuccessful on any of these file systems the system does not boot Booting from removable media and performing maintenance work will then be required before such a system will boot If the fsck command successfully runs on usr var and tmp normal system initialization continues During normal system initialization the fsck command specified with the and p flags runs from the etc rc file Thi
87. node or by name format Renames restored files to the file s i node number as it exists on the archive This is useful if a few files are being restored and you want these files restored under a different file name Since any restored archive members are renamed to their i node numbers directory hierarchies and links are not preserved Directories and hard links are restored as regular files The mflag is used when the archive is in file system format Specifies that the first volume is ready to use and that the restore Command should not prompt you to mount the volume and hit Enter If the archive spans multiple volumes the restore command prompts you for the subsequent volumes Requests a specific volume of a multiple volume file system archive The rR flag allows a previously interrupted restore to be restarted The File parameter is ignored when using the r flag Once restarted the restore command behavior is the same as with the r flag Restores all files in a file system archive The r flag is only used to restore complete level 0 backups or to restore incremental backups after a level 0 backup is restored The restoresymtable file is used by restore to pass information between incremental restores This file should be removed once the last incremental backup is restored The File parameter is ignored when using the r flag 370 AIX Logical Volume Manager from A to Z Troubleshooting and Commands
88. nodes delayed block I O and read write mapped files 380 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note The writing although scheduled is not necessarily complete upon return from the sync subroutine D 30 The sysdumpdev command The following summarizes the options for the sysdumpdev command sysdumpdev Changes the primary or secondary dump device designation in a running system Usage sysdumpdev c C P p Device s Device q Usage sysdumpdev c C p Device s Device q Usage c C d Directory D Directory e k K I L p Device q r Host Path s Device z C Specifies that all future dumps will be compressed before they are written to the dump device The c flag applies to only AIX Version 4 3 2 and later versions C Specifies that dumps will not be compressed The c flag applies to only AIX Version 4 3 2 and later versions D Directory Specifies the Directory the dump is copied to at system boot If the copy fails at boot time using the p flag allows you to copy the dump to an external media Note When using the d Directory or p Directory flags the following error conditions are detected e Directory does not exist e Directory is not in the local journaled file system e Directory is not in the rootvg volume group d Directory Specifies the Directory the dump is copied to at sy
89. o o o o o o o o 0000 0000 0000 0000 0000 0000 0000 0000 Figure 15 Contents of the first physical sectors on an hdisk Block 70 and blocks 71 92 Backup copies of the LVM information record and the LVM bad block directory As can be seen in Figure 15 backup copies of the LVM information record and the bad block directory are stored on the physical volume in blocks 70 92 Since this backup exists there is a last resort method to restore the LVM information block Check the size of the LVM information block using a command of the form echo ib 16 n dd if dev hdiskn skip 7 count 1 od x grep 0000020 awk print 5 gt bc Divide this by 512 and round up to get the number of sectors You can run a command of the form dd if dev hdiskn skip 70 of dev hdiskn seek 7 count NumSectors to restore the 124 AIX Logical Volume Manager from A to Z Troubleshooting and Commands original LVM information record and reattempt the vary on Note that whenever such an unsupported technique as this is attempted the volume group may be left in an unstable state At this point we should backup user data and re creation of the volume group should be done Since we have vgda_psn 0 and vgda_psn 1 we can see our VGDAs Looking at vgda_psn 0 we have oxss or sector 136 decimal If we look at the start of this area by running dd if dev hdisk6 skip 136 count 1 2 gt dev null od cx e gt dd if dev ndisk6 skip 136 count 1 2 gt dev
90. of the backup logical volume Specifying the 1 flag also sets the persistence option allowing applications to access split mirror copy via newlvname Maintains information about the existence of an online split mirror copy across a reboot and also allows other nodes in a concurrent mode environment to be aware of the existence of the online split mirror copy Starts a background syncvg for the logical volume Allows split mirror copy to be writable default is to create the split mirror copy as READ ONLY LVName Allows split mirror copy to be writable default is to create the split mirror copy as READ ONLY 288 1 If persistence is used either by using the r flag or by creating a child backup logical volume device by using the 1 flag it will cause the volume group to be usable only on 4 3 2 0 or later versions of AIX This is true even after removal of split mirror copy designation of the parent logical volume and the child backup logical volumes 2 For chvlcopy to be successful in a concurrent volume group environment all the concurrent nodes must be at AIX Version 4 3 2 or later All partitions of a logical volume must be refreshed before chvlcopy can mark a mirror copy as a split mirror Only one copy may be designated as an online split mirror copy Although the chlvcopy command can mark online split mirror copies on logical volumes that are open including logical volumes containin
91. on hdisk6 We create lowlv2 map raw which only contains the logical partitions for hdisk6 grep 00017d377c3abace lowlv2 map raw bothdisks gt lowlv2 map raw cat lowlv2 map raw 00017d377c3abace 219 1 00017d377c3abace 220 2 We now have maps to pass into lreducelv We get the number of partitions in each map with we 1 and run the lreducelv We will see shortly that this was a bad idea at this point in the recovery procedure lreducelv 1 SVGID 1 s 3 lowlvl map raw lreducelv 1 SVGID 2 s 2 lowlv2 map raw ldeletepv g SVGID p 00017d377c3abace lspv M n hdisk7 hdisk6 0516 022 lspv Illegal parameter or structure value The 1spv M result may be unexpected We attempt to reducevg the bad VGDA on hdisk6 away reducevg lowvg hdisk6 usr sbin reducevg 227 15864 Segmentation fault coredump 0516 882 reducevg Unable to reduce volume group 150 AIX Logical Volume Manager from A to Z Troubleshooting and Commands The above procedure doesn t seem to have helped at all In fact we cannot even get the maps out any more with 1spv M n The other high level commands will also dump core Here we might try varying off varyonvg lowvg usr sbin syncvg 16 18586 Segmentation fault coredump usr sbin syncvg 18 18588 Segmentation fault coredump 0516 932 usr sbin syncvg Unable to synchronize volume group lowvg umount lowfs1 umount lowfs2 varyoffvg othervg
92. parameter with this command mlv Logical volume 1v02 is removed Alternatively we may first rmlv the logical volumes then run the reducevg cleanly rmlv 1v02 Warning all data contained on logical volume lv02 will be destroyed mlv Do you wish to continue y es n o y usr sbin rmlv 458 test 0403 004 Specify a parameter with this command mlv Logical volume 1v02 is removed reducevg workvg hdisk6 In both cases we can ignore the 0403 004 error message Now that the physical volume is removed we can bring it back into the volume group and therefore create new good control data on this disk Running extendvg workvg hdisk6 gives us lspv grep hdisk6 hdisk6 00017d372 67da78 None extendvg workvg hdisk6 0516 796 extendvg Making hdisk6 a physical volume Please wait lspv grep workvg hdisk5 00017d372f1de834 workvg hdisk6 00017d3738adc5d4 workvg Note that hdiske has aquired a new PVID the extendvg doesn t ask for confirmation because there was no PVID on this disk We can vary on and vary off the volume group without difficulties but 1vo2 is still unavailable in fact it disappeared and hence so is our data varyonvg workvg lsvg 1 workvg workvg LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT loglv00 jfslog 1 1 1 closed syned N A AIX Logical Volume Manager from A to Z Troubleshooting and Commands This is where the mapfile we cr
93. s SeekBackup Specifies the backup to seek and restore ona multiple backup tape archive The s flag is only applicable when the archive is written to a tape device To use the s flag properly a no rewind on close and no retension on open tape device such as dev rmt0 1 or dev rmt0 5 must be specified If the s flag is specified with a rewind tape device the restore command displays an error message and exits with a non zero return code If a no rewind tape device is used and the s flag is not specified a default value of s 1 is used The value of the SeekBackup parameter must be in the range of 1 to 100 inclusive It is necessary to use a no rewind on close no retension on open tape device because of the behavior of the s flag The value specified with s is relative to the position of the tape s read write head and not to an archive s position on the tape For example to restore the first second and fourth backups from a multiple backup tape archive the respective values for the s flag would be s1 s1 and s2 Displays information about the backup archive If the archive is in file name format the information contained in the volume header and a list of files found on the archive are written to standard output The File parameter is ignored for file name archives If the archive is in file system format the behavior is identical to the t flag Displays information about the backup archive If the archive
94. share the same physical volume S Sets a super strict allocation policy so that the partitions allocated for one mirror cannot share a physical volume with the partitions from another mirror 252 AIX Logical Volume Manager from A to Z Troubleshooting and Commands u UpperBoundl Sets the maximum number of physical volumes for new allocation The value of the Upperbound variable should be between one and the total number of physical volumes When using striped logical volumes or super strictness the upper bound indicates the maximum number of physical volumes allowed for each mirror copy Note When creating a superstrict logical volume you must specify physical volumes or use the u flag Commands called allocp getlvodm lextendlv lquerylv lquerypv lqueryvg putlvodm and lresyclv m Note The e m s and u flags are not valid with a striped logical volume The mklvcopy command increases the number of copies of each logical partition in LogicalVolume This is accomplished by increasing the total number of physical partitions for each logical partition to the number represented by Copies The LogicalVolume parameter can be a logical volume name or logical volume ID You can request that the physical partitions for the new copies be allocated on specific physical volumes within the volume group with the PhysicalVolume parameter otherwise all the physical volumes within the volume group are a
95. size restrictions shown in Table 1 on page 330 A Specifies the attributes for auto mount yes File system is automatically mounted at system start no File system is not automatically mounted at system start d Device Specifies the device name of a device or logical volume on which to make the file system This is used to create a file system on an already existing logical volume g VolumeGroup Specifies an existing volume group on which to make the file system A volume group is a collection of one or more physical volumes 336 AIX Logical Volume Manager from A to Z Troubleshooting and Commands LogPartitions Specifies the size of the log logical volume expressed as a number of logical partitions This flag applies only to JFS file systems that do not already have a log device m NewMountPoint Specifies the mount point which is the directory where the file system will be made available Note If you specify a relative path name it is converted to an absolute path name before being inserted into the etc filesystems file n NodeName Specifies the remote host name where the file system resides This flag is only valid with remote virtual file systems such as the Network File System NFS Sets the permissions for the file system ro Specifies read only permissions rw Specifies read write permissions Sets the accounting attribute for the specified file syst
96. small VG let delta 8192 hard coded for 4MB partitims 2048 4 blocks HDISK 1 VGID getlvodn j SHDISK VG cet lvodm t D echo SiG exec 2 gt dev mi11 while do MAGIC dd if dev SHDISK bs 512 skip Soffset count 1 2 gt dev mill head c 8 if SMAGIC AIX LVCB then PP echo Soffset 4352 8192 1 n bc echo AIX IVG found at PP PP offset Soffiset TYPE dd if dev SHDISK bs 512 skip Soffset comt 1 dd bs 1 skip 10 comt 31 gt TWWAVE dd if dev SHDISK bs 512 skip soffset camt 1 dd bs 1 skip 67 comt 31 gt set dd if dev SHDISK bs 512 skip Soffset comt 1 dd bs 2 skip 104 camt 2 od d let NUMLPS 2 echo Type TYPE echo Iname SLVNAME NUMLPS SNUMLPS exec 2 gt dev tty read decide Attempt restore y to do if Sdecide ny then echo SHDISK SPP exor SPP SNUMLPS gt tmp map_restoress nklv y SLVNAME m tmp map_restores SVG SNUMLPS m trp map_restares fi exec 2 gt dev mi11 echo fi let offset offset delta Appendix E Scripts used during this residency 399 400 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Appendix F Special notices This publication is intended to help customers and systems engineers to better understand the AIX logical volume manager in order to make better proposals or to develop their storage strategy on RS 6000 systems See the PUBLICATIONS section of the IBM Programming Announceme
97. spans multiple disks and also if an LVM level bad block relocation had occurred The script in Appendix E 8 scraper on page 398 performs this simple laying down of maps Chapter 2 Problem determination and recovery 167 mklv y scrapelv scrapevg 5 scrapelv mklv t jfslog y scrapelog scrapevg 1 scrapelog crfs d dev scrapelv a logname dev scrapelog m scrapefs v jfs Based on the parameters chosen the new scrapefs JFS file system is limited to a maximum size of 134217728 512 byte blocks New File System size is 40960 echo y logform dev scrapelog mount scrapefs echo data gt scrapefs data 1s 1 scrapefs total 16 Yw r r 1 root sys 5 Oct 01 13 58 data drwxrwx 2 root system 512 Oct 01 13 56 lost found umount scrapefs mlv f scrapelog scrapelv mlv Logical volume scrapelog is removed mlv Logical volume scrapelv is removed I J Here rmlv is used but the procedure would be the same if we had destroyed the whole VGDA we would also have to fix etc filesystems 168 AIX Logical Volume Manager from A to Z Troubleshooting and Commands scraper hdisk6 scrapevg ATX LVCB found at PP1 offset 4352 Type jfs LVname ppllv NUMLPS 1 Attempt restore y to do n ATX LVCB found at PP200 offset 1634560 Type jfslog LVname toastlog NUMLPS 1 Attempt restore y to do n ATX LVCB found at PP210 offset 1716480 Type jfs LVname toast NUMLPS 7 At
98. system to defrag Perform Query or report Select the option Perform Run the defragmentation of the file system Query Report the current fragmentation state of the file system Report Produce a report as if the defrag had been run smit dejfs g N Defragment a Journaled File System Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields File System Name Perform Query or Report perform F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image Si F9 Shell F10 Exit Enter Do The smit dejfs can be used to obtain a report AIX Logical Volume Manager from A to Z Troubleshooting and Commands COMMAND STATUS Command OK stdout yes stderr no Before command completion additional instructions may appear below TOP statistics before running defragfs number of free fragments 364911 number of allocated fragments 11921 number of free spaces shorter than a block 0 number of free fragments in short free spaces 0 statistics after running defragfs number of free spaces shorter than a block 0 number of free fragments in short free spaces 0 other statistics number of fragments moved 0 number of logical blocks moved 0 number of allocation attempts 0 number of exact matches 0 BOTTOM F1 Help F2 Refresh F3 Cancel F6 Command F8 Image F9 Shel1 F1O Exit Find 3 n Find Next J 1 4 4 Mount a file system Mounting a fil
99. the command line or the SMIT panels These panel are located under the System Storage Management tree as shown in the following screen a A System Management Move cursor to desired item and press Enter Software Installation and Maintenance Software License Management Devices System Storage Management Physical amp Logical Storage Security amp Users Comunications Applications and Services Print Spooling Problem Determination Performance amp Resource Scheduling System Environments Processes amp Subsystems Applications Using SMIT information only F1l Help F2 Refresh F3 Cancel F8 Image F9 Shel1 F10 Exit Enter Do Figure 1 on page 2 shows the layout of the system storage management SMIT panels Copyright IBM Corp 2000 1 SMIT System Storage Management Logical Volume Manager File Systems Files amp Directories Removable Disk Management System Backup Manager y File Systems Removable Disk Management List All File Systems List All Mounted File Systems List All Mounted File Systems on a Disk Add Change Show Delete File Systems Unmount File Systems on a Disk Mount a File System Remove a Disk form the Operating System Mount a Group of File Systems Remove a Disk Unmount a File System Open Door Unmount a Group of File Systems Verify a File System Backup a File System Restore a File System List Cont
100. the file to be copied to the root directory of the CD file system This file could be a customization script specified in the bosinst data file such as CUSTOMIZATION_FILE filename For example If the file my_script is in tmp on the machine where mkca is running then enter z tmp my_script and specify CUSTOMIZATION_FILE my_script The code copies the script to the root directory of the RAM file system before it executes Note 1 If you are creating a non bootable CD using the s flag you cannot use the por 1 flags 2 If you are creating a non bootable CD with a savevg image using the s or v flags you cannot use the p 1 u i z or b flags The mckd command creates a system backup image mksysb to CD Recordable CD R from the system rootvg or from a previously created mksysb image It also creates a volume group backup image savevg to CD R from a user specified volume group or from a previously created savevg image With mkcd you can create three types of CDs Personal system backup generic backup and a non bootable volume group backup 242 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Personal system backup CDs can only boot and install a specific machine and are similar to using mksysb on tape Generic backup CDs can boot and install any RS 6000 platform rspc rs6k or chrp This backup requires all the necessary device support including the MP kernel to
101. the fsck command This flag lets the fsck command take any action it considers necessary Use this flag only on severely damaged file systems 350 m Note Always run the fsck command on file systems after a system malfunction Corrective actions may result in some loss of data The default action for each consistency correction is to wait for the operator to enter yes or no If you do not have write permission for an affected file system the fsck command defaults to a no response in spite of your actual response AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note 1 The fsck command does not make corrections to a mounted file system 2 The fsck command can be run on a mounted file system for reasons other than repairs However inaccurate error messages may be returned when the file system is mounted The fsck command checks and interactively repairs inconsistent file systems You should run this command before mounting any file system You must be able to read the device file on which the file system resides for example the dev hdO device Normally the file system is consistent and the fsck command merely reports on the number of files used blocks and free blocks in the file system If the file system is inconsistent the fsck command displays information about the inconsistencies found and prompts you for permission to repair them The fsck command is conservative in its
102. the mkszfile command ensure that enough space is available in the tmp file to store a boot image This space is needed during both backup and installation To determine the amount of space needed in the tmp file issue the following command bosboot qad dev hdiskn where hdiskn is the physical volume that contains hd5 use lslv 1 ha5 or bootinfo b A 21 The mkvg command The following summarizes the options for the mkvg command mkvg Creates a volume group Usage mkvg d MaximumPhysicalVolumes B G f i c x m MaxPvSize n s Size t factor V MajorNumber y VolumeGroup Physical_Volume s B Creates a big VGDA volume group This can consist of up to 128 physical volumes and 512 logical volumes and is not supported on AIX 4 3 1 or below C Creates a concurrent capable volume group Can only be used with HACMP and has no effect if HACMP is not being used Not supported prior to AIX Version 4 2 d Max PVs The maximum number of physical volumes in the volume group default is 32 or 128 for a big VGDA volume group This can be used to minimize the space occupied by the VGSA VGDA so the value must be 32 or 128 for a big VGDA Appendix A High level LVM commands 259 Forces the volume group to be created on the specified physical volume unless the physical volume is part of an active volume group or a volume group defined in the configuration database
103. the next volume press the Enter key to continue restoring files 2 If an archive created using the backup command is made to a tape device with the device block size set to 0 it may be necessary for you to have explicit knowledge of the block size that was used when the tape was created in order to restore from the tape 3 Multiple archives can exist on a single tape When restoring multiple archives from tape the restore command expects the input device to be a no retension on open no rewind on close tape device Do not use a no rewind tape device for restoring unless the B s or x flag is specified For more information on using tape devices see the rmt special file File system archives File system archives are also known as i node archives due to the method used to archive the files A file system name is specified with the backup command and the files within that file system are archived based on their structure and layout within the file system The restore command restores the files on a file system archive without any special understanding of the underlying structure of the file system When restoring file system archives the restore command creates and uses a file named restoresymtable This file is created in the current directory The file is necessary for the restore command to do incremental file system restores m Note Do not remove the restoresymtable file if you perform incremental file syste
104. the specifications in the mapfile Usage Imigratelv I LVID s Size Mapfile I LVID The LVID of the logical volume to be moved s Size The size in logical partitions of the logical volume Appendix B Intermediate level commands 299 Mapfile The mapfile describing the migration moves and has the following records on each line Src_PVID Src_LPNum Dest_PVID Dest_LPNum B 19 The Imigratepp command The following summarizes the options for the Imigratepp command Imigratepp An object file command that moves a physical partition to a specified physical volume Usage Imigratepp g VGID p old_PVID n old_PPNum P new_PVID N new_PNum g VGID The VGID of the volume group that contains both the origin and destination physical volumes p old_PVID The PVID of the physical volume that contains the physical partition that is to be moved n old_PPNum The physical partition number of the partition that is to be moved P new_PVID The PVID of the physical volume to which the physical partition should be moved N new_PPNum The destination physical partition number B 20 The Iquerylvy command The following summarizes the options for the 1qureylv command Iquerylv An object file command that queries the attributes of a logical volume Usage Iquerylv L LVID p PVName N G n M S c s P R v o w a d 1 A r
105. the use of putlvodm v and p in the last and the next piece of code These may be used in the recovery situation to tweak the ODM without having to build stanzas for odmadd ae test s tmp pvlistss then while read PVID PVNAME do putlvodm P PVID 2 gt dev null delete old entry putlvodm p SKEY VGID PVID add new entry if 0 then dspmsg s 1 cmdlvm cat 576 lvmmsg 576 n SPROG gt amp 2 fi done lt tmp pvlists else dspmsg s 1 cmdlvm cat 567 lvmmsg 567 gt n SPROG gt amp 2 exit fi test s tmp pvlist22794 0 lt tmp pvlist22794 read PVID PVNAME Finally we rebuild the entries in dev and work with the trusted computing base if necessary 144 AIX Logical Volume Manager from A to Z Troubleshooting and Commands t Remove any existing node for the VG recreate fresh one to insure correctness f dev SVGNAME 2 gt dev null Note lvrelminor is being bypassed m f dev targetvg 2 gt dev null knod dev SVGNAME c MAJOR 0 mknod dev targetvg c 41 0 3 B H if n SLFLAG then chown L _OWNER L GROUP dev S VGNAME 2 gt dev null chmod L_PERMS dev SVGNAME 2 gt dev null fi n Add modify the dev entry to TCB if ON if z TCB OFF a x usr bin tcbck then tcbck 1 dev SVGNAM EE fi z PdAt uniquetype un attribute TCB STATE deflt tcb disabled values width type generic r
106. the volume group then it is returned otherwise a new number is returned Appendix B Intermediate level commands 307 B 31 The lvgenminor command The following summarizes the options for the lvgenminor command Ivgenminor An object file command that gets the next available device minor number Usage Ilvgenminor p PreferredMinNum MajorNum NewDeviceName p PreferredMinNum Used to provide the preferred minor number If this number is not available then an error is returned MajorNum Device major number NewDeviceName Name of device B 32 The lvchkmajor command The following summarizes the options for the lvchkmajor command Ivchkmajor An object file command that checks if the major number is valid for the given volume group Usage Ivchkmajor MajorNum VGName MajorNum The major number for the volume group to check VGName The volume group name If the provided major number is valid for the volume group then it is returned otherwise 1 is returned B 33 The lvistmajor command The following summarizes the options for the 1vlstmajor command Ivistmajor An object file command that returns a list of unused major numbers Usage Ivistmajor 308 AIX Logical Volume Manager from A to Z Troubleshooting and Commands B 34 The lvmmsg command The following summarizes the options for the 1vmmsg command Ivmmsg This is an object fil
107. this allocation alone Again the policy of the logical volume is not changed Maximum number of physical volumes to use for the allocation Sets the upper bound for this allocation Allocate each logical partition on a separate physical volume Sets the strictness policy for this allocation Chapter 1 LVM commands 37 File containing the allocation map A map file can be used to specify which physical partitions are to be used smit extendlv a Increase the Size of a Logical Volume Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields LOGICAL VOLUME name grom_lv Number of ADDITIONAL logical partitions PHYSICAL VOLUME names POSITION on physical volume middle RANGE of physical volumes minimum MAXIMUM NUMBER of PHYSICAL VOLUMES 32 to use for allocation Allocate each logical partition copy yes on a SEPARATE physical volume File containing ALLOCATION MAP F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image F9 Shel1 F10 Exit Enter Do 1 2 12 List the configuration of a logical volume The 1slv command displays information about the logical volume Used without any arguments it returns the configuration of the logical volume Other flags can return information about the physical partition usage or logical partition mapping For the command line options see 1svg in Appendix A 11 The Isvg command on page 229 or use the SMIT fastpath
108. tmp pvlist22794 containing all the physical volumes that belong to our target volume group Next we have a thought provoking piece of commentary Chapter 2 Problem determination and recovery 139 NOTE In importvg we already checked to insure that VGNAME is not already in use if used then print error message and exit redefinevg would not be called at all We decided to have the checking here also because redefinevg is a separated command and in the future it might be used in other places or used by itself When we know for sure that we do not need the ckecking here anymore then we will delete it ttt HOH OH OH OH Another valid solution to the problem of a loss of all ODM entries is simply to run an importvg y targetvg hdiske This will produce the output importvg y targetvg hdisk6 targetvg 0516 012 lvaryoffvg Logical volume must be closed If the logical volume contains a filesystem the umount command will close the LV device 0516 942 varyoffvg Unable to vary off volume group targetvg Despite the above error message caused because the victimfs file system is still mounted running an 1svg 1 would confirm that all ODM objects have been re created Some may prefer to use this one high level command rather than the two intermediate level commands redefinevg and synclvodm On the other hand the intermediate command combination will finish quicker than an importvg Sometimes there s more than one w
109. to AIX Version 4 2 or later Node ID of this node if volume group is varied on in concurrent node Node IDs of other concurrent nodes that have this volume group varied on Maximum number of physical partitions per physical volume allowed for this volume group Maximum number of physical volumes allowed in this volume group 1 Only displayed if volume group is concurrent capable A 12 The Isvgfs command The following summarizes the options for the 1svgfs command Appendix A High level LVM commands 233 Isvgfs This is an object file command that lists the file systems that are in the specified volume group Usage Isvgfs VolumeGroup A 13 The Ilvedit command The following summarizes the options for the 1vedit command Ivedit The logical volume editor is used for interactive definition and placement of logical volumes within a volume group This command does not apply to AIX Version 4 2 or later Usage Ivedit VolumeGroup 234 The 1vedit command invokes the logical volume editor in the context of a particular volume group VolumeGroup From within the editor a user can inspect the current state of logical and physical volumes in the volume group and can make changes to this state interactively A user can make the same changes that are possible using the mklv chlv extendlv and rmlv commands but can have much more precise control over the placement of logical volumes on physical volumes
110. to Z Troubleshooting and Commands Logical Volumes List All Logical Volumes by Volume Group Show Characteristics of a Logical Volume Remove a Logical Volume Copy a Logical Volume Change a Logical Volume Rename a Logical Volume Increase the Size of a Logical Volume Add a Copy to a Logical Volume Remove a Copy from a Logical Volume Figure 3 SMIT logical volume options 1 2 1 Add a logical volume The mklv command creates a new logical volume within the given volume group Physical partitions can be allocated to the logical volume from one physical volume or many Allocation policies can also be used to determine which physical partitions should be used to make up the logical volume if there are enough free partitions in the volume group to allow choice When the logical volume is created it can be created with extra copies as a mirrored logical volume or additional copies can be added later on Note AIX Version 4 3 3 supports logical volumes that are both striped and mirrored These logical volumes cannot be used on systems with lower versions of AIX For the command line options see mklv in Appendix A 17 The mklv command on page 243 or use the SMIT fastpath smit mklv SMIT provides the following fields Logical volume name A system wide unique one to 15 character name Chapter1 LVMcommands 23 24 Number of logical partitions The number of lo
111. verify that it is not already in another volume group If the system believes the physical volume belongs to a volume group that is varied on it exits But if the system detects a description area from a volume group that is not varied on it prompts the user for confirmation in continuing with the command The previous contents of the physical volume are lost so the user must be cautious when using the override function A 8 The importvg command The following summarizes the options for the importvg command importvg Imports a new volume group definition from a set of physical volumes R PhysicalVolume Usage importvg V MajorNumber y VolumeGroup f c x L VolumeGroup n F C Imports the volume group and creates it as a Concurrent Capable volume group Only use the c flag with the HACMP It has no effect on volume groups and systems not using the HACMP product This flag only applies to AIX Version 4 2 or later Forces the volume group to be varied online 216 AIX Logical Volume Manager from A to Z Troubleshooting and Commands LVolumeGroup Takes a volume group and learns about possible changes performed to that volume group Any new logical volumes created as a result of this command emulate the ownership group identification and permissions of the dev special file for the volume group listed in the y flag The 1 flag performs the functional equivalent of the F
112. volume groups VGID putlvodm An object file command that puts physical volume values into the Configuration Database Appendix B Intermediate level commands 313 Usage putlvodm p VGID PVID p VGID The volume group ID for which the physical volume should be added PVID The physical volume ID to add putlvodm An object file command that removes a volume group from the Configuration Database Usage putlvodm V VGID V VGID The volume group ID for which the volume group information is to be removed putivodm An object file command that removes logical volume values from the Configuration Database Usage putlvodm L LVID L LVID The LVID of the logical volume for which logical volume data is to be removed putlvodm An object file command that removes physical volume values from the Configuration Database Usage putlvodm P PVID P VGID The PVID of the physical volume for which physical volume data is to be removed 314 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Appendix C ODM commands A large part of the information that the Logical Volume Manager relies on is included in the ODM This appendix lists the commands that allow you to work with the ODM adding removing or modifying items C 1 The odmadd command The following summarizes the options for the odmadd command odmadd An object file
113. which to mount TYPE of file system FORCE the mount no REMOTE NODE containing the file system to mount Mount as a REMOVABLE file system no Mount as a READ ONLY system no Disallow DEVICE access via this mount no Disallow execution of SUID and sgid programs no in this file system F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image C F9 Shell F10 Fxit Enter Do 56 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 1 4 5 Unmount a file system Unmounting a file system makes it unavailable for use There can be no files open This will cause the unmount to fail unless the force option is used For the command line options see umount in Appendix D 32 The umount command on page 391 or use the SMIT fastpath smit umountfs SMIT provides the following fields Unmount all mounted file systems Choosing yes will unmount all local file systems except tmp and usr Unmount all remotely mounted file systems Choosing yes will unmount all remote file systems Name of file system to unmount Specify the device directory or file system to be unmounted Remote node containing the file system s to unmount Choose the node that is holding the file system to be unmounted smit umountfs C N Unmount a File System Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields Unmount ALL mounted file syst
114. within the group PVstate State of the physical volume Total PPs Total number of physical partitions on the physical volume Free PPs Number of free physical partitions on the physical volume Distribution The number of physical partitions allocated within each section of the physical volume Outer edge outer middle center inner middle and inner edge of the physical volume Appendix A High level LVM commands 229 Lists the following information for each logical volume within the group specified by the VolumeGroup parameter LV A logical volume within the volume group Type Logical volume type LPs Number of logical partitions in the logical volume PPs Number of physical partitions used by the logical volume PVs Number of physical volumes used by the logical volume Logical volume State of the logical volume Opened stale indicates the logical volume is open but contains partitions that are not current Opened syncd indicates the logical volume is open and synchronized Closed indicates the logical volume has not been opened Mount point File system mount point for the logical volume if applicable i Reads volume group names from standard input 230 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Lists the following fields for each logical volume on the physical volume PVname PPnum LVname LPnum Copynun PPstate PVname Name of the physic
115. y The volume group is automatically activated during system startup Appendix A High level LVM commands 205 B Changes the volume group to big VG format This can accommodate up to 128 physical volumes and 512 logical volumes Notes 1 The x flag cannot be used if there are any stale physical partitions or there are any open logical volumes in the volume group 2 Once the volume group is converted it cannot be imported into AIX Version 4 3 1 or lower versions 3 The B flag cannot be used if the volume group is varied on in concurrent mode 4 There must be enough free partitions available on each physical volume for the VGDA expansion for this operation to be successful 5 Since the VGDA resides on the edge of the disk and it requires contiguous space for expansion the free partitions are required on the edge of the disk If those partitions are allocated for user usage they will be migrated to other free partitions on the same disk The rest of the physical partitions will be renumbered to reflect the loss of the partitions for VGDA usage This will change the mappings of the logical to physical partitions in all the PVs of this VG If you have saved the mappings of the LVs for a potential recovery operation you should generate the maps again after the completion of the conversion operation Also if the backup of the VG is taken with the map option and you plan to restore using those maps the restore operatio
116. you to change the DestinationPV The replacepv command replaces allocated physical partitions and the data they contain from the SourcePV to DestinationPV The specified SourcePV cannot be the same as DestinationPV 266 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note 1 The DestinationPV size must be at least the size of the SourcePV 2 The replacepy command cannot replace a SourcePV with a stale logical volume unless this logical volume has a non stale mirror The allocation of the new physical partitions follows the policies defined for the logical volumes that contain the physical partitions being replaced A 28 The rmlv command The following summarizes the options for the xmlv command rmlv Removes logical volumes from a volume group Usage rmlv B f p Physical Volume LogicalVolume s B Issues a chlvcopy B s for the parent logical volume if the logical volume was created using the 1 flag If itis a regular logical volume then the B flag is ignored f Removes the logical volumes without requesting confirmation p PhysicalVolume Removes only the logical partition on the PhysicalVolume The logical volume is not removed unless there are no other physical partitions allocated Commands called getlvodm ldeletelv lqueryly lreducelv putlvodm and lvrelminor m Note This command destroys all data in the specified logic
117. 0 This done further checks are made Chapter 2 Problem determination and recovery 141 142 Va check to see whether dev entry for the volume group already existed if r dev SVGNAME then VGEXISTED 1 node was there else VGEXISTED 0 node was not there fi r dev targetvg VGEXISTED 1 if n SLFLAG then some code for learning mode to protect the volume group s permissions has been ommitted here fi n use user supplied major number or system supplied major number if z SvFLAG then MAJOR lvgenmajor VGNAME else MAJOR lvchkmajor vVAL S VGNAMI fi z lvgenmajor targetvg MAJOR 41 eel if eq 1 then dspmsg s 1 cmdlvm cat 568 lvmmsg 568 n SPROG gt amp 2 exit fi 0 eq 1 We are using the system supplied major number release the minor number before trying to get it back lvrelminor VGNAME gt dev null 2 gt amp 1 lvrelminor targetvg 1 gt dev null 2 gt amp 1 MINOR lvgenminor p 0 SMAJOR S VGNAME lvgenminor p 0 41 targetvg MINOR 0 if 0 then dspmsg s 1 cmdlvm cat 570 Ivmmsg 570 n SPROG gt amp 2 exit fi 0 0 We save the minor number in SMINOR and sanity check p The next and final piece of information gathering sets the AUTO_ON and QUORUM variables AIX Logical Volume Manager from A to Z Troubleshooting and Com
118. 1 LVM commands 21 Mount point The mount point of the associated file system smit lsvg Physical volume list option C N COMVAND STATUS Commend OK stdout yes stderr no Before command carpletio additional instructions may appear below asgard vg PV NAVE PV SIATE TOAL PPs FREE PPs FREE DISTRIBUTION hdisk11 active 537 492 108 62 107 107 108 hdisk9 active 537 517 108 87 107 107 108 hdisk10 active 537 537 108 107 107 107 108 Fl Help F2 Refresh F3 Cancel F6 Comend F8 Inage F9 Shel1 FLO Bxit Find Uri Next y Explanation of the fields displayed PV Name Name of the physical volume PV State The physical volume state active in use missing physical volume missing removed for a removable physical volume varied off volume group varied off Total PPs The total number of physical partitions on the physical volume Free PPs The number of physical partitions currently not allocated to a logical volume Free distribution The number of physical partitions in each region Regions are shown from right to left as outer edge outer middle center inner middle and inner edge 1 2 Logical volume related commands The commands in this section relate to creating and modifying logical volumes These commands can be run from the command line or from the SMIT System Storage management gt Logical Volume Manager gt Logical Volumes menu See Figure 3 22 AIX Logical Volume Manager from A
119. 120 hdisk11 BOTTOM F1 Help F2 Refresh F3 Cancel F6 Command F8 Image F9 Shel1 F1O Exit Find ke n Find Next Explanation of the fields displayed LP The logical partition number PP1 3 Physical partition number for copy 1 to 3 PV1 3 Physical volume that the physical partition resides on for copy 1through 3 1 3 Physical volume related commands The commands in this section relate to creating and modifying physical volumes These commands can be run from the command line or from the SMIT System Storage management gt Logical Volume Manager gt Physical Volumes menu See Figure 4 Chapter 1 LVM commands 43 Physical Volumes List All Physical Volumes in System Add a Disk Change Characteristics of a Physical Volume List Contents of a Physical Volume Move Contents of a Physical Volume Figure 4 SMIT physical volume options 1 3 1 List all physical volumes in system 44 The getlvodm c command will list all the configured physical volumes in the system For the command line options see getlvodm in Appendix B 7 The getlvodm command on page 291 or use the SMIT fastpath smit pv smit pv List all physical volumes in system a N COMMAND STATUS Command OK stdout yes stderr no Before command completion additional instructions may appear below hdisko hdisk1 hdisk2 hdisk3 hdisk5 hdisk4 hdisk6 hdisk9 hdisk10 hdisk11 hdisk7 F1 Help F2
120. 23 4128 If no criteria are specified all object entries in the object class are changed InputFile The InputFile file has the same format as the InputFile file for the odmadd command C 3 The odmcreate command The following summarizes the options for the odmcreate command odmcreate An object file command that produces the c source and h include files necessary for ODM application development and creates empty object classes Usage odmcreate p c h ClassDescriptionFile C Creates empty object classes only it does not generate the C language h and c files h Generates the c and h files only it does not create empty classes p Runs the C language preprocessor on the ClassDescriptionFile file ClassDescriptionFile The ClassDescriptionFile parameter specifies an ASCII file that contains descriptions of one or more object classes The odmcreate command is the ODM class compiler The command takes as input an ASCII file that describes the objects a user wishes to use in a 316 AIX Logical Volume Manager from A to Z Troubleshooting and Commands specific application The odmcreate command can create empty object classes as part of its execution The output of the odmcreate command is an h file an include file that contains the C language definitions for the object classes defined in the ASCII ClassDescriptionFile file The resulting include file is used by the
121. 305 B 27 The lresyncpv command 0 0 0 cee 305 B 28 The Ilvaryonvg command 0 0 00 cee ee 306 B 29 The Ivaryoffvg command 0 0 00 cee ees 307 B 30 The lvgenmajor command 0 00 e ee 307 B 31 The lvgenminor command 0 00 ee 308 B 32 The Ivchkmajor command 0 308 B 33 The Ivistmajor command 0 00 ee 308 B 34 The lvmmsg command 0 cee ee 309 B 35 The lvrelmajor command 0 00 eee eee 309 B 36 The lvrelminor command oo eresse veseri anrea erea es 309 B 37 The migfix command 0000 eee 309 B 38 The putlvcb command 0 ee ee 310 B 39 The putlvodm command 0 0 eee es 311 Appendix C ODM commands 020 eee eee eee 315 C 1 The odmadd command 0 cece eee 315 vi AIX Logical Volume Manager from A to Z Troubleshooting and Commands C 2 The odmchange command sssaaa saasa aaraa aeea 315 C 3 The odmcreate command 0 0 ee ee eens 316 C 4 The odmdelete command 0 0 cc eee ee ee eens 318 C 5 The odmdrop command 00 tee 318 C 6 The odmget command 0 0 ees 318 C 7 The odmshow command 0 0 00 cece eee eee 319 Appendix D Other related commands 321 D 1 The backup command 0 000 eee 321 D 2 The chfScommand 0 0 cee eee ee eee eens 327 D 3 The chps command 000 eee eee 330 D 4 The
122. 37 1 open syncd home 0516 022 lsvg Illegal parameter or structure value 0516 304 lsvg Unable to find device id 0000000000000000 in the Device Configuration Database corrlv jfs 3 6 3 closed stale N A LL 421 J This could occur if migratepv or reorgvg commands fail and create a false mirror Note the error message occurs before the problem logical volume For AIX 4 3 however ordinarily the error message is the more explicit 0516 1147 lsvg Warning logical volume corrlv may be partially mirrored In fact any situation where a lextendlv command either ran partially or was given a map that does not cover the entire physical volume could create a false mirror and give this error S lslv m corrlv corrlv N A LP PP1 PV1 PP2 PV2 PP3 PV3 0001 0216 hdisk6 0216 hdisk7 0002 0217 hdisk6 0217 hdisk7 0003 0218 hdisk6 lsvg 1 corrvg corrvg LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT 0516 1147 lsvg Warning logical volume corrlv may be partially mirrored corrlv jfs 3 5 3 closed stale N A J This can occur if a mirrored logical volume is increased but not enough partitions are available to accommodate all mirrors However allocp should prevent this at AIX 4 3 3 All of these partial or false mirrored conditions can 160 AIX Logical Volume Manager from A to Z Troubleshooting and Commands be fixed by simply running a rmlvcopy against the appropriate disk unless the inadvisable practice of having b
123. 45 LV count 2 PV count 2 Total VGDAs 3 Conc Allowed 0 MAX PPs per 2032 MAX PVs 16 Logical 00017d37383736ec 1 loglv00 1 00017d37383736ec 2 lv02 1 Physical 00017d372f1de834 2 0 00017d372f67da78 1 0 Total PPs 2150 lqueryvg g getlvodm v workvg At p hdisk6 Max LVs 256 PP Size 22 Free PPs 2145 LV count 2 PV count 2 Total VGDAs 3 Conc Allowed 0 MAX PPs per 2032 MAX PVs 16 Logical 00017d37383736ec 1 loglv00 1 00017d37383736ec 2 lv02 1 Physical 00017d372f 1de834 2 0 00017d372 67da78 1 0 Total PPs 2150 CH This confirms that the VGDAs themselves are intact In fact all the aa command did was erase the first 512 bytes of the disk 512 is the default blocksize for dd thus destroying hdisk6 s PVID Because we have some intact copies of the VGDA we can pull out the LP maps for each physical volume with 1spv m in the present case This is not always possible after the corruption event We now create a mapfile for Iv02 the unusable logical volume based upon these maps in the format used by mklv 80 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 1spv M hdisk5 hdisk5 1 215 hdisk5 216 hdisk5 217 1075 1spv M hdisk6 loglv00 1 hdisk6 1 215 hdisk6 216 1v02 1 hdisk 6 217 1v02 2 hdisk6 218 1v02 3 hdisk 6 219 1v02 4 hdisk6 220 1075 echo hdisk6 216 219 gt map _1v02 i Since the PVID information in the first block of the
124. 76c135837 00017d376b7ee896 read VGID ON MAJOR NUM ee 00017d375251870e 00017d376b7ee896 read VGID_ON MAJOR NUM 000174373eb17718 00017d376b7ee896 read VGID ON MAJOR NUM 00017d37e1762ac7 00017d376b7ee896 read VGID_ON MAJOR NUM J XN We have passed that test We now come to the core functionality of the script We are seeing the standard pattern for LVM high level scripts First parse and check the requested action is possible then gather required data and perform actions with that data This was seen for mkvg back in Figure 6 on page 61 The following section gathers up all the relevant logical volume ODM information for this volume group Get all the of the logical volume id numbers Remove all evidence of their existence from object data manager getlvodm L SVGNAME gt tmp lvinfo S 2 gt dev null getlvodm L victimvg 1 gt tmp lvinfol7482 2 gt dev null if s tmp lvinfoss then odmget q name SVGNAME CuDep grep dependency cut d f2 gt tmp lvinfoss LVID fi s tmp lvinfo17482 J This has been created with a temporary file tmp lvinfo17482 containing the following two lines victimlv 00017d376b7ee896 2 victimlog 00017d376b7ee896 1 The script continues information gathering Chapter 2 Problem determination and recovery 117 C N Is tcb turned on TCB OFF odmget q attribute TCB STAT
125. ADME rdwrtest should be read carefully before using this tool This tool first shipped with AIX4 3 2 Below AIX4 3 2 the following APARs should be installed AIX Level APAR 4 2 1 IX71128 4 1 5 IX70680 Another indication of this situation is if many DISK_ERR Or SCSIT_ERR are logged in the error report An example of using the error report was considered in 2 6 1 Checking the errorlog on page 88 2 7 5 Setting up notification of LYM_MISSPVADDED errors It is possible to configure the error logging subsystem to automatically notify the system administrator of LVM MISSPVADDED conditions To do this we first back up the errnctify ODM class cp etc objrepos errnotify etc objrepos errnotify bak Next we create a file tmp pvmiss add containing the following ODM object stanza Chapter 2 Problem determination and recovery 185 errnotify en pid 0 en name LVM MISSPVADDED en persistenceflg 1 en label LVM_MISSPVADDED en crcid 0 en class S en type UNKN en alertflg en resource LIBLVM en rtype NONE en rclass NONE G en method usr lib ras pvmiss notify 1 2 3 4 5 6 7 8 9 P The parameters of en method will be filled in by the error notification daemon A script to be run when the error occurs is also needed usr lib ras pwmiss notify Goce exec 3 gt dev console print u3 007 print u3
126. AIX Logical Volume Manager from A to Z Troubleshooting and Commands Laurent Vanel Ronald van der Knaap Dugald Foreman Keigo Matsubara Antony Steel International Technical Support Organization www redbooks ibm com SG24 5433 00 SG24 5433 00 International Technical Support Organization AIX Logical Volume Manager from A to Z Troubleshooting and Commands March 2000 Take Note Appendix F Special notices on page 401 Before using this information and the product it supports be sure to read the general information in First Edition March 2000 This edition applies to AIX Version 4 3 Program Number 5765 C34 This document created or updated on March 20 2000 Comments may be addressed to IBM Corporation International Technical Support Organization Dept JN9B Building 003 Internal Zip 2834 11400 Burnet Road Austin Texas 78758 3493 When you send information to IBM you grant IBM a non exclusive right to use or distribute the information in any way it believes appropriate without incurring any obligation to you Copyright International Business Machines Corporation 2000 All rights reserved Note to U S Government Users Documentation related to restricted rights Use duplication or disclosure is subject to restrictions set forth in GSA ADP Schedule Contract with IBM Corp Contents PretaCe n si du nae Rah a acha te 2 ae ete Ye aS Gita oh Bao eaten ag ix The team that wro
127. ATOR getlvodm d SVGNAVE 2 gt dev ni11 if SVGVAIR eq SROOIMATR J then dsmsg s 1 ardlvm cat 762 lvmmsg 762 gt n SPROGDVE SVGAAVE gt 82 exit fi else dgomsg s 1 ardlvm cat 606 lvmmeg 606 gt n SPROGAVE gt amp 2 dsgomsg s 1 ardlvm cat 760 lvmmeg 760 gt n SPROGAVE gt amp 2 exit fi n adw VQ AVESbadvg sd s g 1s 1 dev IPL rootvg 1 gt tmp xootdevicel7818 read skipfld skipfld skipfld skipfld ROOIMAIOR skipfld 0 lt tmp rootdevicel7818 getlvodn v badvg 0516 306 getlvodm Unable to find badvg in the Device Configuraticn Database VGID test return 3 0516 772 exportvg Unable to export volume group badvg cleanup i y Here we clearly see the getlvodm v badvg fail get1vodm will only read from the logical volume control blocks LVCB and the ODM The v flag obtains the VGID from the ODM so we can rule out problems with the LVCBs Therefore there must be some kind of problem with information stored in the ODM Trying an odmget q name badvg AND attribute vgserial id Cuat returns nothing A general test for any relevant information is a command of the form odmget CuAt grep ip badvg This still returns nothing So we try the other Chapter 2 Problem determination and recovery 69 parts of the customized database However odmget CuDvDr grep ip badvg and odmget CuDv grep ip badvg both also give us nothing There are two choices at this point In the context of
128. Actions to perform in dealing with corrupt VGDA problems and common situations are discussed in LVM control data corruption on page 146 We now complete our examination of the low level control structures by looking at the LVCB The start of an LV is shown here for reference Running od a on the logical volume device is a quick sanity check for LVCB corruption od c dev hd1 2 0000000 A T x L V C B o 0 j f s o o o 0000020 o NO NO o i VO NG AG AO VO AG AO AG AG AO NO 6 0000040 o o o o o o o o o o 0 0 0 1 7 d 0000060 3 7 e ey 7 6 A a 7 R 8 o o o o 0000100 o o o h d 1 AO O o 0 o VO AG NO o i o 0000120 o o o o o o o o o o o o o o o o 0000200 o o o F saj i A u g 2 7 2 1 0000220 t 4 9 t 2 0 L 9 9 9 n o o o o 0000240 o Ss u n S e p 1 2 1 3 0 0000260 0 3 Y at 9 9 9 n o o o o Oo 1 0000300 7 D 3 7 C 0 o o y m c 0 y o 0000320 o 004 O 001 h fe m e o o o o o o o 0000340 o AO o NO AG AG AO AO AO ONO AO i o XO o AO XO 0000520 o o o o 1 o g d e v h d 8 0000540 m o u n t t r u e c h e c 0000560 k t r u e v o 1 h ome 0000600 f r e e f a 1 s e O0 Oo O O O 0000620 o o o o o o o o o o o o o o o o _ snipped P The text AIX LVCB should be at the start of the logical volume Note that when this is information is corrupt it is not apparent for big VGs since the remaini
129. B94 LVM HWREL UNKN H HARDWARE DISK BLOCK RELOCATION ACHIEVED 26120107 LVM MISSPVADDED UNKN S PHYSICAL VOLUME DEFINED AS MISSING 320B8ED9 LVM BBEPOOL UNKN H PV NO LONGER RELOCATING NEW BAD BLOCKS 33604AEF LVM BBDIR90 UNKN H BAD BLOCK DIRECTORY OVER 90 FULL 41BF2110 LVM MWCWFAIL UNKN H MIRROR WRITE CACHE WRITE FAILED 41E36337 LVM SA FRESHPP UNKN S PHYSICAL PARTITION MARKED ACTIVE 438E027E LVM BBDIRERR UNKN H PV NO LONGER RELOCATING NEW BAD BLOCKS 4523CAA9 CVMDLVM PERF H DISK OPERATION ERROR 52715FA5 LVM SA WRTERR UNKN H FATLED TO WRITE VOLUME GROUP STATUS AREA 56116BF9 LVM MWCENTRY NOT FO TEMP O MWC ENTRY NOT FOUND 613E5F38 LVM IO FATL PERM H I O ERROR DETECTED BY LVM 688B4101 LVM BBDIRFUL UNKN H BAD BLOCK RELOCATION FAILURE 80D3764C LVM BBFATI UNKN H PV NO LONGER RELOCATING NEW BAD BLOCKS 88453987 LVM BBDIRBAD UNKN H PV NO LONGER RELOCATING NEW BAD BLOCKS 9811EB50 LVM HWFATL UNKN H HARDWARE DISK BLOCK RELOCATION FAILED AF6582A7 LVM MISSPVRET UNKN S PHYSICAL VOLUME IS NOW ACTIVE CAD234BE LVM SA QUORCLOSE UNKN H QUORUM LOST VOLUME GROUP CLOSING D8CF8401 LVM SWREL UNKN H SOFTWARE DISK BLOCK RELOCATION ACHIEVED DAFEE4D6 LVM BBRELMAX UNKN H PV NO LONGER RELOCATING NEW BAD BLOCKS EFAA3D429 LVM SA STALEPP UNKN S PHYSICAL PARTITION MARKED STALE F7DDA124 LVM_SA PVMISS UNKN H PHYSICAL VOLUME DECLARED MISSING Li J In the following example we fake a disk failure A volume group is created with quorum set to off mkvg f
130. Commands Free PPs Used PPs Free distribution Used distribution Allocatable Logical volumes VG descriptors Number of free physical partitions on the physical volume Number of used physical partitions on the physical volume Number of free partitions available in each intra physical volume section Number of used partitions in each intra physical volume section Allocation permission for this physical volume Number of logical volumes using the physical volume Number of volume group descriptors on the physical volume 1 3 3 List contents of a physical volume The 1spv command will also display information about the logical volumes on the physical volume and the physical partition usage by region and logical volume As seen in the previous section the logical volumes or physical partitions can be selected Explanation of the fields displayed LV Name LPs PPs Distribution Mount point The logical volume name The number of logical partitions used by the logical volume The number of physical partitions used by the logical volume The number of physical partitions laying in each region on the physical volume shown from outer edge to inner edge The mount point of the file system if appropriate Chapter 1 LVM commands 47 smit lspv Show logical volumes on the physical volume gt N COMMAND STATUS Command OK stdout yes stderr no Before command compl
131. Commands smit lsvg Volume group configuration fy N COMVAND STATUS Command OK stdout yes stderr no Before command carpletio additional instructions may appear below VOLUME GROUP asard vg VG HENTIFIER 000170B78edE7c9d VG SIATE active PP SIZE 4 megabyte s VG PERMISSION read write TOAL PPs 1611 6444 megabytes MAX LVS 256 FREE PFs 1546 6184 megabytes IVs 5 USED PPs 65 260 megabytes OPEN IVs 0 QUORLM 2 TAL Fs 3 VG DESCRIPICRS 3 STALE PVs 0 STALE PPs 0 ACTIVE PVs 3 AUIO ON yes MAX PPs per PV 1016 MAX PVS 32 Fl Help F2 Refresh F3 Cancel F Comend F8 Inrage F9 Shel1 FLO Exit Find n Find Next S Explanation of the fields displayed Volume group Volume group state Permission Max LVs LVs Open LVs Total PVs Name of the volume group State of the volume group If the volume group is activated with the varyonvg command the state is either active complete indicating all physical volumes are active or active partial indicating some physical volumes are not active If the volume group is not activated with the varyonvg command the state is inactive Access permission read only or read write Maximum number of logical volumes allowed in the volume group Number of logical volumes currently in the volume group Number of logical volumes within the volume group that are currently open Total number of physical volumes within the volume group Ch
132. D t VGID Returns the volume group name for this VGID u VGDescriptor Returns the auto on value for this volume group name or VGID The value returned is as follows y Auto vary on set n Auto vary on disabled v VGDescriptor Returns the VGID for this volume group name or VGID w VGDescriptor Returns the physical volume names and PVIDs for this volume group name or VGID y LVID Returns the logical volume type for this LVID AIX Logical Volume Manager from A to Z Troubleshooting and Commands X VGDescriptor Returns the concurrent capable flag for the volume group name or VGID x VGDescriptor Returns the auto on concurrent flag for the volume group name or VGID The get1lvodm command gets logical volume data from the Configuration Database and writes it to standard out stdout The command line option specifies what information will be retrieved The LVDescriptor PVDescriptor or VGDescriptor can be either an ID ora name B 8 The getvgname command The following summarizes the options for the getvgname command getvgname An object file command that returns a new unused volume group name This command is undocumented Usage getvgname n Y Prefix VGName N Returns whether this volume group name is in use or not Y Prefix The getvgname command returns a volume group name The name is written to standard out std
133. DD H define PSN MWC RECO 2 PSN of the first MWC cache record define PSN MWC RECI 3 PSN of the second MWC cache record define PSN LVM REC 7 PSN of IM information record define PSN BB DIR 8 beginning PSN of bad block directory define LEN BB DIR 22 length in sectors of bad block dir define PSN CFG RAK 64 PSN of the backup config record define PSN LVM BAK 70 PSN of backup IM information record define PSN BB BAK 71 PSN of backup bad block directory define PSN CFG IMP 120 PSN of concurrent config work record define PSN NONRSRVD 128 PSN of first non reserved sector Figure 10 usr include sys hd_psn h Since a physical sector is 512 bytes each of them begins at offset of the Physical Sector Number PSN 512 bytes We can look at these sectors by running od cx dev hdisk9 note that od displays offsets in octal and 512 is octal 1000 Note that not every block in this region is used and some blocks are reserved for other code than LVM Chapter 2 Problem determination and recovery 121 od cx devw hdisk9 ooococoo0 A A o o 0 o 0 o o Wo o o io o c9c2 d4c1 0000 0000 0000 0000 0000 0000 0000020 0 o o o o 0 0 0 0 o o o o 0 o 0 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 Figure 11 Contents of the first physical sectors on an hdisk Blocks 1 The IPL record Note that the LVM maintains t
134. E and deflt tcb disabled PdAt odmget q attribute TCB STATE and deflt tcb disabled PdAt TCB_OFF PdAt uniquetype attribute TCB STATE deflt tcb disabled values width type un generic reo G nls _ index 0 A LVM code has to work with the trusted computing base if its turned on We check for this in PdAt At this point we have all the data we need We loop through each logical volume in tmp lvinfo17482 and perform the relevant actions for it while read LVNAME LVID do remove entries from TCB database if TCB is On if z TCB_OFF a x usr bin tcbck then tebck d dev LVNAME dev r LVNAME gt dev null 2 gt amp 1 fi imfs lx SLVNAM GW done lt tmp lvinfos 0 lt tmp lvinfol7482 read LVNAME LVID z PdAt uniquetype attribute TCB STATE deflt tcb disabled values width type generic rep nn nls index 0 a x usr bin tcbck imfs 1x victimlv S Calling imfs like this deletes any relevant entries from etc filesystems for the LV This line is why exportvg should not be used as the ultimate solution for every LVM problem 118 AIX Logical Volume Manager from A to Z Troubleshooting and Commands read LVNAME LVID z PdAt uniquetype 00 attribute TCB STATE deflt tcb disabled values width
135. GKILL kill 9 to interrupt any LVM command This opens a window to ODM or LVM database corruption Again looking at chvg traced with trclvm we see exit trap will handle cleanup exit cleanup getlvodm R 0 eq 0 savebase rm f tmp pvmap 00091974d81f 431 18298 exit 2 A couple of useful observations can be made here e The cleanup routine tidies up various files that are generated by high level command scripts in the tmp directory Since these scripts are created by high level commands using the syntax filenamess they all end in the process ID of the creator script The importance of this is that if we modify the scripts so temporary files are not removed they are available for later examination and are easily identifiable by process ID This facility is part of the example trclvm script when run with the t flag e As part of cleanup a savebase may be run to synchronize the mini ODM within the boot logical volume with the runtime ODM database get1lvodm R checks for the presence of the runtime attribute within PdAt It is often the Chapter 2 Problem determination and recovery 65 case that after LVM problems have been fixed manually that is by working at a lower level than the high level command scripts do running savebase is neglected This may lead to a reoccurrence of the problem that has been fixed the next time the machine is rebooted 2 3 Corruption example 1 Simple OD
136. LLELFLAG a n SPFLAG then dspmsg s 1 cmdlvm cat 1149 lvmsg 1149 gt n SPROG gt amp 2 fi L 0 an o n a n All the flags were compatible Here we only had 1 so we now parse the command line arguments E N Parse command line arguments if i option then read values in from standard in and save in name list if n SIFLAG then while read LINE do namelist Snamelist LINE C done P Chapter 2 Problem determination and recovery 175 This allows us to pipe in arguments to syncvg with traditional UNIX semantics We are now building the namelist mentioned at the start of this commentary on the script else get arguments from command line and save in namelist D else shift skip past from getopt if i 1 J then namelist else dspmsg s 1 cmdlvm cat 618 lvmmsg 618 n SPROG gt amp 2 dspmsg s 1 cmdlvm cat 1145 lvmmsg 1145 gt n PROG gt amp 2 exit fi fi n shift n stalelv Mest namelist stalelv J We have build our namelist containing the single argument stalelv The next section is a switch to the appropriate synchronization function for the type of LVM object we are targeting if n if sycronizing volume group s then syncvolgrp syncronize the volume group s elif n SLFLAG if syncronizing logical volume s then synclogvol syncronize the logical vo
137. LPs 20 PPs 60 STALE PPs 0 BB POLICY relocatable INTER POLICY minimum RELOCATABLE yes INTRA POLICY middle UPPER BOUND 32 MOUNT POINT N A LABEL None MIRROR WRITE CONSISTENCY on EACH LP COPY ON A SEPARATE PV yes lslv testlv LOGICAL VOLUME testlv VOLUME GROUP testvg LV IDENTIFIER 00041631d7475818 1 PERMISSION read write VG STATE active complete V STATE closed syncd TYPE jfs WRITE VERIFY off MAX LPs 512 PP SIZE 8 megabyte s COPIES 1 SCHED POLICY striped LPs 21 PPs 21 STALE PPs 0 BB POLICY relocatable INTER POLICY maximum RELOCATABLE no INTRA POLICY middle UPPER BOUND 3 MOUNT POINT home mirror LABEL None MIRROR WRITE CONSISTENCY on EACH LP COPY ON A SEPARATE PV yes STRIPE WIDTH 3 STRIPE SIZE 128K lslv m testlv testlv N A LP PP1 PV1 PP2 PV2 PP3 PV3 0001 0056 hdisk3 0028 hdisk2 0002 0028 hdisk1 0028 hdisk4 0003 0057 hdisk3 0029 hdisk2 0004 0029 hdisk1 0029 hdisk4 0005 0058 hdisk3 0030 hdisk2 0006 0030 hdisk1 0030 hdisk4 0007 0059 hdisk3 0031 hdisk2 0008 0031 hdisk1 0031 hdisk4 0009 0060 hdisk3 0032 hdisk2 0010 0032 hdisk1 0032 hdisk4 0011 0061 hdisk3 0033 hdisk2 0012 0033 hdisk1 0033 hdisk4 0013 0062 hdisk3 0034 hdisk2 0014 0034 hdisk1 0034 hdisk4 0015 0063 hdisk3 0035 hdisk2 0016 0035 hdisk1 0035 hdisk4 0017 0064 hdisk3 0036 hdisk2 0018 0036 hdisk1 0036 hdisk4 0019 0065 hdisk3 0037 hdisk2 0020 0037 hdisk1 0037 hdisk4 SD Chapter 3 Replacing a drive in a mirrored configuration 195
138. M corruption 66 A very simple example will illustrate some initial problem determination Note A number of recovery examples are presented in this chapter These are not recipes to be followed blindly Many LVM problems are unique The aim is to develop strategies and techniques for solving LVM problems Suppose we have a system comprising rootvg and datavg However when lsvg is run svg rootvg datavg badvg we seem to have acquired an extra volume group badvg We could try an exportvg badvg exportvg badvg 0516 306 getlvodm Unable to find badvg in the Device Configuration Database 0516 772 exportvg Unable to export volume group badvg Matters have not improved as a second 1svg would confirm We think exportvg only references the ODM We also think 1svg only references the ODM In fact both of these are true at least for 1svg run without any flags However faced with a situation like this we may feel some doubt as to what is actually going on particularly when faced with the more complex scripts such as mirrorvg At this point some people might try a synclvodm often used as a general way of fixing LVM problems For the record in this situation it would fail synclvodm v badvg AIX Logical Volume Manager from A to Z Troubleshooting and Commands 0516 306 synclvodm Unable to find volume group badvg in the Device Configuration Database 0516 502 synclvodm Unable to a
139. Manager from A to Z Troubleshooting and Commands smit reducevg After entering volume group name fz N Remove a Physical Volume from a Volume Group Type or select values in entry fields Press Enter AFTER making all desired changes Entry Field VOLUME GROUP name park vg PHYSICAL VOLUME names FORCE deallocation of all partitions on no this physical volume F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image C F9 Shell FLO Exit Enter Do 1 1 5 Remove a volume group A volume group can be removed from the system by removing all physical volumes from the volume group and then deleting it The reducevg command executed with each physical volume will remove all logical volume information from each physical volume then remove the physical volume from the volume group When the last physical volume is removed For the command line options see reducevg in Appendix A 25 The reducevg command on page 263 or use the SMIT fastpath smit reducevg2 SMIT provides the following fields Volume group name The name of the volume group to be removed F4 for list Chapter 1 LVM commands 9 smit reducevg2 gt N Remove a Volume Group Type or select a value for the entry field Press Enter AFTER making all desired changes Entry Fields VOLUME GROUP name park_vg F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image F9 Shel1 F10 Exit Enter Do
140. ONS read write Enable BAD BLOCK relocation yes Enable WRITE VERIFY no Mirror Write Consistency yes BOTTOM F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image _ F9 Shell F10 Exit Enter Do J Chapter 1 LVM commands 31 1 2 7 Set strictness for a logical volume The strictness policy can be changed for a logical volume but it will not affect the existing physical partitions It will only affect the allocation of new partitions or if the logical volume is reorganized using the reorgvg command For the command line options see chivin Appendix A 1 The chlv command on page 199 or use the SMIT fastpath smit chlv SMIT provides the following fields Logical volume name The name of the logical volume that will have its strictness changed F4 for list Maximum number of physical volumes to use for allocation This physical volume upper bound is used with super strictness to determine how many physical volumes each mirror will span Allocate each logical partition copy on a separate physical volume Sets the strictness for the logical volume to yes Sets a strict allocation policy so that copies of a logical partition cannot share the same physical volume default no Does not set a strict allocation policy 32 AIX Logical Volume Manager from A to Z Troubleshooting and Commands smit chlv N Change a Logical Volume Type or select values in entry fields Pr
141. OR WRITE CONSISTENCY on EACH LP COPY ON A SEPARATE PV yes STRIPE WIDTH 3 STRIPE SIZE 8K F1 Help F2 Refresh F3 Cancel F6 Command F8 Image F9 Shel1 F1O Exit Find n Find Next N S Chapter 1 LVM commands 39 40 Explanation of the fields displayed Logical volume Volume group Logical volume identifier Permission Volume group state Logical volume state Type Write verify Mirror write consistency Max LPs PP size Copies Schedule policy LPs PPs Name of the logical volume Name of the volume group Identifier of the logical volume Access permission read only or read write State of the volume group If the volume group is activated with the varyonvg command the state is either active complete indicating all physical volumes are active or active partial indicating all physical volumes are not active If the volume group is not activated with the varyonvg command the state is inactive State of the logical volume The Opened stale status indicates the logical volume is open but contains physical partitions that are not current Opened syncd indicates the logical volume is open and synchronized Closed indicates the logical volume has not been opened Logical volume type Write verify state on or off Mirror write consistency state yes or no Maximum number of logical partitions the logical volume can hold Size of each physical partition Number of physical partitions created for
142. PhysicalVolume The logical volume can be a name or identifier m Note If the 1slv command cannot find information for a field in the Device Configuration Database it will insert a question mark in the value field As an example if there is no information for the LABEL field the following is displayed LABEL The command attempts to obtain as much information as possible from the description area when it is given a logical volume identifier When no flags are used the following characteristics of the specified logical volume are displayed Logical volume Name of the logical volume Logical volume names must be unique system wide and can range from 1 to 15 characters Volume group Name of the volume group Volume group names must be unique system wide and can range from 1 to 15 characters Logical volume identifier Identifier of the logical volume AIX Logical Volume Manager from A to Z Troubleshooting and Commands Permission Volume group state Logical volume state Type Write verify Mirror write consistency Max LPs PP size Copies Schedule policy LPs PPs Stale partitions BB Policy Inter policy Access permission read only or read write State of the volume group If the volume group is activated with the varyonvg command the state is either active complete indicating all physical volumes are active or active partial indicating all physical volumes are not
143. Refresh F3 Cancel F6 Command F8 Image F9 Shel1 F1O Exit Find n Find Next Ne S AIX Logical Volume Manager from A to Z Troubleshooting and Commands 1 3 2 List configuration of a physical volume The 1spv command displays information about the physical volume Used without any arguments it returns the configuration of the physical volume Other flags can return information about the logical volumes and physical partitions For the command line options see 1spv in Appendix A 10 The Ispv command on page 224 or use the SMIT fastpath smit 1spv SMIT provides the following fields Physical volume name The name of the physical volume to report F4 to list List option Status The configuration of the physical volume Logical volumes List the logical volumes on the physical volume their size and distribution Physical partitions The physical partition usage by logical volume smit lspv List contents of a physical volume N List Contents of a Physical Volume Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields PHYSICAL VOLUME name hdisk9 List OPTION status F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Conmand F7 Edit F8 Image F9 Shel1 F10 Exit Enter Do smit lspv can also be used to show the configuration of a physical volume Chapter 1 LVM commands 45 a N COMVAND STATUS Commend OK
144. T CODE 0 oranch to cleanup with success EXIT COLE 0 We reset SEXIT CODE so no spurious errors appear when cleanup is nm cleanup Now we are inside cleanup we reset the signal trapping ad nna savebase so that the mini AM in the boot logical volume is consistent with the run time ODM Unlike same of the other high level camends this savebase is always run whether ve are in run time or mt there is no test with getlvodm R here trap 01215 n savebase exit 0 J 120 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 2 6 9 Examining raw data on physical volumes and logical volumes Unfortunately an examination of the low level structures comprising the VGDA and VGSA is beyond the scope of this book By this point in the investigation it is to be hoped that a service call has been raised with AlX support this applies for watching system calls with trace and examining internal memory structures as well However we can use od as a simple test for corruption by examining what the raw structures laid out on a disk for the LVM should look like This section will briefly examine the structures mentioned in usr include sys hd_psn h hard disk physical sector numbers define PSN IPL REC 0 PSN of the IPL record define PSN CFG REC 1 PSN of the configuratia record The Mirror Write Consistency MNC records must stay contiguous The MC cache is written to each altemately by the LV
145. The following summarizes the options for the dumpfs command dumpfs Dumps file system information Usage dumpfs FileSystem Device The dumpfs command prints out the superblock i node map and disk map information for the file system or special device specified This listing is used to find out file system information Primarily the dumpfs command is for debugging purposes Appendix D Other related commands 343 D 10 The ff command The following summarizes the options for the command ff Lists the file names and statistics for a file system Usage ff a Number c Number 1 I m Number n File o Option p Prefix s u V VFSName i Number Number s FileSystem a Number Displays the file if it has been accessed within the number of days specified by the Number parameter c Number Displays the file if its inode has been changed within the number of days specified by the Number parameter i Number Displays the files corresponding to the i node numbers specified by the l Number parameter The i node numbers listed must be separated by a comma l Does not display the i node after each path name l Additionally displays a list of pathnames for files with more than one link m Number Displays the file if it has been modified within the number of days specified by the Number parameter n File Displays the file if it has been modif
146. Usage Ilvaryonvg a VGname V VGmajor g VGID o r n p f t filename a VGname The volume group name of the volume group to be varied on V VGmajor The major number of the volume group to be varied on g VGID The VGID of the volume group to be varied on O Selects if logical volumes in the volume group will be opened If this flag is used then access to the logical volumes in the Volume Group is permitted If omitted then access to the logical volumes is not permitted but queries and other system management functions can be performed on the volume group If selected automatic resynchronization of logical and physical Volumes in the Volume Group containing stale partitions will be performed If selected the volume group will be varied on if a quorum is available even though the names of one or more physical volumes are missing from the input list If selected the volume group will be varied on if a quorum is available even though one or more physical volumes in the volume group is missing or removed If selected the volume group will be varied on even if quorum is not present Caution data consistency may be lost if normal operations are attempted without a quorum t Include tags in the output filename The name of the file that contains the list of physical volumes in the volume group to be varied on This file contains one name per line or the n
147. VGSA LVCB JFS problems e Hardware failures e Special considerations for rootvg RAID and SSA issues e Escalating a call to AlX support The following sections attempt to tackle LVM issues in detail The sections dealing with other subsystems are briefer 2 7 1 ODM corruption The most common form of ODM corruption is the ODM and the LVM control information becoming out of sync In this situation generally it is the control information the VGDAs and VGSAs on the disk surfaces that are correct while the ODM becomes incorrect but we should not assume this is always the case The two commands to correct this situation are redefinevg and synclvodm redefinevg will rebuild volume group and physical volume information in the ODM and should give enough information to vary on It will also restore some knowledge of logical volumes to the system synclvodm will restore the remaining logical volume information from the VGDA and LVCBs so as to allow access to the logical volume groups in the volume group For an example of the distinction between these two commands we will create a volume group targetvg containing logical volumes targetlv and targetlog with a JFS file system targetts Chapter 2 Problem determination and recovery 129 mkvg ft2 y targetvg hdisk6 targetvg crfs l 2 m targetfs a size 8192 g targetvg v jfs Based on the parameters chosen the new targetfs JFS file system is limited to a maximum size of
148. a e mail address usib6fpl ibmmail com Contact information is in the How to Order section at this site http www elink ibmlink ibm com pb1 pb1 1 800 879 2755 1 800 IBM 4YOU Country coordinator phone number is in the How to Order section at this site http www elink ibmlink ibm com pbl pb1 1 800 445 9269 1 403 267 4455 Fax phone number is in the How to Order section at this site http www elink ibmlink ibm com pb1 pb1 This information was current at the time of publication but is continually subject to change The latest information may be found at the Redbooks Web site m IBM Intranet for Employees IBM employees may register for information on workshops residencies and Redbooks by accessing the IBM Intranet Web site at http w3 itso ibm com and clicking the ITSO Mailing List button Look in the Materials repository for workshops presentations papers and Web pages developed and written by the ITSO technical professionals click the Additional Materials button Employees may access MyNews at http w3 ibm com for redbook residency and workshop announcements Copyright IBM Corp 2000 407 IBM Redbooks fax order form Please send me the following Title Order Number Quantity First name Last name Company Address City Postal code Country Telephone number Telefax number VAT number Invoice to cus
149. a member of the system group The unmirrorvg command may take a significant amount of time to complete because of complex error checking and the number of logical volumes to unmirror in a volume group AIX Logical Volume Manager from A to Z Troubleshooting and Commands A 36 The updatelv command The following summarizes the options for the updatelv command updatelv Updates the logical control block and the ODM It is an undocumented command Usage updatelv c D L P R Lyname VGname C Exit script if there is a name conflict D This flag keeps the existing entries in dev L Do a learning importvg P Preserve permissions and ownership of special files R Restore permissions from the VGDA Commands called getlvname getlvodm lvgenminor lvrelminor lquerylv lchangelv putlvodm putlvchb getlvcb and mknod The volume group must be varied on when this command is executed A 37 The updatevg command The following summarizes the options for the updatevg command updatevg synchronizes volume group information in the ODM if the ODM has at least a valid volume group identifier It is an undocumented command usage updatevg VGname Commands called getlvodm getlvcb lqueryvg putlvcb and putlvodm The LVM information and the logical volume control block are both used to do the resynchromization The volume group must be varied when this command is execu
150. a_psn 0 and vgda_psn 1 are pointers to the primary and secondary VGDAs If this disk only has one VGDA the secondary will be used for bad block relocation vgsa_psn 0 and vgsa_psn 1 serve the same purpose for bad block relocation The version field indicating LVM version level may serve to explain the cause when a cross system importvg fails Chapter 2 Problem determination and recovery 123 0010000 D E F E C T o o o o o o o o o o 4445 4645 4354 0000 0000 0000 0000 0000 0010020 o o o o o o o o o o o o o o o o 0000 0000 0000 0000 0000 0000 0000 0000 Figure 14 Contents of the first physical sectors on an hdisk Blocks 8 to 29 The LVM bad block directory The bad block directory structure is described by usr include sys bbdir h This is also backed up in sectors 71 to 92 0106000 L V M O 001 7 R Q207 016 O o o o 5f4c 564d 0001 7037 5251 870e 0000 0000 0106020 o o o dO o o 020 t o o b 2 o o O 210 0000 0000 0000 1074 0000 0832 0000 0088 A 0106040 o o b A o C 3 210 O 0 001 oO o 004 0 026 0000 08c2 0043 3388 0000 0100 0004 0016 0106060 o o o b o o o 200 oO O b e o 006 O O 0000 0008 0000 0080 0000 08ba 0006 0000 0106100 o o o o o o o o o o o o o o o o 0000 0000 0000 0000 0000 0000 0000 0000 0107000 D E F E C T o o o o o o o o o o 4445 4645 4354 0000 0000 0000 0000 0000 0107020 o o o o o o o o
151. active If the volume group is not activated with the varyonvg command the state is inactive State of the logical volume The Opened stale status indicates the logical volume is open but contains physical partitions that are not current Opened syncd indicates the logical volume is open and synchronized Closed indicates the logical volume has not been opened Logical volume type Write verify state of On or Off Mirror write consistency state of Yes or No Maximum number of logical partitions the logical volume can hold Size of each physical partition Number of physical partitions created for each logical partition when allocating Sequential or parallel scheduling policy Number of logical partitions currently in the logical volume Number of physical partitions currently in the logical volume Number of physical partitions in the logical volume that are not current Bad block relocation policy Inter physical allocation policy Appendix A High level LVM commands 223 Relocatable Intra policy Upper bound Mount point Label Indicates whether the partitions can be relocated if a reorganization of partition allocation takes place Intra physical allocation policy If the logical volume is super strict upper bound is the maximum number of disks in a mirror copy File system mount point for the logical volume if applicable Specifies the label field for the logical volume Each LP c
152. al Volume Manager from A to Z Troubleshooting and Commands V VFSName Uses the description of the virtual file system specified by the VFSName variable for the file system instead of using the etc filesystems file to determine the description If the v VfsName flag is not specified on the command line the etc filesystems file is checked and the vfs Attribute of the matching stanza is assumed to be the correct file system type y Assumes a yes response to all questions asked by the fsck command This flag lets the fsck command take any action it considers necessary Use this flag only on severely damaged file systems The dfsck command lets you simultaneously check two file systems on two different drives Use the FlagList1 and FlagList2 parameters to pass flags and parameters for the two sets of file systems For a list of valid flags for FlagList1 and FlagList2 see the above table Use a minus sign to separate the file system groups if you specify flags as part of the arguments The dfsck command permits you to interact with two fsck commands at once To aid in this the dfsck command displays the file system name with each message When responding to a question from the dfsck command prefix your response with a 1 or a 2 to indicate whether the answer refers to the first or second file system group Note Do not use the dfsck command to check the root file system D 9 The dumpfs command
153. al copy including the striped block size and striping width of the source logical volume is created and then the data is copied 2 If you are copying a striped logical volume and you have created the destination logical volume with the mklv command using a different stripe block size and striping width or the destination is not a striped logical volume the new characteristics are maintained and the data is copied from the source logical volume A 5 The exportvg command The following summarizes the options for the exportvg command exportvg Exports the definition of a volume group from a set of physical volumes exportvg VolumeGroup Commands called getlvodm lvaryonvg lqueryvgs lvaryoffvg putlvodm lvrelminor and lvrelmajor The exportvg command removes the definition of the volume group specified by the VolumeGroup parameter from the system Since all system knowledge of the volume group and its contents are removed an exported volume group can no longer be accessed The exportvg command does not modify any user data in the volume group A volume group is a non shared resource within the system It should not be accessed by another processor until it has been explicitly exported from its current processor and imported on another The primary use of the exportvg command coupled with the importvg command is to allow portable volumes to be exchanged between processors Only a complete volume g
154. al functions called and so forth syncvg may be run in three ways Targeting a volume group a logical volume or a physical volume depending on whether the v 1 or p flags are used Each of these will call a different internal korn shell function to do the work syncvolgrp synclogvol or syncphyvol These functions are passed to nameliist a list of LVM objects to synchronize which is then process either lresyncpv or lresynclv respectively We omit the first pass of the shell reading the functions in and proceed directly to main Since we are running with 1 we will only see synclogvol executing There is also a cleanup function read in here a N HHHH main HHHH Synchronizes all stale partitims within either a volure gap wyar ical volume Ivdescript ar physical volure pvrene Trput Commend line optics and argurents snog bil f v 1 p N Qtpt Exror Messages Standard eror tt HH HH HOH PATH usr bin etc usr soin usv ucb usr oin X11 doin usr bin etc usr sbin usr ucb home dug bin u hash getlvodm Iqueryvg Iresynclv lresynopv cat alias t getlvodn Iqueryvg lresynclv lresyngov cat EXIT CIF 1 Initialize exit This will be reset to 0 before EXIT COE 1 exiting aly if syncvg completes ancessfully Trap a exit interrupt break to clean up trap cleanp 01215 trap cdenp 0 1 2 15 x oe We have seen the above startup code befo
155. al partitions to add to the logical volume Commands called allocp getlvodm lextendlv lquerylv lquerypv lqueryvg putlvodm and resynclv Note The e m s and u flags are not valid with a striped logical volume The extendlv command increases the number of logical partitions allocated to the LogicalVolume by allocating the number of additional logical partitions represented by the LPartitions parameter The LogicalVolume parameter can 214 AIX Logical Volume Manager from A to Z Troubleshooting and Commands be a logical volume name or a logical volume ID To limit the allocation to specific physical volumes use the names of one or more physical volumes in the PhysicalVolume parameter otherwise all the physical volumes in a volume group are available for allocating new physical partitions By default the logical volume is expanded using the existing characteristics that are displayed when you use the 1slv command To temporarily override these existing characteristics for the new partitions only choose different values for these characteristics by using the flags The characteristics of the logical volume do not change The default maximum number of partitions for a logical volume is 512 Before extending a logical volume more than 512 logical partitions use the chlv command to increase the default value The default allocation policy is to use a minimum number of physical volumes per logical volume
156. al volume as specified by the system PPnum Physical partition number Physical partition numbers can range from 1 to 1016 LVname Name of the logical volume to which the physical partitions are allocated Logical volume names must be system wide unique names and can range from one to 64 characters LPnum Logical partition number Logical partition numbers can range from 1 to 64 000 Copynum Mirror number PPstate Only the physical partitions on the physical volume that are not current are shown as stale Accesses information from the descriptor area specified by the DescriptorPhysicalVolume variable The information may not be current since the information accessed with the n flag has not been validated for the logical volumes If you do not use the nflag the descriptor area from the physical volume that holds the most validated information is accessed therefore the information displayed is current The volume group need not be active when you use this flag Lists only the active volume groups those that are varied on An active volume group is one that is available for use The lsvg command displays information about volume groups If you use the VolumeGroup parameter only the information for that volume group is displayed If you do not use the VolumeGroup parameter a list of the names of all defined volume groups is displayed When information from the Device Configuration database is unavailable
157. al volumes The rmlv command removes a logical volume The LogicalVolume parameter can be a logical volume name or logical volume ID The logical volume first must be closed If the volume group is varied on in concurrent mode the logical volume must be closed on all the concurrent nodes on which volume group is varied on For example if the logical volume contains a file system it must be unmounted However removing the logical volume does not notify Appendix A High level LVM commands 267 the operating system that the file system residing on it has been destroyed The xmfs command updates the etc filesystems file A 29 The rmlvcopy command The following summarizes the options for the xmlvcopy command rmlvcopy Removes copies from a logical volume Usage rmlvcopy LogicalVolume Copies PhysicalVolume s LogicalVolume The LogicalVolume parameter can be the logical volume name or logical volume ID Copies The copies parameter determines the maximum number of physical partitions that remain PhysicalVolume The PhysicalVolume parameter can be the physical volume name or the physical volume ID Commands called getlvodm putlvodm putlvcb lquerylv allocp and lreducelv The xmlvcopy command removes copies from each logical partition in the LogicalVolume Copies are the physical partitions which in addition to the original physical partition make up a logical partition You can ha
158. ame of the block device and the Directory parameter as the directory on which to mount the file system If you enter the mount command without flags the command displays the following information for the mounted file systems e The node if the mount is remote e The object mounted e The mount point e The virtual file system type e The time mounted e Any mount options If you specify only the Directory parameter the mount command takes it to be the name of the directory or file on which a file system directory or file is usually mounted as defined in the etc filesystems file The mount command looks up the associated device directory or file and mounts it This is the most convenient way of using the mount command because it does not require you to remember what is normally mounted on a directory or file You can also specify only the device In this case the command obtains the mount point from the etc filesystems file The etc filesystems file should include a stanza for each mountable file system directory or file This stanza should specify at least the name of the file system and either the device on which it resides or the directory name If the stanza includes a mount attribute the mount command uses the associated values It recognizes five values for the mount attributes automatic true false removable and read only The mount all command causes all file systems with the mount true attribute to be mounted in t
159. ames can be fed via standard input 306 AIX Logical Volume Manager from A to Z Troubleshooting and Commands After varying a Volume Group on line the user gets permission to do the following 1 If the o flag is not invoked queries and other system management functions can be performed on the Volume Group but opens to the Logical Volumes in the Volume Group is not permitted 2 If the o is invoked opens and thereby access to the Logical Volumes in the Volume Group is permitted The IDs of Physical Volumes comprising the Volume Group are output from this command This output is produced only when the supplied Physical Volume names did not match in number and identity of the Volume Group s Physical Volume s stored in the VGDA B 29 The lvaryoffvg command The following summarizes the options for the lvaryoffvg command Ivaryoffvg An object file command that varies off the volume group Usage lvaryoffvg g VGID f g VGID The VGID of the volume group to be varied off f Vary the volume group off into system management mode B 30 The lvgenmajor command The following summarizes the options for the 1vgenmajor command Ivgenmajor An object file command that gets the next available device major number for a volume group Usage lvgenmajor VGName VGName The name of the volume group for which the major number should be generated If a major number already exists for
160. an AlX Support line call support personnel would typically ask for full odmget CuAt odmget Cubv and odmget CuDvDr output and possibly also the predefined database at this point It would then be a more efficient use of the customer s time for support to browse these off line after all we are notin a data availability situation here Alternatively we may tackle this problem from another angle Review the types of information contained within the ODM specific to volume groups as opposed to information for logical volumes and physical devices N Cuat name rootvg attribute vgserial_iq value 00017d37e1762ac7 type R generic D rep n nls index 637 Cuat name rootvg attribute timestamp value 37cec3d704a471db type R generic DU rep Won nls index 0 name rootvg status 0 chgstatus 1 ddins location parent connwhere PdDvLn logical _volume vgsubclass vgtype CuDvDr resource ddins valuel rootvg value2 10 value3 That CuDvDr resource devno valuel eb value2 ou C value3 rootvg J 70 AIX Logical Volume Manager from A to Z Troubleshooting and Commands CuAt CuAt CuDvDr CuDvDr name rootvg attribute vgserial id value 00017d37e1762ac7 type R generic D rep n nls_index 637 name rootvg attribute timestamp value 37cec3d704a471db type R g
161. ana 381 D231 The tar commandi ns Sse de kee acd ye Gail he ae eich a ede 384 D 32 The umount command 0 eee ee eee 391 Appendix E Scripts used during this residency 393 E trem 2 530532 cose sa a Ge re pe ee ele Se oe a 393 vii viii E 2 dSPMSGQSINdEX ecesna tirait ratat a Ra ase E e Gea db ahaeat des 395 ES CRDVIG Macs tea ist teeta mates a T aah aah Sw hae LE eae NOOR BES 395 EA gatheriMaps 2 cecye sd ardei e ante ee ee he ee eae ee aa 396 ES TIM CIV Shine Bice by ten aes ey we Ne cnet ecard a te hana ee oh ee eee 397 EO MaKe ona ete anata len aie cacy Sah oan ita ee oe Sete at hee Oe ime hese 397 Ese PVSINVG anche see eemhdnaa Ge gin heat ee and srs Seep een Ves EAE a a ara 398 E8 SCFAD ER aise k Mint a Getta tn En atid piss E gaan Colas Gog AAA aetna A 398 Appendix F Special notices 0 0 0 cee eee 401 Appendix G Related publications 04 405 G 1 IBM Redbooks 0 000000 eee eee 405 G 2 IBM Redbooks collections 0 cc eee ees 405 G 3 Other reSourceS 0 0 i d ee eee eee eee 406 G 4 Referenced Web sites 0 0 ccc ee eee 406 How to get IBM Redbooks 0 00 0 eee ee ees 407 IBM Redbooks fax order form 0 0 nananana anaana 408 Glossary js5 sc ecr ea ey See en a ee ei ee ee 409 index sarina crdcene a haw ah r a a a Died te Sea IOS BAL eared ean 411 IBM Redbooks review 0 00 0 cee ee ee eee
162. and The mkvg command creates a new volume group using the physical volume names supplied Since AIX Version 4 1 this command also varies on the volume group using the varyonvg command The exception to this is when the volume group is created as concurrent capable By default this command creates a volume group that is capable of handling up to 255 logical volumes on 32 physical volumes These limits can be extended to 512 logical volumes and 128 disks and the design allows space for 1024 physical volumes The mkvg script checks each physical volume to verify that it is not already part of another volume group If a physical volume is part of a varied on volume group then mkvg exits If it is part of anon varied on volume group the user is warned that data will be lost and is given a chance to back out The script also checks that the selected partition size will not violate the 1016 partition limit In this case either the partition factor can be set or the partition size increased For the command line options see mkvg in Appendix A 21 The mkvg command on page 259 or use the SMIT fastpath smit mkvg Chapter 1 LVM commands 3 4 SMIT provides the following fields Volume group name A system wide unique 1 15 character name that will be assigned to the volume group Physical Partition size The physical partition size for the volume group from 1 to 1024 in powers of 2 Physical volume names The name s of the
163. and allocates the number of logical partitions to the new logical volume If you specify one or more physical volumes with the PhysicalVolume 248 AIX Logical Volume Manager from A to Z Troubleshooting and Commands parameter only those physical volumes are available for allocating physical partitions otherwise all the physical volumes within the volume group are available The default settings provide the most commonly used characteristics but use flags to tailor the logical volume to the requirements of your system Once a logical volume is created its characteristics can be changed with the chlv command The default allocation policy is to use a minimum number of physical volumes per logical volume copy to place the physical partitions belonging to a copy as contiguously as possible and then to place the physical partitions in the desired region specified by the a flag Also by default each copy of a logical partition is placed on a separate physical volume The m flag specifies exact physical partitions to be used when creating the logical volume If the volume group in which the logical volume is being created is in big VG format U G and P flags can be used to set the ownership group and permissions respectively of the special device files Only root user will be able to set these values If the volume group is exported these values can be restored upon import if R flag is specified with the importvg command Physi
164. ange the relocation flag to y default n Prevents the logical volume from being relocated during reorganization S Strict Determines the strict allocation policy Copies of a logical partition can be allocated to share or not to share the same physical volume The Strict variable is one of the following y Sets a strict allocation policy so that copies of a logical partition cannot share the same physical volume default n Does not set a strict allocation policy so that copies of a logical partition can share the same physical volume s Sets a super strict allocation policy so that the partitions allocated for one mirror cannot share a physical volume with other partitions from another mirror Note When changing a non superstrict logical volume to a superstrict logical volume you must use the u flag Appendix A High level LVM commands 201 t Type Sets the logical volume type jfs boot jfslog or paging The maximum size is 31 characters If the logical volume is striped you cannot change type to boot U Userid Specifies user ID for the logical volume special file u Upperbound Sets the maximum number of physical volumes for new allocation The value of the Upperbound variable should be between one and the total number of physical volumes When using striped logical volumes or super strictness the upper bound is the maximum number of physical volumes allowed for each mirror copy v Verify Set
165. another high level command script since high level commands frequently run other high level commands some high level commands even export variables in sub shells running other high level commands The importvg flag parsing loop is shown as follows 112 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Parse command line options set getopt y V fcxL FnR if 0 if there is a syntax error then dspmsg s 1 cmdlvm cat 1205 lvmmsg 1205 n PROG gt amp 2 exit fi yFLAG yWAL VFLAG VVAL fFLAG CFLAG xFLAG LFLAG 1VALl FFLAG nFLAG L DEVICE RFLAG while 1 While there is a command line option do case 1 in y yFLAG n yVAL 2 shift shift vgname V vi V vVAL S2 shift shift major number L LF L IVAL 2 shift shift learn option FLAG f shift force varyon c CFLAG c shift import a concurrent vg x XFLAG x shift f auto on conc vg F FFLAG F shift fast import n nFLAG n shift no auto on after import R RFLAG R shift Restore uid gid perms of lvs and auto_on and quorum esac LL done It is a good idea to become familiar with the FLAG variables uses within a particular script before attempting to tackle the rest of the script These may include undocumented flags as some flags should never be passed
166. ant additional considerations This is discussed Corruption example 3 Low level VGDA corruption on page 146 We will still be able to access our data until something happens that requires accessing the affected VGDA such as bringing the volume group back up after a reboot or a varyoffvg varyonvg combination Note that running shutdown will not perform an exportvg 76 AIX Logical Volume Manager from A to Z Troubleshooting and Commands a 1s 1 workfs total 16 Yw r r 1 root sys 5 Sep 11 20 03 fs ok drwxrwx 2 root system 512 Sep 11 20 03 lost found umount workfs varyoffvg workvg varyonvg workvg PV Status hdisk5 00017d372f1de834 PVACTIVE hdisk6 0000000000000000 INVPVID 00017d372 67da78 NAMIDMTCH varyonvg Volume group workvg is varied on gi J We can still vary on the volume group however when we try to mount the file system to access our data we get mount workfs Unable to read superblock TERMINATED Unable to read superblock TERMINATED mount 0506 318 jfs is not a known vfs type for object dev 1v02 This kind of JFS error can indicate a problem either with the jfslog or with the file system itself We display the logical partition maps for each of the disks in our volume group With some kinds of corruption this will not be possible It is recommended that regular snapshots are made of these maps to assist in potential recovery situatio
167. application for accessing objects stored in ODM The odmcreate command also produces an c file that must be compiled and bound in with the application The c file contains structures and definitions that are used internally by ODM at run time The ClassDescriptionFile parameter specifies an ASCII file that contains descriptions of one or more object classes The general syntax for the ClassDescriptionFile parameter is as follows file classes classes class classes class class head body tail head struct ClassName tail body elements elements elements elements element element char DescriptorName DescriptorSize vchar DescriptorName DescriptorSize binary DescriptorName DescriptorSize short DescriptorName long DescriptorName long64 or int64 or OOM_LONG_LONG DescriptorName method DescriptorName link StdClassName StdClassName ColName DescriptorName The default suffix for a ClassDescriptionFile file is cre If no suffix is specified on the odmcreate command then an cre suffix is appended The file can have C language comments if run with the p flag and can include define and include lines that can be preprocessed if the p flag is used to run the C language preprocessor on the file Appendix C ODM commands 317 m Note ODM data bases are 32 bit data bases The long type when used in the class description file is a 32 bit data item The long64 or int64 type when used in the cl
168. apter 1 LVM commands 19 20 Active PVs VG identifier PP size Total PPs Free PPs Alloc PPs Quorum VGDS Auto on Concurrent Auto Concurrent VG Mode Node ID Active Nodes Max PPs Per PV Max PVs Number of physical volumes that are currently active The volume group identifier Size of each physical partition Total number of physical partitions within the volume group Total number of physical partitions not allocated Number of physical partitions currently allocated to logical volumes Number of physical volumes needed for a majority Number of volume group descriptor areas within the volume group Automatic activation at IPL yes or no States whether or not the volume group is Concurrent Capable or Non Concurrent Capable Applies to AIX Version 4 2 or later States whether you should autovary the Concurrent Capable volume group in concurrent or non concurrent mode For volume groups that are Non Concurrent Capable this value defaults to Disabled Applies to AIX Version 4 2 or later The vary on mode of the volume group Concurrent or Non Concurrent Applies to AIX Version 4 2 or later Node ID of this node if volume group is varied on in concurrent node Node IDs of other concurrent nodes that have this volume group varied on Maximum number of physical partitions per physical volume allowed for this volume group Maximum number of physical volumes allowed in this volu
169. are problem is a vital initial step in problem determination precisely speaking although this chapter refers loosely to problem determination problem determination is the distinction between hardware and software problems and problem source identification is the isolation of a problem to a particular hardware or software component The errorlog s contents can be displayed by running errpt a On the hardware side it alerts system administrators events ranging from complete disk failures bus problems such as SCSI and SSA errors and high levels of bad block relocation Such events may be accompanied by sense data which can be analyzed by hardware support Some sites may wish to replace disks once it is judged that an unacceptable level of bad block relocation has been reached It is not unknown to see system crashes accompanied by large amounts of bad block relocation when a disk is dying On the software side as well as seeing JFS errors for conditions such as full file systems we can benefit from LVM s own error logging The relevant error template names are shown as follows AIX Logical Volume Manager from A to Z Troubleshooting and Commands errpt t grep LVM 03913
170. ass description file is a 64 bit data item The generated files will function the same for both 32 and 64 bit applications C 4 The odmdelete command The following summarizes the options for the odmdelete command odmdelete An object file command that deletes selected objects from a specified object class Usage odmdelete o ObjectClass q Criteria o ObjectClass Specifies the object class to delete data from q Criteria Specifies the criteria used to select objects from the object class C 5 The odmdrop command The following summarizes the options for the odmdrop command odmdrop An object file command that removes an object class Usage odmdrop o ClassName o ClassName Specifies the object class to remove The odmdrop command removes an entire object class and all of its objects No checking is done to see if other object classes are linked to this one C 6 The odmget command The following summarizes the options for the odmget command 318 AIX Logical Volume Manager from A to Z Troubleshooting and Commands odmget An object file command that retrieves objects from the specified object classes into an odmadd input file Usage odmget q Criteria ObjectClass s q Criteria Specifies the search criteria used to select objects from the object class or classes C 7 The odmshow command The following summarizes the options for t
171. ations Syntax chpvid lt pvid gt lt hdiskn gt Flags none Note It is most strongly recommended that the chpvid script is not used as a production tool in a production environment as its effects are undefined IBM will not be held responsible if you use this tool Syntax chpvid new pvid hdisk Flags None Appendix E Scripts used during this residency 395 ee usr bin ksh pvid 1 disk 2 set A a echo pvid awk for f 1 f lt length 0 f f 2 print ibase 16 nobase 8 n toupper substr 0 f 2 FIN be 2 gt dev null usr bin echo o a 0 o a 1 o a 2 o a 3 o a 4 o a 5 o a 6 o a 7 0 0 0 0 0 0 0 0 c _ da bs 1 seek 128 of dev disk E 4 gather_maps gather maps saves the map files for all available logical volumes in all available volume groups on a system The administrator should check that all LVs are in a known good state before gathering the maps and should verify the snapshot afterwards for accuracy Syntax gather_maps Flags none N usr bin ksh m r tmp lweps cld my tmp lwteps tmp lvreps old mkdir tmp lvreps od tmp lvreps for IV in 1svg 1svg il k 1 IV s amp 1 2 forint 1 db NunCopies Islv SiV ank COPTES pint 2 gt Cpyi if eNinCopies gt 1 than COPY1 cgoy fi lslv m SV Jak NR gt 2 printf s e n 3 62 gt SLVSCOPYL if
172. ay to tackle an LVM problem Continuing with the trace of redefinevg the script checks for ODM corruption making sure KEY_VGID is not already known to the ODM by another name 140 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Va if vonane is already in database but used for another VG then fatal error else if voname is in database save the VC s auto vale else set fault rnme any existing VG entry from database in or to build fresh ae store all VG info in catebase including all the associated PV info if z SLFIAG then EXIST re ING VGID getvodm v SVGNAVE 2 gt dev n11 if c eq 3 then VME not found in database EXISTING VQ BME getlvodm t SKEY VGID 2 gt dev mi11 is VGID in database if eq 0 witha different VQWE then dms s 1 ardlvm cat 572 lvmsg 572 gt n SPROG SEXISTING WNME gt amp 2 exit fi elif c eq 0 then fi fi z a SEXIST ING VGID SKEY VGID dgonsg s 1 ardlvm cat 574 lvmmnsg 574 gt n SPROG SVQAVE gt amp 2 if z then ING VGID J need a new msg saying that adm is inconsistent exportvg then importvg dams s 1 mdlw cat 621 lvmsg 621 gt n SPROG SVGAVE gt amp 2 fi exit getlvodn v tarcetvg 2 gt dev ul1 EXISTING WD 3 3 eq3 getlvodm t 0001743770963601 2 gt amp wv all EXISTING VQ AVE 4 3 lt
173. ble in the tmp file to store a boot image This space is needed during both backup and installation To determine the amount of space needed in the tmp file issue the following command bosboot q a d device If you use the x flag with the mksysb command you do not need to run the bosboot command to determine the amount of space needed in the tmp file Calls the mkszfile command with the m flag to generate map files Disables software packing of the files as they are backed up Some tape drives use their own packing or compression algorithms This flag only applies to AIX Version 4 2 or later V Verbose mode Lists files as they are backed up This flag only applies to AIX Version 4 2 or later Specifies to automatically expand the tmp file system if necessary The tmp file system may need to be extended to make room for the boot image when creating a bootable backup to tape This flag only applies to AIX Version 4 2 or later Commands called bosboot mkinstape mkszfile and backup The mksysb command creates a backup of the operating system that is the root volume group You can use this backup to reinstall a system to its original state after it has been corrupted If you create the backup on tape the tape is bootable and includes the installation programs needed to install from the backup The file system image is in backup file format The tape format includes a boot image a bosins
174. block the device configuration database and the volume group descriptor areas on the physical volumes Usage synclvodm v VolumeGroup LogicalVolume s V verbose Commands called getlvodm lqueryvg updatevg and updatelv During normal operations the device configuration database remains consistent with the logical volume manager information in the logical volume control blocks and the volume group descriptor areas on the physical volumes If for some reason the device configuration database is not consistent with Logical Volume Manager information the synclvodm command can be used to resynchronize the database The volume group must be active for the resynchronization to occur see varyonvg If logical volume names are specified only the information related to those logical volumes is updated m Note Do not remove the dev entries for volume groups or logical volumes Do not change the device configuration database entries for volume groups or logical volumes using the object data manager 276 AIX Logical Volume Manager from A to Z Troubleshooting and Commands A 34 The syncvg command The following summarizes the options for the syncvg command syncvg Synchronizes logical volume copies that are not current VGname s Usage syncvg f i H P NumParallelLPs I LVName s p PVName s v f Specifies a good physical copy is chosen and propagated to a
175. by these procedures Once a task is selected a resource menu may be presented showing all resources supported by the task Resource Selection This selection will list the resources in the system that are supported by these procedures Once a resource is selected a task menu will be presented showing all tasks that can be run on the resource s C Fl Help Esc 0 Exit F3 Previous Menu S Move the cursor to the boxed area Task Selection and press return TASKS SELECTION LIST 801004 From the list below select a task by moving the cursor to the task and pressing Enter To list the resources for the task highlighted press List TOP Run Diagnostics Run Error Log Analysis Run Exercisers Display or Change Diagnostic Run Time Options Add Resource to Resource List Backup and Restore Media Certify Media Change Hardware Vital Product Data Configure ISA Adapter Configure Reboot Policy Configure Remote Maintenance Policy MORE 30 Re F1 Help F4 List Esc 0 Exit 182 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Select certify media SSA disk certification is performed from the SSA service aids further down this menu RESOURCE SELECTION LIST N 801006 From the list below select any number of resources by moving the cursor to the resource and pressing Enter To cancel the selection press E
176. cal partitions are numbered starting at the outermost edge with number one Appendix A High level LVM commands 249 m Note 1 Changes made to the logical volume are not reflected in the file systems To change file system characteristics use the chfs command 2 Each logical volume has a control block This logical volume control block is the first few hundred bytes within the logical volume Care has to be taken when reading and writing directly to the logical volume to allow for the control block Logical volume data begins on the second 512 byte block 3 A mirrored or copied logical volume is not supported as the active dump device System dump error messages will not be displayed and any subsequent dumps to a mirrored logical volume will fail 4 When creating a striped logical volume using the s flag you must specify two or more physical volumes or use the u flag 5 When creating a striped logical volume the number of partitions must be an even multiple of the striping width 6 To create a striped logical volume with more than one copy all active nodes should be at least AIX Version 4 3 3 or later when the volume group is in the concurrent mode Note There is an undocumented option to the mklv command that suppresses the display of the warning when you try to create a mirrored and striped logical volume Use this command if you want to automate the creation of mirrored and striped logical volumes wi
177. called lsvg lslv lsjfs df bootinf and bosboot The mkszfile command overwrites an existing image data file with new information The mkszfile command saves the system state for reinstallation on the current system or on another system The information saved includes the following e System installation information e Logical volume information for the root volume group e File system information The saved information allows the bosinstall routine to re create the logical volume information as it existed before the backup The mkszfile command creates the image data file The contents of this file are defined by the system in which the image was created The user can edit the image data file before calling the mksysb command The mksysb command in turn only backs up the file systems specified in the image data file which reflects the requirements of the rootvg file system All the saved information is obtained using AIX list commands The commands are listed in the image data file as comments for the user s reference when editing this file 258 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Files on tape cannot be changed However in order to override the data files on the tape the user can create a diskette with the desired files The mkszfile command checks to be sure there is at least 8 MB of free space available in the tmp file system for the boot image Note Before running
178. ccess volume group badvg Later on the reasons for the failure will become clear It is a bad idea to simply throw commands at LVM problems Reasons for this are also made clear later on We run a quick sh xv exportvg badvg to see what is going on sh xv will actually display the contents of all functions as the shell reads them into memory before they are run Like all the other high level LVM scripts there is also a preamble setting out the command name a brief description of its purpose return codes and requisite external programs To keep this document brief these initial sections are omitted here As an aside there are often considerable numbers of comment lines spread throughout the LVM scripts that improve their readability However the reader should always bear in mind that the code of the scripts itself is the best documentation Chapter 2 Problem determination and recovery 67 var esrin jeta faaw asin arf fob A abin seers PATH usr bin etc usr sbin usr ucb usr bin X11 sbin usr bin etc usr sb in usr ucb usr bin X11 sbin home dugald nome dugald home dugald ham e dugald hore dugald scripts EXIT CODE 1 Initialize exit code This will be reset to 0 before EXTT CODR 1 exiting aly if exportvg completes successfully Trap cn exit interrupt break to clean up trap cleamup 012 15 trap cler 012 15 ERGE basename 0 basename export
179. cified by one or more File parameters to the beginning of the archive Makes separate entries for block files special character files and first in first out FIFO piped processes Normally the tar command will not archive these special files When writing to an archive with the d flag the tar command makes it possible to restore empty directories special files and first in first out FIFO piped processes with the x flag Note Although anyone can archive special files only a user with root user authority can extract them from an archive Appendix D Other related commands 385 Checks the file type before archiving Source Code Control Systems SCCS Revision Control Systems RCS files named core errs a out and files ending in o dot o are not archived f Archive Uses the Archive variable as the archive to be read or written When this flag is not specified the tar command uses a system dependent default file name of the form dev rmt0O If the Archive variable specified is minus sign the tar command writes to standard output or reads from standard input If you write to standard output the c flag must be used Forces the tar command to follow symbolic links as if they were normal files or directories Normally the tar command does not follow symbolic links Ignores header checksum errors The tar command writes a file header containing a checksum for each file in the archive Wh
180. command to a previously defined volume group the QUORUM and AUTO ON values will be reset to volume group default values You should verify the parameters of the newly imported volume group with the 1svg command and change any values with the chvg command A volume group with a mirrored striped logical volume cannot be back ported into a version of AIX older than 4 3 3 For the command line options see importvg in Appendix A 8 The importvg command on page 216 or use the SMIT fastpath smit importvg SMIT provides the following fields Volume group name The system wide unique name to be assigned to the volume group when imported into the system Physical volume name The name hdiskn of any of the physical volumes in the volume group The remaining physical volumes will be accessed using the VGDA information on the given physical volume Volume group major number The next available will be allocated or one can be chosen Concurrent options Only used if the volume group is concurrent and the appropriate software installed Chapter 1 LVM commands 11 smit importvg fe N Import a Volume Group Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields VOLUME GROUP name PHYSICAL VOLUME name B Volume Group MAJOR NUMBER Make this VG Concurrent Capable no Make default varyon of VG Concurrent no F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Co
181. compressing the tar file Gathers dump and unix information The primary dump device is used Attention If bosboot k was used to specify the running kernel to be other than unix the incorrect kernel will be gathered Make sure that unix is or is linked to the kernel in use when the dump was taken Gathers dump and unix information The primary dump device is used Identifies the optional snap command output directory tmp ibmsupt is the default Gathers file system information Includes predefined Object Data Manager ODM files in general information collected with the g flag Gathers the output of the 1slpp hBc command which is required to re create exact operating system environments Writes output to the tmp ibmsupt general lslpp hBc file Also collects general system information and writes the output to the tmp ibmsupt general general snap file Gathers installation debug vital product data VPD information Gathers kernel information Gathers LVM information Gathers programming language information Suppresses the check for free space Gathers Network File System NFS information o OutputDevice Copies the compressed image onto diskette or tape 378 AIX Logical Volume Manager from A to Z Troubleshooting and Commands p Gathers printer information r Removes snap command output from the tmp ibmsupt directory S Includes security
182. copy command The following summarizes the options for the readlvcopy command readivcopy Reads a specific mirror copy of a logical volume s skip S seek Usage readlvcopy d device c copy C copy b n number_of_blocks o outfile d device Logical volume special device file to be read from c copy Requested mirror copy to read from Valid values are 1 2 or 3 for the first second or third copy of the data Data is read even if the logical partition has been marked stale The default is the first copy of the data C copy Requested mirror copy to read from Valid values are 1 2 or 3 for the first second or third copy of the data Stale logical partitions are not read b Read mirror copy marked as online backup n number_of_blocks Number of 128 K blocks to read o outfile Destination file The default is stdout s skip Number of 128 K blocks to skip into device S seek Number of 128 K blocks to seek into outfile 262 AIX Logical Volume Manager from A to Z Troubleshooting and Commands A 24 The redefinevg command The following summarizes the options for the redefinevg command redefinevg Redefines the set of physical volumes of the given volume group in the device configuration database Usage redefinevg d Device i Vgid VolumeGroup d Device The volume group ID Vgid is read from the specified physical volume device Yo
183. cpio command 00 0 cee 331 D 5 The crfs Command samia eee ed ae edd Bae ee i 335 D 6 The defragfs command 0 00 ee eee 338 D 7 Ther df COMMAND sede Ss oa eG ees hea tee Ae an 339 D 8 The dfsck command 0 cc cece ee eee eens 341 D 9 The dumpfs command 0 0 c eee eee 343 D310 The ff command derei wei be Sites ecw bie es eee Tek Wea Certs ed Das 344 D 11 The file command 0 0 0 c eee eee eee ees 345 D 12 The fileplace command 000 eee eee 346 D 13 The fsck command 0 00 eee eens 348 D 14 The fsdb command 0 00 cee eee eens 354 D 15 The imfs command 0 00 cece eee eee 354 D 16 The ipl_varyon command sierras ereke aa es 355 D 17 The istat command anaana naaa aaaea eee eee 355 D 18 The logform command oe ess edass e tees 356 D 19 The logredo command 0 c ee ees 357 D 20 The Isfs command 00 ccc eee teens 357 D 21 The lsps command 0 0 cee eee 358 D 22 The mkfs command 0c eee ee eee ee eee 360 D 23 The mount command 00 0 cc teens 362 D 24 The ncheck command 00 cece eee eee eee 365 D 25 The restore command nnana eens 365 D 26 The rmfs command 0 00 cece ee eee tees 376 D 27 The savebase command 0c cc teens 376 D 28 The snap command 000 ees 377 D 29 The synccommand a A eee 380 D 30 The sysdumpdev command sanaan a
184. d SMIT settings and seven logical partitions in size After these manipulations our maps will look as follows Chapter 2 Problem determination and recovery 163 lsvg M allocpvg more allocpvg hdisk6 1 108 hdisk6 109 allocplv4 1 hdisk 110 allocplv4 2 hdisk 111 allocplv4 3 hdisk 112 allocplv4 4 hdisk6 113 allocplv4 5 hdisk 114 allocplv4 6 hdisk6 115 allocplv4 7 hdisk6 116 118 hdisk 119 allocplv3 1 hdisk 6 120 allocplv3 3 hdisk 121 allocplv3 5 hdisk6 122 537 hdisk7 1 108 hdisk7 109 allocplv2 1 hdisk7 110 allocplv2 2 hdisk7 111 allocplv2 3 hdisk7 112 allocplv2 4 hdisk7 113 allocplv2 5 hdisk7 114 allocplv2 6 hdisk7 115 allocplv2 7 hdisk7 116 allocplv2 8 hdisk7 117 allocplv2 9 hdisk7 118 allocplv2 10 hdisk7 119 allocplv2 11 hdisk7 120 allocplv2 12 hdisk7 121 allocplv2 13 hdisk7 122 allocplv2 14 hdisk7 123 allocplv2 15 hdisk7 124 allocplv3 2 hdisk7 125 allocplv3 4 hdisk7 126 537 Ne We place some dummy data into our logical volumes to recover echo allocplv2_data dd seek 1 of dev allocplv2 0 1 records in 0 1 records out echo allocplv3 data dd seek 1 of dev allocplv3 0 1 records in 0 1 records out echo allocplv4 data dd seek 1 of dev allocplv4 0 1 records in 0 1 records out y We now assume our VGDAs are now destroyed 164 AIX Logical Volume Manager from A to Z Troubleshooting and Commands dd if dev zero c
185. d Subsystems Technical Reference Volume 1 SC23 4163 Recall however that BigVG functionality will not be documented by IBM lquerylv lquerypv lqueryvg The low level VGDA interrogation commands are often the only way short of examination of the raw data on disk to detect subtle corruption of the VGDA structures It is recommended that the reader become very familiar with these commands They are of use in recovering maps checking VGDAs for consistency between disks examining VGDA timestamps and so forth As with high level commands these commands can be run against specific physical volumes as shown in the example above for get1vodm 2 6 8 korn shell debug Before running any LVM command its effects should be understood As was already stated the reader should be comfortable with the material presented in the first volume of this redbook AIX Logical Volume Manager from A to Z Introduction and Concepts SG24 5432 As we have seen it is sometimes useful to go lower and look at the script of a particular high level with sh xvor trclvm Examining scripts at this level of detail is also very useful to build an understanding of LVM In Rebuilding and repair on page 128 we will break down a couple of the high level commands used in the recovery situation As preparation for this a commentary for exportvg follows we saw this may do things that might not be appropriate in some situations in Corruption example 2
186. d in conjunction with the u flag the Level parameter provides a method of maintaining a hierarchy of incremental backups for each file system Specify the u flag and set the Level parameter to n to back up only those files that have been modified since the n 1 level backup Information regarding the date time and level of each incremental backup is written to the etc dumpdates file The possible backup levels are 0 to 9 A level 0 backup archives all files in the file system If the etc dumpdates file contains no backup information for a particular file system specifying any level causes all files in that file system to be archived The FileSystem parameter can specify either the physical device name block or raw name or the name of the directory on which the file system is mounted The default file system is the root file system Users must have read access to the file system device such as dev hd4 or have Backup authorization in order to perform backups by i_node AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note 1 You must first unmount a file system before backing it up by i node If you attempt to back up a mounted file system a warning message is displayed The backup command continues but the created backup may contain inconsistencies because of changes that may have occurred in the file system during the backup operation 2 Backing up file systems by i node truncates the uid or gi
187. d of files having a uid or gid greater than 65535 When restored these files may have different values for the uid and gid attributes To retain the values correctly always back up by name files having a uid or gid greater than 65535 3 You can archive only JFS Journaled File System file systems when backing up by i node Back up any non JFS file systems by file name or by using other archive commands such as the pax tar Or cpio command D 2 The chfs command The following summarizes the options for the chfs command chfs Changes attributes of a file system Usage chfs n NodeName m NewMountPoint u MountGroup A yes no p ro rw t yes no a Attribute Value d Attribute FileSystem a Attribute Value Specifies a virtual file system dependent attribute value pair To specify more than one attribute value pair provide multiple a Attribute Value parameters The following attribute value pairs are specific to the Journaled File System JFS a ag 8 16 32 64 Specifies the allocation group size in megabytes An allocation group is a grouping of inodes and disk blocks similar to BSD cylinder groups The default ag value is 8 This attribute only applies to AIX Version 4 2 or later Appendix D Other related commands 327 a bf true false Specifies a large file enabled file system If you do not need a large file enabled file system set this option to fal
188. d propagate the corruption to all your other VGDAs It is extremely difficult to recover from such a situation It is suggested you work through this example before attempting the procedure If you are in any way unsure of what you are doing do not try this call AlX support instead In this example we create the volume group lowvg and associated logical volumes and file systems using the script supplied in Appendix E 6 maker on page 397 This will create volume group lowvg and logical volumes lowlv1 and lowlv2 containing file systems lowfs1 and lowfs2 respectively lowlv1 is a simple logical volume containing three logical partitions and lowlv2 contains two logical partitions mirrored across two disks In our case we are using hdisk6 and hdisk7 This example should be read very carefully It first shows what not to do A very bad outcome of problem determination is presented from which there really is no recovery apart from the theoretical possibility of using dd to adjust the raw bytes in a VGDA after that the maker script will be re run and the recovery performed correctly To begin the example the maker script is run The LVM is now in the following state AIX Logical Volume Manager from A to Z Troubleshooting and Commands Isvg rootvg lowvg lsvg o lowvg rootvg lsvg l lowg lowvg LV NAM TYPE LPs PPs PVs LV STATE MOUNT POINT lowlv1 jfs open syncd lowfs1 lowlv2 jfs open syncd lowfs2 lowlog jf
189. devicel 7482 placed in the VGD variable for exportvg s use Ruming this command also allows us to to take a quick sanity check for a valid SVGNAME e that is present in the M If the SVGQAME is bad then the following call of the test retum function will fail we know that samrething is wrong getlvodm v victinvg VGID 00017d376b7ee896 test return 0 0 0 We are inside test retum function called by the line above we have passed the test getlvodm d victinvg We now get our target volume group s major mmber fram the COM We Check we aren t trying to export rootvg since this is never pemmissible VOQVATOR 43 2 gt dev oul1 4 eq 10 J Now the latest test is passed the next section of code tries to get the volume group lock and makes a further sanity check We cannot export volume groups that are varied on Chapter 2 Problem determination and recovery 115 A N Get volume group information from object data manager if n VGID then putlvodm k SVGID 2 gt dev null lock the volume group so only we change it now test_retur check for error return LOCKED y Determine if the volume group is varied on If the volume group is varied on then output an error message since a volume group that is varied off can only be exported lqueryvgs gt tmp vginfo 2 gt dev null if 0 then while read VGID_ON MAJOR NUM do if VGID_ON VGID
190. dex 642 CuAt name mirrorlv attribute stripe width value 0 type R generic DU rep y nls_index 1100 CuDv name mirrorlv status 1 chgstatus 1 ddins location parent odmvg connwhere PdDvin logical_volume lvsubclass lvtype Chapter 2 Problem determination and recovery 103 Mirrored logical volume ODM objects CuDep name odmvg dependency mirrorlv CuDvDr resource devno valuel 41 value2 3 value3 mirrorlv 2 6 6 5 Logical volumes striped Striped logical volume ODM objects Striped logical volume ODM objects Cuat name stripelv attribute lvserial_iq value 00017d375251870e 2 type R generic D rep n nls_index 648 Cuat name stripelv attribute inter value x type R generic DU rep nyu nls index 643 Cuat name stripelv attribute relocatable value n type R generic DU rep nyu nls index 644 104 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Striped logical volume ODM objects CuAt name stripelv attribute stripe width value 2 type R generic DU rep y nls_index 1100 CuAt name stripelv attribute strictness value s type R generic DU rep nyu nls index 645 CuaAt name stripelv attribute stripe size va
191. disk has become corrupted we will remove the physical volume from the volume group then re add it thus giving the high level LVM commands the opportunity to rebuild the damaged control data In the following commands we are removing the LVM s control data It must be understood that the following commands will not touch the user data held within the logical partitions reducevg workvg hdisk6 0516 016 ldeletepv Cannot delete physical volume with allocated partitions Use either migratepv to move the partitions or reducevg with the d option to delete the partitions 0516 884 reducevg Unable to remove physical volume hdiske Note that this first attempt at reducing the volume group has failed LVM is still aware that there is data on this physical volume but cannot understand that a considered recovery effort is underway We may force the removal with the a flag Although we are warned that all data on the volume group will be destroyed in fact for our purposes this is better read as all data contained on logical volume lv02 will be made inaccessible IN normal LVM operations the original message of course makes a lot of sense 81 Chapter 2 Problem determination and recovery 82 reducevg d workvg hdisk6 0516 914 mlv Warning all data belonging to logical volume 1v02 on physical volume hdisk6 will be destroyed mlv Do you wish to continue y es n o y usr sbin rmlv 458 test 0403 004 Specify a
192. e Max LVs Maximum number of logical volumes allowed in the volume group LVs Number of logical volumes currently in the volume group Open LVs Number of logical volumes within the volume group that are currently open Total PVs Total number of physical volumes within the volume group Active PVs Number of physical volumes that are currently active 232 AIX Logical Volume Manager from A to Z Troubleshooting and Commands VG identifier PP size Total PPs Free PPs Alloc PPs Quorum VGDS Auto on Concurrent Auto Concurrent VG Mode Node ID Active Nodes Max PPs Per PV Max PVs The volume group identifier Size of each physical partition Total number of physical partitions within the volume group Number of physical partitions not allocated Number of physical partitions currently allocated to logical volumes Number of physical volumes needed for a majority Number of volume group descriptor areas within the volume group Automatic activation at IPL yes or no States whether or not the volume group is Concurrent Capable or Non Concurrent Capable Applies to AIX Version 4 2 or later States whether you should auto vary the Concurrent Capable volume group in concurrent or non concurrent mode For volume groups that are Non Concurrent Capable this value defaults to Disabled Applies to AIX Version 4 2 or later The vary on mode of the volume group Concurrent or Non Concurrent Applies
193. e We have already seen the use of the logform command for example in 2 5 Corruption example 2 PVID corruption on page 75 This command should only be run when the log device is known to be corrupt and the file system cannot be mounted If it is run on the log of a mounted file system undefined behavior may result 2 7 3 4 File system cannot be unmounted When a file system cannot be unmounted it indicates that some process is using the file system The fuser cx command can be used to detect this The x flag is an enhancement that shows executable and loadable objects in addition to the standard fuser output For this functionality to be present in pre AlX4 3 3 the following APARs should be installed AIX Level 4 3 2 IX78523 AIX Logical Volume Manager from A to Z Troubleshooting and Commands AIX Level APAR 4 2 1 IX78941 4 1 5 IX78943 2 7 4 Hardware failures Hardware failures are a common cause of LVM problems In the case of a disk failing completely obviously the data on it cannot be recovered The question becomes how to clean up references to the disk from the ODM and the remaining VGDAs The alternative situation is when a disk is failing but not yet dead and the issue is what are the correct sequence of steps to perform to replace the failing disk These procedures are considered in the AIX Version 4 3 System Management Guide Operating System and Devices C23 4126 in the section titled
194. e etc filesystems file with the stanza for the newly imported logical volume You should manually edit the etc filesystems file to add a new stanza for this logical volume The importvg command changes the name of a logical volume if the name already exists in the system It prints a message and the new name to standard error and updates the etc filesystems file to include the new logical volume name m Note 1 AIX Version 4 changed the behavior of importvg so that as part of the importvg process the volume group is automatically varied on by the system after it is imported However if the volume group is Concurrent Capable or was imported with the c flag then the importvg command prompts you to varyonvg the imported volume group manually 2 A volume group with a mirrored striped logical volume cannot be back ported into a version of AIX older than 4 3 3 A 9 The Islv command The following summarizes the options for the 1slv command Islv Displays information about a logical volume Usage Islv L ll m n PhysicalVolume LogicalVolume Islv Displays information about a logical volume allocation map Appendix A High level LVM commands 219 Usage Islv L n PhysicalVolume p PhysicalVolume LogicalVolume L Specifies no waiting to obtain a lock on the Volume group Note If the volume group is being changed using the L flag gives unreliable date Lists the fo
195. e name of the volume group The savevg command uses this file to create a backup image that can be used by the restvg command to remake the user volume group To create a backup of the operating system to CD please refer to the mkcd command A 32 The splitlvcopy command The following summarizes the options for the splitlvcopy command splitlvcopy Splits copies from one logical volume and creates a new logical volume from them Usage splitlvcopy f y NewLVName Y Prefix LVCopies PhysicalVolume s f Specifies to split open logical volumes without requesting confirmation By default split1lvcopy requests confirmation before splitting an open logical volume This includes open raw logical volumes and logical volumes containing mounted file systems Copies Specifies the maximum number of physical partitions that remain in LogicalVolume after the split LogicalVolume Specifies the logical volume name or logical volume ID to split PhysicalVolume Specifies the physical volume name or the physical volume ID to remove copies from y NewLVName Specifies the name of the new logical volume to move copies to from LogicalVolume 274 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Y Prefix Specifies the Prefix to use instead of the prefix ina system generated name for the new logical volume The prefix must be less than or equal to 13 characters A name cannot begi
196. e physical volume from the volume group If a logical volume spans multiple physical volumes the removal of any of those physical volumes may jeopardize the integrity of the entire logical volume f Removes the requirement for user confirmation when the d flag is used Commands called getlvodm ldeletepv lquerylv lquerypv lreducelv putlvodm and rmlv m Note You can use the reducevg command while the volume group is in concurrent mode However if you run this command while the volume group is in concurrent mode and the end result is the deletion of the volume group then the reducevg command will fail The reducevg command removes one or more physical volumes represented by the PhysicalVolume parameter from the VolumeGroup When you remove all physical volumes in a volume group the volume group is also removed The volume group must be varied on before it can be reduced All logical volumes residing on the physical volumes represented by the PhysicalVolume parameter must be removed with the rmlv command or the a flag before starting the reducevg command AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note Sometimes a disk is removed from the system without first running reducevg VolumeGroup PhysicalVolume The VGDA still has this removed disk in its memory but the PhysicalVolume name no longer exists or has been reassigned To remove references to this missing disk you can s
197. e Manager from A to Z Troubleshooting and Commands Relocate the logical volume during reorganization Logical volume label Max number of partitions Sets whether a reorganization of the volume group will affect this logical volume A one to 127 character label for the logical volume If a jfs is using this logical volume then the mount point will be stored here Sets the maximum number of logical partitions that can be allocated to this logical volume The default is 512 Note the logical volume may never have this many logical partitions Enable bad block relocation Scheduling policy Enable write verify Specifies whether the logical volume manager bad block relocation will be used for this logical volume Sets how mirror reads writes are done parallel writes to logical volume copies performed at once reads from the most accessible copy sequential writes to logical volume copies performed sequentially and waits for completion before proceeding to the next Reads are from the primary copy Sets if each write is verified by a read File containing allocation map Stripe size Specifies the name of the file that contains the allocation map for the logical volume The stripe width 4 16 32 64 128 KB or none Chapter 1 LVM commands 25 smit mklv a Type or select values in entry fields TOP Logical volume NAVE VOLUME GROUP name Number of LOGICAL PARTITIONS PHYSICAL VOLUME names
198. e Performance of the logical volume would deteriorate if allocation went beyond a given size e Other parts of an organization may need to be aware if logical volumes are increased For the command line options see chivin Appendix A 1 The chlv command on page 199 or use the SMIT fastpath smit chlv SMIT provides the following fields Logical volume name The name of the logical volume that will have its maximum size changed F4 for list Maximum number of logical partitions This sets the maximum size in logical partitions that a logical volume can be The default is 512 smit chlv aa N Change a Logical Volume Type or select values in entry fields Press Enter AFTER making all desired changes TOP Entry Fields Logical volume NAME odin 1v Logical volume TYPE jfs POSITION on physical volume middle RANGE of physical volumes minimum MAXIMUM NUMBER of PHYSICAL VOLUMES 32 to use for allocation Allocate each logical partition copy yes on a SEPARATE physical volume RELOCATE the logical volume during yes reorganization Logical volume LABEL home odin MAXIMUM NUMBER of LOGICAL PARTITIONS 512 SCHEDULING POLICY for reading writing parallel logical partition copies PERMISSIONS read write Enable BAD BLOCK relocation yes Enable WRITE VERIFY no Mirror Write Consistency yes BOTTOM F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image _ F9 Shell F10 Exit Ente
199. e Volume Group Descriptor Area s SizePP The size of the physical partitions in this volume group This size is expressed as an integer n between 20 and 30 The partition size is 2 bytes Use the force flag to create a volume group on a physical volume when the physical volume appears to be a member of another volume group that is not varied on Caution All data on the physical volume is destroyed Include VGID in the output B 14 The Ideletelyv command The following summarizes the options for the 1deletelv command Appendix B Intermediate level commands 297 Ideletelv An object file command that deletes a logical volume from its parent volume group Usage Ideletelv I LVID I LVID The logical volume ID B 15 The Ideletepv command The following summarizes the options for the 1deletepv command Ideletepv An object file command that deletes a physical volume from its parent volume group Usage Ideletepv g VGID p PVID g VGID The volume group ID p PVID The physical volume ID B 16 The lextendlv command The following summarizes the options for the lextendlv command lextendlv An object file command that extends or allocates additional partitions to a logical volume Usage lextendlv I LVID s ExtendSize filename I LVID The logical volume ID of the logical volume to extend s ExtendSize The number of logical partition
200. e command determines the current type of boot by accessing the boot_mask attribute in the CuAt object class The value of this attribute is the bit mask to apply to the PdDv base field to determine which devices are base Note Base devices are those devices that get configured during phase 1 boot they may vary depending on the type of boot mask For example if the mask is NETWORK_BOOT network devices are considered base For DISK_BOOT disk devices are considered base The type of boot masks are defined in the usr include sys cfgdb h file D 28 The snap command The following summarizes the options for the snap command snap Gathers system configuration information Usage snap a A b c D f g G i k L n N p r s S t o OutputDevice d Dir v Component A Gathers asynchronous TTY information a Gathers all system configuration information This option requires approximately 8 MB of temporary disk space b Gathers SSA information Appendix D Other related commands 377 Creates a compressed tar image snap tar Z file of all files in the tmp ibmsupt directory tree or other named output directory Note Information not gathered with this option should be copied to the snap directory tree before using the c flag If a test case is needed to demonstrate the system problem copy the test case to the tmp ibmsupt testcase directory before
201. e command that is used by other Ivm commands to generate error messages Usage lv mmsg MessageNum MessageNum The LVM message number B 35 The lvrelmajor command The following summarizes the options for the lvrelmajor command Ivrelmajor An object file command that releases a volume group s major number Usage lvrelmajor VGName VGName The name of the volume group for which the major number will be released from the CuDvDr class B 36 The Ivrelminor command The following summarizes the options for the 1vrelminor command Ivrelminor An object file command that releases a logical volume s minor number Usage Ivrelminor DeviceName DeviceName The device for which the minor number will be released All entries in dev associated with this minor number will also be deleted B 37 The migfix command The following summarizes the options for the migfix command Appendix B Intermediate level commands 309 physical partitions migfix migfix is an object file command that is used by the reorgvg command to help determine the proper order of physical partition moves It builds a set of migration moves that will move all the source physical partitions to their destinations without over writing any Usage migfix FreePPList MigratePPList MigrateMoves FreePPList The list of free physical partitions to use in the migration MigratePPList The lis
202. e created on another system The images remain in the directory marked by the r flag or in the mkcd cd_images directory if the 1 flag is not used If multiple volumes are required the final images are uniquely named using the process ID and volume suffixes s savevg_image Indicates a previously created savevg image See the Notes section u bosinst data Specifies the user supplied bosinst data file This data file takes precedence over the bosinst data file in the mksysb image If you do not give the u flag then mkcd restores bosinst data from the given mksysb image or generates a new bosinst data file during the creation of mksysb Appendix A High level LVM commands 241 V cdfs_vol_group Indicates the volume group used when creating the file systems needed for the mkcd command If the v flag is not given and a file system is needed but not there because it was not supplied with other flags then rootvg is the default volume group for creating the file systems If mkcd creates the file systems in the backup volume group those file systems are not included as part of the backup image mkcd created file systems are removed upon the command s completion v savevg_vol_group Denotes the volume group to be backed up using the savevg command See the Notes section See the m flag for more information about where the savevg image is placed z custom_file States the full pathname of
203. e current since the information accessed with the n flag has not been validated for the logical volumes If you do not use the nflag the descriptor area from the physical volume that holds the validated information is accessed therefore the information displayed is current The volume group need not be active when using this flag 226 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Lists the following fields for each physical partition on the physical volume Range A range of consecutive physical partitions contained on a single region of the physical volume State The current state of the physical partitions free used stale or vgda Note If a volume group is converted to a big VG format it may be necessary to use some data partitions for volume group descriptor area These partitions will be marked vgda Region The intra physical volume region in which the partitions are located LVname The name of the logical volume to which the physical partitions are allocated Type The type of the logical volume to which the partitions are allocated Mount point File system mount point for the logical volume if applicable Accesses information based on the VGID variable This flag is needed only when the 1spv command does not function due to incorrect information in the Device Configuration Database The VolumeGroupID variable is the hexadecimal representation of the volume group ident
204. e device name Size Size of the copy that is the smaller logical volume B 5 The getlvcb command The following summarizes the options for the getlvcb command getlvcb This undocumented command queries the lvcb Usage getlvcb A a c f L r P S s t u LYName A Display contents of lvcb a Intra physical volume allocation policy C copies Appendix B Intermediate level commands 289 f File system name L Label r Inter physical volume allocation policy P Stripe expn S Striping width S Strictness t Display type of file system u Upper limit LVName Logical Volume Name B 6 The getlvname command The following summarizes the options for the getlvname command getivname Is an object file command that generates or checks a logical volume name This command is undocumented Usage getlvname Y Prefix n LVName Type n LVName Logical volume name to check Y Prefix Prefix for new logical volume name Type Type of logical volume The getlvname command generates a logical volume name for the logical volume block device If the type argument is contained in the Configuration Database PdAt then the corresponding name prefix is taken from the PdAt class and used to build the logical volume name If the type is not contained in the PdAt class then the default prefix Iv is
205. e mutually exclusive 346 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Displays more information about the file and its placement including statistics on how widely the file is spread across the volume and the degree of fragmentation in the volume The statistics are expressed in terms of either the logical or physical volume fragment numbers depending on whether the 1 or p flag is specified File space efficiency is calculated as the number of non null fragments N divided by the range of fragments R assigned to the file and multiplied by 100 or N R x 100 Range is calculated as the highest assigned address minus the lowest assigned address plus 1 or MaxBlk MinBlk 1 For example the logical blocks written for the file are 01550 through 01557 so N equals 8 The range R 01557 01550 1 also equals 8 Space efficiency for this file is 100 or 8 8 x 100 The v flag message prints the results of the N R 100 equation According to this method of calculating efficiency files greater than 32 KB are never 100 percent efficient because of their use of the indirect block Sequential efficiency is defined as 1 minus the number of gaps nG divided by number of possible gaps nPG or 1 nG nPG The number of possible gaps equals N minus 1 nPGEN 1 If the file is written to 9 blocks greater than 32 KB and the logical fragment column shows 01550 01557 01600 the file is stored in two
206. e name raw device name or file system name If the parameter specifies a file system name the mkfs command uses this name to obtain the following parameters from the applicable stanza in the etc filesystems file unless these parameters are entered with the mkfs command dev vol size boot vfs options Device name Volume ID File system size Program to be installed in the startup block Definition of the virtual file system File system implementation specific options of the form Keyword Keyword Value Appendix D Other related commands 361 m Note 1 The file system is created with the setgid set group ID bit enabled The setgid bit determines the default group permissions All directories created under the new file system have the same default group permissions 2 The mkfs command does not alter anything in a mounted file system including the file system label The file system label changes when you change the mount point unless the file system is mounted Prototype files The mkfs command requires an extended prototype file to create a journaled file system JFS A prototype file is a formatted listing of the contents and structure of a file system A prototype file describes the file system by a series of tokens separated by spaces and new lines The main body of a prototype file defines the objects of the file system A JFS prototype file consists of the main body which can be created by the prot
207. e new copies If not specified physical volumes in the volume group will be used that best meet the allocation policies of the logical volume F4 for list Chapter1 LVMcommands 29 Position on physical volumes Set the intra physical volume allocation policy for this copy copies Range of physical volumes Set the inter physical volume allocation policy for this copy copies Allocate each copy on a separate physical volume Set the strictness policy for this copy copies File containing the allocation map Use a map file for allocating the new logical partitions Synchronize the data in the new logical partition copies This will force the synchronization of the data between the existing logical partitions and the new logical partitions smit mklvcopy N Add Copies to a Logical Volume Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields LOGICAL VOLUME name loki 1v NEW TOTAL number of logical partition 1 copies PHYSICAL VOLUME names POSITION on physical volume middle RANGE of physical volumes minimum MAXIMUM NUMBER of PHYSICAL VOLUMES 32 to use for allocation Allocate each logical partition copy yes on a SEPARATE physical volume File containing ALLOCATION MAP SYNCHRONIZE the data in the new no logical partition copies F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset Fe6 Command F7 Edit F8 Image F9 Shell1 F10 Exit Enter Do 1 2 6 Se
208. e system makes the file system available for use at its defined mount point The mount point becomes the root directory of the newly mounted file system For the command line options see mount in Appendix D 23 The mount command on page 362 or use the SMIT fastpath smit mountfs SMIT provides the following fields File system name The current defined mount point Directory over which to mount If you wish to mount the file system over another directory Type of file system Can be cdrfs jfs nfs sfs nfs3 cachefs or autofs Chapter 1 LVM commands 55 Force the mount Requests a forced mount during system initialization to enable mounting over the root file system Remote node containing the file system to mount If the file system exists on a remote system Mount as a removable file system Treat the file system as a removable file system While files are open it will be treated the same as a normal file system however once all files are closed the system will not return errors if it is removed Mount as read only Set permissions to read only not read write Disallow device access via this mount Set the nodev security option Disallow execution of setuid and setgid programmes in this system Set the nosuid security option smit mountfs a N Mount a File System Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields FILE SYSTEM name DIRECTORY over
209. e track This option applies to by name backups only See the format command for information on sectors per diskette track 0 Creates a Version 2 compatible backup by name This flag is required for compatibility with Version 2 systems because backups by name that are created by a version higher than 2 cannot be restored on Version 2 systems To create a Version 2 compatible backup by name use the o flag along with other flags required for backups by name Files with attributes and values such as user IDs and group IDs that are too large for Version 2 systems will not be backed up A message is displayed for each such file and each value that is too large P Specifies that the files be packed or compressed before they are archived Only files of less than 2 GB are packed Note This option should only be used when backing up files from an inactive file system Modifying a file when a backup is in progress may result in corruption of the backup and an inability to recover the data When backing up to a tape device which performs compression this option can be omitted q Indicates that the removable medium is ready to use When you specify the q flag the backup command proceeds without prompting you to prepare the backup medium and press the Enter key to continue This option applies only to the first volume you are prompted for subsequent volumes The qflag applies only to backups by name Appendix D
210. eSystem n Node a Unmounts all mounted file systems all Unmounts all mounted file systems Appendix D Other related commands 391 allr Unmounts all remotely mounted file systems Note For remote mounts specify the device directory file or file system parameters If you specify the alir flag the umount command unmounts all remote mounts f Forces an unmount in a remote environment Use to free a client when the server is down and server path names cannot be resolved The f flag is not supported for journaled file systems Note This forced unmount works only when there is no hold on the directory that is when the directory is not some user s working directory and there is no process started and still running there n Node Specifies the node holding the mounted directory you want to unmount The umount n Node command unmounts all remote mounts made from the Node parameter t Type Unmounts all stanzas in the etc filesystems file that contain the type Type flag and are mounted The Type parameter is a string value such as the remote value that specifies the name of the group Note You cannot use the umount command on a device in use A device is in use if any file is open for any reason or if a user s current directory is on that device 392 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Appendix E Scripts used during this resid
211. eated earlier becomes of use We re create 1v02 On hdiske using the mapfile so that the logical volume resides on exactly the same partitions as before Note that we have to specify the size of our logical volume the number of logical partitions even though this is implied by the mapfile mklv y 1v02 m map 1v02 workvg 4 1v02 1svg 1 workvg workvg LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT loglv00 jfslog 1 1 1 closed syned N A 1v02 jfs 4 4 1 closed syncd N A We cannot mount the file system just yet Remember that the original entry in etc filesystems was destroyed when the exportvg was run We need to re add this entry with vi to etc filesystems workfs dev dev 1v02 vfs jfs log loglv00 account false mount false Figure 9 Example etc filesystems entry This should be sufficient to get our data back online However as a matter of best practice we should run fsck y on the file systems before mounting it File system corruption is a cause of system crashes and since mount will access the file system we expose ourselves here It is a good idea to guarantee integrity in any case when there may have been a corruption window when access to the file system broke down Chapter 2 Problem determination and recovery 83 84 k p kk kk kk kk kk H u g g d g kk 1s f scl fs_ok nase log redo nase nase nase nase nase nase k
212. ect otherwise the logical volume is ignored Note 1 The reorgvg command does not reorganize the placement of allocated physical partitions for any striped logical volumes 2 At least one free physical partition must exist on the specified volume group for the reorgvg command to run successfully 3 In AIX Version 4 2 or later if you enter the reorgvg command with the volume group name and no other arguments the entire volume group is reorganized In earlier versions of AIX if you enter the reorgvg command with the volume group name and no other arguments it will only reorganize the first logical volume in the volume group The first logical volume is the one listed by the 1svg 1 VolumeName command A 27 The replacepvy command The following summarizes the options for the replacepv command replacepv Replaces a physical volume in a volume group with another physical volume Usage replacepv f SourcePV SourcePVID DestinationPV Usage replacepv R dir_name DestinationPV f Forces the replacement of the DestinationPV by the SourcePV unless the DestinationPV is part of another volume group in the Device Configuration Database ora volume group that is active R dir_name Recovers replacepv if it is interrupted by lt ctrl c gt a system crash or a loss of quorum When using the xr flag you must specify the directory name given during the initial run of replacepv This flag also allows
213. ed within the volume group the volume group back reference to the PdDv database and the ddins and devno driver entries We have not yet pulled in the information from the LVCBs This is why we still see question marks in the 1svg 1 type fields At this stage commands such as extendlv will fail We do not have any policy information in the ODM extendlv targetlv 1 0516 306 getlvodm Unable to find targetlv in the Device Configuration Database 0516 788 extendlv Unable to extend logical volume 134 AIX Logical Volume Manager from A to Z Troubleshooting and Commands To complete the repair run synclvodm This will restore the LVCB control information to the ODM the command can also perform other tidying up such as removing stale ODM information and running imfs x for logical volumes that are not in the VGDA and thus must have been deleted At this point all ODM information is restored synclvodm targetvg lsvg l targetvg targetvg LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT targetlog jfslog 2 2 1 open syncd N A targetlv jfs 1 1 1 open syncd targetfs savebase Note that we run savebase at this point It is useful to look at the commands the redefinevg actually ran when it succeeded Hopefully this will prepare the reader for cases when it fails The trclvm output for redefinevg d hdisk targetvg follows Chapter 2 Problem determination and recovery 135 bin
214. egin checking and input validation Chapter 2 Problem determination and recovery 137 4 Read the VGID from the key devices descriptor area if n SdFLAG then KEY VGID lqueryvg p dVAL v key device defines the VG if 0 then dspmsg s 1 cmdlvm cat 562 lvmmsg 562 n SPROG SdVAL gt amp 2 exit fi elif n SiFLAG then KEY_VGID iVAL else dspmsg s 1 cmdlvm cat 560 lvmmsg 560 gt n SPROG gt amp 2 exit fi n d lqueryvg p hdisk6 v KEY_VGID 00017d3770963601 st 0 0 J At this point in the code we will be in the same state whether we ran with a Or i KEY_VGID is set to 0001743770963601 With parsing and checking of input arguments done we begin to gather data for the reconstruction process 4 Get the list of all configured physical volumes in the database PVS getlvodm C getlvodm C PVS hdisk0 hdisk1 snipped hdisk11 if test z PVS couldn t find any configured PVs then dspmsg s 1 cmdlvm cat 566 lvmmsg 566 n PROG gt amp 2 exit fi test z hdisko hdisk1 snipped hdisk11 J This has just given us a list of all available physical volumes in PVS A sanity check has also been taken 138 AIX Logical Volume Manager from A to Z Troubleshooting and Commands if n SLELAG o n SFFLAG then the shell reads in a lang section of code here which will not actually be executed because we are not runni
215. eisas r Kas rer cee eee 233 A 13 The Ivedit command 000 eee 234 A 14 The migratepv command 00 ee 235 A 15 The mirrorvg command evers eee eee 236 A 16 The mkcd command 00 eee eee 238 A 17 The mklv command 0000 cc eee 243 A 18 The mklvcopy command 00 eee tee 250 A 19 The mksysb command 00 c eee eee 254 A 20 The mkszfile command 0 000 c ee eee 257 A 21 The mkvg command 000 r E a ee 259 A 22 The mkvgdata command eass reier esere aw a eee 261 A 23 The readilvcopy command 0 eee eee 262 A 24 The redefinevg command 00 eee eee 263 A 25 The reducevg command 0 0c eee eee eee 263 A 26 The reorgvg command ee ee 265 A 27 The replacepv command 0 ees 266 A 28 The rmlvcommand 0 eee ee 267 A 29 The rmilvcopy command saaana aaa 268 A 30 The restvg command ssassn aaraa anaana 268 A 31 The savevg command 000 eee 270 A 32 The splitlvcopy command 0 ce es 274 A 33 The synclvodm command 00 cece eee 276 A 34 The syncvg command 0 ee ees 277 A 35 The unmirrorvg command 0 00 eee 277 A 36 The updatelv command 0 0 eee eee 279 A 37 The updatevg command 0 0 ce es 279 A 38 The varyoffvg command 00 eee ee 280 A 39 The varyonvg command sssaaa neeaae 280 Appendix B Intermed
216. em yes File system accounting is to be processed by the accounting subsystem no File system accounting is not to be processed by the accounting subsystem this is the default u MountGroup Specifies the mount group Mount groups are used to group related mounts so that they can be mounted as one instead of mounting each individually For example if several scratch file systems always need to be mounted together when performing certain tests they can each be placed in the test mount group They can then all be mounted with a single command such as the mount t test command v VfsType Specifies the virtual file system type The crfs command creates a file system on a logical volume within a previously created volume group A new logical volume is created for the file system unless the name of an existing logical volume is specified using the d flag An entry for the file system is put into the etc filesystems file Appendix D Other related commands 337 Note The file system is created with the setgid set group ID bit enabled This determines the default group permissions All directories created under the new file system will have the same default group permissions The volume group in which the file system resides defines a maximum logical volume size and also limits the file system size Only journaled file systems created with the default ag bf compress frag and nbpi values and a
217. em types It is displayed in addition to other file system characteristics reported by the 1sfs command u MountGroup Reports on all file systems of a specified mount group v VFSType Reports on all file systems of a specified type The 1sfs command displays characteristics of file systems such as mount points automatic mounts permissions and file system size The FileSystem parameter reports on a specific file system The following subsets can be queried for a listing of characteristics e All file systems e All file systems of a certain mount group e All file systems of a certain virtual file system type e One or more individual file systems The 1sfs command displays additional journaled file system JFS characteristics if the q flag is specified D 21 The Ilsps command The following summarizes the options for the 1sps command Isps Displays the characteristics of paging spaces 358 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Usage Isps s c I I a t lv nfs PagingSpace a Specifies that the characteristics of all paging spaces are to be given The size is given in megabytes C Specifies that the output should be in colon format The colon format gives the paging space size in physical partitions l Specifies that the output should be in list format S Specifies that the summary characteristics of all paging spaces are to be giv
218. emove the stanzas from etc filesystems that contain information pertaining to the specified object The object is a logical volume D 16 The ipl_varyon command The following summarizes the options for the ipl_varyon command ipl_varyon An object file command that is used to vary on the root volume group during system boot processing Usage ipl_varyon d IPLDevice i v d IPLDevice The name of the device from which to perform the ipl i Inquiry mode skips ipl device processing V Set verbose output D 17 The istat command The following summarizes the options for the istat command istat Examines i node numbers Usage istat FileName i nodeNumber Device The istat command displays the i node information for a particular file You can specify the file either by providing a file or directory name with the FileName parameter or by providing an i node number with the i nodeNumber parameter and a device name with the Device parameter You can specify the Device parameter as either a device name or as a mounted file system name Appendix D Other related commands 355 If you specify the FileName parameter the istat command writes the following information about the file e Device where the file resides e i node number of the file on that device e File type such as normal directory and block device e File access permissions e Name and identification nu
219. ems no except tmp usr OR Unmount all REMOTELY mounted file systems no NAME of file system to unmount H REMOTE NODE containing the file system s to unmount F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image a F9 Shel1 F1O Exit Enter Do J Chapter 1 LVM commands 57 58 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Chapter 2 Problem determination and recovery This chapter sets out techniques for Logical Volume Manager problem determination and discusses some internal details of the high level commands After setting out an underlying methodology for problem determination It will also present some common issues encountered and appropriate solutions or recommendations for these situations There will be considerable reference to the contents of LVM scripts not only to highlight specific details of the scripts operation but also to promote an overall awareness of the structure of LVM It is strongly recommended that the reader work through this chapter as a whole Performing recovery techniques on LVM without understanding what you are doing may well lead to further corruption of LVM control data It is also recommended that the reader works through the examples given on a test system Some LVM recovery situations require a good feel for LVM Unless otherwise stated all the tests in this chapter were performed on non bigVG volume groups as these are weaker than the newer
220. en This information consists of the total paging space in megabytes and the percentage of paging space currently assigned used If the s flag is specified all other flags are ignored Note Setting the environment variable PSALLOC early causes the use of early paging space algorithm In this case the value the s flag specifies is different from the value returned for a single paging space or when using the a flag for all the paging spaces The value the s flag displays is the percentage of paging space allocated reserved whether the paging space has been assigned used or not Therefore the percentage reported by the s flag is usually larger than that reported by the a flag when PSALLOC is set to early t Specifies the characteristics of the paging space One of the following variables is required Iv Specifies that the characteristics of only logical volume paging spaces are to be given nfs Specifies that the characteristics of only NFS paging spaces are to be given The heading of the output will be changed to display the host name of the NFS server and the path name of the file that resides on the server that is being used for NFS paging The 1sps command displays the characteristics of paging spaces such as the paging space name physical volume name volume group name size percentage of the paging space used whether the space is active or inactive and whether the paging space is set to automatic The
221. en this flag is not specified the system verifies the contents of the header blocks by recomputing the checksum and stops with a directory checksum error when a mismatch occurs When this flag is specified the tar command logs the error and then scans forward until it finds a valid header block This permits restoring files from later volumes of a multi volume archive without reading earlier volumes L InputList Writes the files and directories listed in the InputList variable to the archive Directories from the InputList variable are not treated recursively For directories contained in the InputList variable the tar command writes only the directory to the archive not the files and subdirectories rooted in the directory If additional files and directories follow the InputList variable on the command line the contents of the InputList variable are archived after these files and directories These additional files or directories are archived with their default behavior which is to treat them recursively 386 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Writes an error message to standard output for each file with a link count greater than 1 whose corresponding links were not also archived For example if file1 and file2 are hard linked together and only file1 is placed on the archive then the 1 flag will issue an error message Error messages are not displayed if the 1 flag is not specif
222. ency This appendix provides a list of the scripts referred to in Chapter 2 Problem determination and recovery on page 59 All these scripts are intended as examples only and have minimal or no error checking Note The following scripts were produced during this residency as aides to problem determination and for understanding the structure and internals of LVM Please note that they are supplied on an as is basis and have not been submitted to any formal IBM testing procedure IBM can take no responsibility for the effects of the scripts in a customer environment E 1 trclvm trclvm makes a temporary copy of an Ivm high level command and instruments it with set xv lines so that all functions within the script produce debug output Syntax trcelvm 1 lt logfile gt t lt command gt lt parameters gt Flags t do not delete temporary files 1 logfile log to file For example trclvm unmirrorvg rootvg will produce shell debug output that can be captured by the script command Copyright IBM Corp 2000 393 usr bin ksh set getopt tl if 0 then echo Usage trclw t 1 logfile comend optims exit fi tFLAG while 1 do case 1 in t tFLAG t shift 1 1FIAG 1 1FILE 2 shift shift esac doe shift file 1 shift panms wkdir tmp lvmtraces cp whence file tmp Ivmtrace file orig awk v tFLAG StFIAG
223. eneric DU rep Won nls index 0 name rootvg status 0 chgstatus 1 ddins location parent connwhere PdDvin logical _volume vgsubclass vgtype resource ddins valuel rootvg value2 10 value3 resource devno valuel 10 value2 0 value3 rootvg Figure 7 Review of volume group specific information in the ODM We can target these categories Chapter 2 Problem determination and recovery 71 odmget q attribute vgserial_id CUAt CuAt name rootvg attribute vgserial_ id value 00017d37e1762ac7 type pR generic D rep n nls index 637 CuAt name datavg attribute vgserial_id value 00017d371e38 579 type R generic D rep n nls index 637 _ There is nothing unexpected here so we look in CuDv Va odmget q PdDvin logical _volume vgsubclass vgtype CuDv CuDv name rootvg status 0 chgstatus 1 ddins location parent connwhere PdDvin logical _volume vgsubclass vgtype name badvg status 0 chgstatus 1 ddins un location parent connwhere PdDvin logical _volume vgsubclass vgtype name datavg status 1 chgstatus 1 ddins un location parent connwhere C PdDvin logical_volumme vgsubclass vgtype 72 AIX Logical Volume Manager from A to Z Tr
224. enough entries in mapfile to fulfill allocation allocp is picky about the way the map file is organized The partitions must be in order from lower to higher and if multiple drives the drives must appear in the same order from bottom to top B 2 The cfgvg command The following summarizes the options for the cfgvg command cfgvg Is a Bourne shell script command that is called by etc rc to vary on volume groups that have the auto varyon flag set This undocumented script requires no flags Usage cfgvg B 3 The chlvcopy command The following summarizes the options for the chlvcopy command chlvcopy Marks or unmarks mirror copy as read only backup Usage chlvcopy f B s b c copy f P newlvname w LV name b Marks a mirror copy as a split mirror copy C Copy Mirror copy to mark as split mirror copy The allowed values of copy are 1 2 or 3 If this option is not specified the default for copy is the last mirror copy of the logical volume B Unmarks a mirror as split mirror copy It will also attempt to remove the child backup logical volume if one was created with the 1 option Appendix B Intermediate level commands 287 Forces split mirror copy even if there are stale partitions If used with the B option the child backup logical volume if one was created with the 1 option will be removed with the force option newlvname New name
225. ents of a Backup System Backup Manager Backup Up the System List Files in a System Image A Restore Files in a System Image y Files amp Directories Logical Volume Manager Backup a File or Directory Volume Groups Restore a File or Directory Logical Volumes List Contents of a Backup Physical Volumes Paging Space Figure 1 SMIT system storage management menu 1 1 Volume group related commands 2 The commands in this section relate to creating and modifying volume groups These commands can be run from the command line or from the SMIT System Storage management gt Logical Volume Manager gt Volume Groups menu See Figure 2 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 1 1 1 Adda Volume Groups List All Volume Groups List Contents of a Volume Group Remove a Volume Group Activate a Volume Group Change a Volume Group Deactivate a Volume Group Add a Physical Volume to a Volume Group Import a Volume Group Remove a Physical Volume from a Volume Group Export a Volume Group Reorganize a Volume Group Mirror a Volume Group Unmirror a Volume Group Synchronize LVM Mirrors Back Up a Volume Group Remake a Volume Group List Files in a Volume Group Backup Restore Files in a Volume Group Backup Figure 2 SMIT volume group options volume group A volume group is created by using the mkvg comm
226. ep nn C nls index 0 a x usr bin tcbck As usual we finish by resetting the EXIT_CODE and calling cleanup EXIT _CODE 0 EXIT CODE 0 exit SEXIT_CODE exit 0 cleanup trap 01215 test 0 eq 1 getlvodm R 0 eq 0 savebase 1 gt dev null umask 022 exit 0 tet ete te tt Va A final piece of ODM manipulation to consider is that you may rmdev dl hdisks and then run mkdev to bring in specific physical volumes to adjust your hdisk numbering This is particularly useful when making both sides of a Chapter 2 Problem determination and recovery twin tailed configuration appear the same or when working with EMC hardware mirrored disks 2 7 2 LVM control data corruption 146 This section first presents an example of low level VGDA corruption Other VGDA problems are then discussed Corruption example 3 Low level VGDA corruption Fortunately VGDA corruption is very rare Generally you can tell that corruption is present when LVM commands start to dump core an alternative possibility is that you have hit a defect in LVM code Sometimes corruption may have occurred on the VGDA but only becomes apparent when the VGDA is restructured for example chvg Bis run or the factor size changed with t In the following example we will deliberately corrupt and then repair a VGDA In a production environment this procedure must be performed extremely carefully otherwise you coul
227. er is specified the df command displays information for the file system on which the file or directory resides If you do not specify the FileSystem or File parameter the af command displays information for all currently mounted file systems File system statistics are displayed in units of 512 byte blocks by default The df command gets file system space statistics from the statfs system call However specifying the s flag gets the statistics from the virtual file system VFS specific file system helper If you do not specify arguments with the s flag and the helper fails to get the statistics the statfs system call statistics are used Under certain exceptional conditions such as when a file system is being modified while the af command is running the statistics displayed by the df command might not be accurate AIX Logical Volume Manager from A to Z Troubleshooting and Commands m Note Some remote file systems such as the Network File System NFS do not provide all the information that the af command needs The df command prints blanks for statistics that the server does not provide D 8 The dfsck command The following summarizes the options for the dfsck command dfsck Checks and repairs two file systems simultaneously on different drives Usage dfsck FlagList1 FileSystem1 FlagList2 FileSystem2 Where FlagList d BlockNumber f i InodeNumber n o Options p t File V VFSName
228. ersion 4 3 Kernel Extensions and Device Support Programming Concepts SC23 4125 AIX Version 4 3 System Management Guide Operating System and Devices SC23 4126 AIX Version 4 3 Installation Guide SC23 4112 AIX Version 4 3 General Programming Concepts Writing and Debugging Programs SC23 4128 G 4 Referenced Web sites e http www rs6000 ibm com support 406 AIX Logical Volume Manager from A to Z Troubleshooting and Commands How to get IBM Redbooks This section explains how both customers and IBM employees can find out about IBM Redbooks redpieces and CD ROMs A form for ordering books and CD ROMs by fax or e mail is also provided e Redbooks Web Site http www redbooks ibm com Search for view download or order hardcopy CD ROM Redbooks from the Redbooks Web site Also read redpieces and download additional materials code samples or diskette CD ROM images from this Redbooks site Redpieces are Redbooks in progress not all Redbooks become redpieces and sometimes just a few chapters will be published this way The intent is to get the information out much quicker than the formal publishing process allows e E mail Orders Send orders by e mail including information from the IBM Redbooks fax order form to In United States Outside North America e Telephone Orders United States toll free Canada toll free Outside North America e Fax Orders United States toll free Canada Outside North Americ
229. esidency 397 E 7 pvsinvg pvsinvg examines the in kernel volume group structures to determine which physical volumes the logical volume device driver thinks are associated with a particular volume group Sy ntax pvsinvg lt volume group name gt Flags none 1 esefoin cch VG S1 DEVND 1s 1 dev S VG awk print 5 cut d f1 MAJOR 0 echo cb 16 n DEVNO be DASDPIR echo devsw SVMAJOR kdb tail 1lawk print 2 for DEVI in echo volgyp SDASDPIR kdb grep pvole avk print 6 sort u db if SDEVI 00000000 J then HEXMAJ echo SDEVT cut c 1 4 HEXMIN echo SDEVI cut c 5 MAJOR echo ib 16 nSHEXMA bc MINOR echo ib 16 n HEXMIN bc cdmget q valuel SVAJOR and vallue2 SMINOR CuDvDr grep value3 cut d f 2 fi done sort u E 8 scraper This last resort script will search for orphan LVCBs on a disk that has lost its VGDAs It will then offer the option of re creating a contiguous logical volume equal in size to the number of logical partitions held in the LVCB Syntax scraper lt hdisk name gt Flags none 398 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Va usr bin ksh Note this is a very minimal exanple of a scammer repairer it has no understanding of the user data and ly Jays down contiguous maps change the folowing variables for bigVG or nm 4MB PPSIZE let offset 4352 hard coded for a
230. ess Enter AFTER making all desired changes TOP Entry Fields Logical volume NAME odin 1v Logical volume TYPE jfs POSITION on physical volume middle RANGE of physical volumes minimum MAXIMUM NUMBER of PHYSICAL VOLUMES 32 to use for allocation Allocate each logical partition copy yes on a SEPARATE physical volume RELOCATE the logical volume during yes reorganization Logical volume LABEL home odin MAXIMUM NUMBER of LOGICAL PARTITIONS 512 SCHEDULING POLICY for reading writing parallel logical partition copies PERMISSIONS read write Enable BAD BLOCK relocation yes Enable WRITE VERIFY no Mirror Write Consistency yes BOTTOM F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image _ F9 Shell F10 Exit Enter Do J 1 2 8 Set write verify for a logical volume The write verify can be changed for a logical volume and it will take effect immediately Setting the write verify for a logical volume will mean that each write will be followed by a read to verify the data written This obviously has performance implications For the command line options see chlv in Appendix A 1 The chlv command on page 199 or use the SMIT fastpath smit chlv SMIT provides the following fields Logical volume name The name of the logical volume that will have its write verify flag changed F4 for list Enable write verify Set write verify on or off for the logical volume Chapter 1 LVM commands 33 smit c
231. etermines this by inspecting the s_fmod flag in the file system superblock This flag is set whenever a file system is mounted and cleared when it is unmounted successfully If a file system is unmounted successfully it is unlikely to have any problems Because most file systems are unmounted successfully not checking those file systems can reduce the checking time i i Node Number Searches for references to a specified i node Whenever the fsck command encounters a directory reference to a specified i node it displays the full path name of the reference n Assumes no response to all questions asked by the fsck command It does not open the specified file system for writing o Options Passes comma separated options to the fsck command These options are assumed to be file system implementation specific except that the following are currently supported for all file systems mountable Causes the fsck command to exit with success returning a value of 0 if the file system in question is mountable clean If the file system is not mountable the fsck command exits returning with a value of 8 mytype Causes the fsck command to exit with success 0 if the file system in question is of the same type as either specified in the etc filesystems file or by the v flag on the command line Otherwise 8 is returned For example fsck o mytype v jfs exits witha value of 0 if the root file system is a journaled f
232. etion additional instructions may appear below hdisk9 LV NAME LPs PPs DISTRIBUTION MOUNT POINT loki 1v 4 4 00 04 00 00 00 home 1loki odin_lv 12 12 00 12 00 00 00 N A thor 1v 15 30 00 30 00 00 00 home thor F1 Help F2 Refresh F3 Cancel F6 Command F8 Image F9 Shel1 F1O Exit Find n Find Next Ne J Explanation of the fields displayed PP range The physical volume name State The current state of the physical partitions free used stale or VGDA If a volume group is converted to a big VG format it may be necessary to use some data partitions for the volume group descriptor area These partitions will be marked as VGDA Region The intra physical volume region in which the partitions are located LV ID The name of the logical volume to which the physical partitions are allocated Type The type of the logical volume to which the partitions are allocated Mount point The mount point of the file system if appropriate 48 AIX Logical Volume Manager from A to Z Troubleshooting and Commands smit lspv Show physical partition usage for the physical volume a N COMMAND STATUS Command OK stdout yes stderr no Before command completion additional instructions may appear below hdisk9 PP RANGE STATE REGION LV ID TYPE MOUNT POINT 1 108 free outer edge 129 132 used outer middle loki _lv jfs home loki 133 144 used outer middle odin 1v jfs N A 145 174 used outer middle thor 1v jfs home tho
233. evice 21 dumpfs 343 E errorlog 88 errpt 87 88 exportvg 13 66 83 118 211 extendlv 37 212 extendvg 61 215 extra volume group 66 F ff 344 file 345 file system 51 permissions 51 type 51 55 file system corruption 180 fileplace 346 fragment size 51 fsck 83 348 fsdb 354 G getlvcb 87 108 289 getlvname 290 getlvodm 44 65 108 291 getvgname 293 H hardware failure 181 header files 59 411 l imfs 118 135 354 importvg 10 11 62 112 216 ipl_varyon 355 istat 355 J journaled file system 21 K korn shell 59 62 L Ichangelv 293 Ichangepv 295 Ichlvcopy 294 Icreatelv 296 Icreatevg 61 297 Ideletelv 297 Ideletepv 149 298 lextendlv 298 linstallpv 299 Imigratelv 299 Imigratepp 300 log device corruption 180 logform 356 logical partition mapping 38 logical volume allocation characteristics 14 control block 69 label 25 relocatable flag 14 state 21 40 type 21 24 logredo 357 Iquerylv 87 Iquerypv 87 302 Iqueryvg 79 87 125 132 302 Iqueryvgs 303 Iqureyly 300 Ireducelv 151 304 lresynclp 305 Iresyncly 305 lresyncpv 305 Isattr 87 Isdev 87 Isfs 357 Islpp 87 Islv 38 41 219 Isps 358 Ispv 45 47 80 87 96 224 Isvg 14 16 18 21 66 229 Isvgfs 233 lvaryoffvg 307 lvaryonvg 306 LVCB 75 lvchkmajor 308 Ilvedit 234 lvgenmajor 307 lvgenminor 308 lvistmajor 308 LVM control data corruption 129 LVM_MISSPVADDED 91 185 lvmmsg 309 lvrelmajor 309 Ivre
234. f the restore command when the archived files are restored from the tape Backing up files by name To back up by name use the i flag The backup command reads standard input for the names of the files to be backed up File types can be special files regular files or directories When the file type is a directory only the directory is backed up The files under the directory are not backed up unless they are explicitly specified Appendix D Other related commands 325 326 m Note 1 Files are restored using the same path names as the archived files Therefore to create a backup that can be restored from any path use full path names for the files that you want to back up 2 When backing up files that require multiple volumes do not enter the list of file names from the keyboard Instead pipe or redirect the list from a file to the backup command When you enter the file names from the keyboard and the backup process needs a new tape or diskette the command loses any file names already entered but not yet backed up To avoid this problem enter each file name only after the archived message for the previous file has been displayed The archived message consists of the character a followed by the file name 3 If you specify the p flag only files of less than 2 GB are packed Backing up file systems by i node To back up a file system by i node specify the Level and FileSystem parameters When use
235. fault compress value is no Selecting compression requires a fragment size of 2048 or less Appendix D Other related commands 335 a frag 512 1024 2048 4096 Specifies the JFS fragment size in bytes A file system fragment is the smallest unit of disk storage that can be allocated to a file The default fragment size is 4096 bytes a logname LVName Specifies the log logical volume name The specified logical volume will be the logging device for the new JFS The LVName logical volume must already exist The default action is to use an existing logging device in the target volume group a nbpi 512 1024 2048 4096 8192 16384 32768 65536 131072 Specifies the number of bytes per i node nbpi The nbpi affects the total number of i nodes on the file system The nbpi value is inversely proportional to the number of i nodes on the file system The default nbpi value is 4096 bytes The values 32768 65536 and 131072 only apply to AIX Version 4 2 or later a size Value Specifies the size of the JFS in 512 byte blocks If the specified size is not evenly divisible by the physical partition size it is rounded up to the closest number that is evenly divisible This attribute is required when creating a JFS file system See Understanding JFS Size Limitations for more information The maximum size of a JFS file system is a function of its fragment size and the nbpi value These values yield the
236. files in general information collected with the g flag S Gathers Systems Network Architecture SNA information Gathers Transmission Control Protocol Internet Protocol TCP IP information v Component Displays the output of the commands executed by the snap command Use this flag to view the specified name or group of files The snap command gathers system configuration information and compresses the information into a tar file The file can then be downloaded to disk or tape or be transmitted to a remote system The information gathered with the snap command may be required to identify and resolve system problems Use the snap o dev rfdo command to copy the compressed image to diskette Use the snap o dev rmto command to copy the image to tape Approximately 8 MB of temporary disk space is required to collect all system information including contents of the error log If you do not gather all system information with the snap a command less disk space may be required depending on the options selected Note that if you intend to use a tape to send a snap image to IBM for software support the tape must be one of the following formats e 8mm 2 3 Gb capacity e 8mm 5 0 Gb capacity e 4mm 4 0 Gb capacity Using other formats prevents or delays IBM software support from being able to examine the contents The snap g command gathers general system information including the following e Error re
237. flag When specifying path names that are greater than 100 characters for the United States Tape Archiver USTAR format remember that the path name is composed of a prefix buffer a slash and a name buffer The prefix buffer can be a maximum of 155 bytes and the name buffer can hold a maximum of 100 bytes If the path name cannot be split into these two parts by a slash it cannot be archived This limitation is due to the structure of the tar archive headers and must be maintained for compliance with standards and backwards compatibility In addition the length of a destination for a hard or symbolic link the link name cannot exceed 100 bytes When writing to an archive the tar command uses a temporary file the tmp tar file and maintains in memory a table of files with several links You receive an error message if the tar command cannot create the temporary file or if there is not enough memory available to hold the link tables Two groups of flags exist for the tar command The required flags and the optional flags The required flags control the actions of the tar command and include the c r t u and x flags At least one required flag must be selected for the tar command to function Having selected a required flag you can select an optional flag but none are necessary to control the tar command 390 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 1 Note When the storage de
238. following test of Frequent error checking is a feature of the high level commands If the putlvodm had failed dspmsg would have been run This will attempt to get an Chapter 2 Problem determination and recovery 63 appropriate NLS message for the situation but if one is not available in the message catalogues it will fall back onto the default set of English language messages generated by 1vmsg If for some reason the putlvodm had failed the following message would have been printed 0516 624 mkvg Warning cannot update device configuration database for volume group Execute redefinevg to synchronize the database This is not considered a fatal error so we continue execution LVM commands will sometimes output more than one error message when failure occurs Some LVM high level commands call an errhandler function or a similar function serving this purpose For example if we have a 9 1 GB disk with physical partition size of 4 MB there are 2275 physical partitions Running chvg t2 c testvg will ask for 2032 physical partitions Since this is insufficient 0516 1158 chvg The t flag parameter value 2 is too small to accommodate the largest disk in the volume group testvg Specify a value between 3 and 16 OR do not specify any value for default minimum value lvmmsg 1158 0516 732 chvg Unable to change volume group testvg lvmmsg 732 will be displayed This is generated by the following code Va convert1016
239. fragments out of a possible nine fragments The sequential efficiency calculation for this file is nG 1 nPG 9 1 8 1 1 8 x 100 87 5 The fileplace command displays the placement of a specified file within the logical or physical volumes containing the file By default the fileplace command lists to standard output the ranges of logical volume fragments allocated to the specified file The order in which the logical volume fragments are listed corresponds directly to their order in the file A short header indicates the file size in bytes the name of the logical volume in which the file lies the block size in bytes for that volume the fragment size in bytes and the compression indicating if the file system is compressed or not Appendix D Other related commands 347 Occasionally portions of a file may not be mapped to any fragments in the volume These areas whose size is an integral number of fragments are implicitly zero filled by the file system The fileplace command indicates which areas in a file have no allocated fragments Optionally the fileplace command also displays e Statistics indicating the degree to which the file is spread within the volume e The indirect block addresses for the file e The file s placement on physical as opposed to logical volume for each of the physical copies of the file m Note 1 The fileplace command is not able to display the placement of remote Network
240. g and the mkcd cd_fs directory exists mkcd uses that directory If you do not give the c flag and the mkcd cd_fs directory does not exist mkcd Creates the file system mkcd cd_fs and removes it when the command finishes executing The command creates the file system in the volume group indicated with the v flag or rootvg if that flag is not used Each time you invoke the rkcd command a unique subdirectory using the process ID is created under the mkcd cd_fs directory or in the directory specified with the c flag D Turns on the debug output information feature The default is no debug output d cd_device Indicates the CD R device dev cd1 for instance This flag is required unless you use the s flag Excludes the files and or directories from the backup image listed in the etc exclude volume_group file You cannot use this flag with the mor s flags Generates a generic bootable mksysb CD The CD contains all three RS 6000 platform chrp rs6k rspc boot images In conjunction with the cflag you must specify the mand p flags Appendix A High level LVM commands 239 cd_image_dir Specifies the directory or file system where the final CD images are stored before writing to the CD R device If this flag is not used mkcd uses the mkcd cd_images directory if it already exists If not the command creates the mkcd cd_images file system in the volume group given with the v flag
241. g mounted file systems this is not recommended unless the application is at a known state at the time the copy is marked as a split mirror The split mirror copy is internally consistent at the time the chlvcopy command is run but consistency is lost between the logical volume and the split mirror copy if the logical AIX Logical Volume Manager from A to Z Troubleshooting and Commands volume is accessed by multiple processes simultaneously and the application is not at a known state When marking an open logical volume data may be lost or corrupted Logical volumes should be closed before marking online split mirror copies in order to avoid a potential corruption window If the persistence flag is not set to prevent the loss of backup data the volume group should be set to not automatically vary on and the nflag should be used with varyonvg to prevent stale partitions from being resynced If the persistence flag P is set the following applies In the event of a crash while an online split mirror copy exists or multiples exist the existence of copies is retained when the system is rebooted B 4 The copyrawlv command The following summarizes the options for the copyrawlv command copyrawlv Is an object file command that is used by cplv to do the actual copying on the physical volume Usage copyrawlv LVNamefrom LVNameto Size LVNamefrom Source logical volume device name LVNameto Destination logical volum
242. g sections should core dump before a low level modification command is run However if you choose to operate at the low level it is assumed that you know what you are doing Remember that the low level commands are unsupported when used directly by customers Chapter 2 Problem determination and recovery 151 We now resume before we run the lreducelv To get back to this situation we can rerun maker and retrace our steps fe N maker hdisk6 hdisk7 output skipped echo c dd bs 1 seek echo ib 8 n200000 11000 be of dev hdisk6 1 0 records in 1 0 records out lsvg n hdisk7 grep ID VOLUME GROUP lowvg VG IDENTIFIER 00017d378099d25d VGID 00017d378099d25a lspv M n hdisk7 hdisk6 grep lowlvl awk print 1 gt lowlv1 map lspv M n hdisk7 hdisk6 grep lowlv2 awk print 1 gt lowlv2 map If you are using the same disks as before you will find that there is a certain predictability to the way allocp lays out partitions on a disk The LV maps should actually be the same as before lquerylv L VGID 1 r gt lowlvl map raw lquerylv L VGID 2 r grep 00017d377c3abace gt lowlv2 map raw Qe J The key point in this procedure is that we must remove hdisk6 from the wheel before it is spun We can do this by removing dev hdisk6 If the LVM device driver cannot read from the disk it cannot propagate it to any other devices on the wheel Since we have to vary off the volume group before we can
243. ge of consecutive physical partitions PPnum is the physical partition number Appendix A High level LVM commands 213 S Strict Determines the strict allocation policy Copies of a logical partition can be allocated to share or not to share the same physical volume The Strict variable is one of the following y Sets a strict allocation policy so that copies of a logical partition cannot share the same physical volume default n Does not set a strict allocation policy so that copies of a logical partition can share the same physical volume S Sets a super strict allocation policy so that the partitions allocated for one mirror cannot share a physical volume with other partitions from another mirror Note When changing a non superstrict logical volume to a superstrict logical volume you must use the u flag u Upperbound Sets the maximum number of physical volumes for new allocation The value of the Upperbound variable should be between one and the total number of physical volumes When using striped logical volumes or super strictness the upper bound is the maximum number of physical volumes allowed for each mirror copy PhysicalVolume s The physical volumes to be used in allocating additional partitions to the logical volume can be restricted to those named If no physical volumes are named then all will be available depending on which best meet the allocation policies LPartitions The number of logic
244. ger from A to Z Troubleshooting and Commands Breaks disk reservations on disks locked as a result of a normal varyonvg command Use this flag on a volume group that is already varied on This flag only applies to AIX Version 4 2 or later Note This flag unlocks all disks in a given volume group Varies the volume group on in concurrent mode This is only possible if the volume group is Concurrent Capable and the system has the HACMP product loaded and available If neither is true the volume group will fail the vary on This flag only applies to AIX Version 4 2 or later If the vary on process detects that there is a new logical volume in the volume group whose name is already being used for one of the existing logical volumes then the vary on will fail You will need to rename the existing logical volume before attempting the vary on again Allows a volume group to be made active that does not currently have a quorum of available disks All disk that cannot be brought to an active state will be putina removed state At least one disk must be available for use in the volume group Disables the synchronization of the stale physical partitions within the VolumeGroup All physical volumes must be available to use the varyonvg command Makes the volume group available in System Management mode only Logical volume commands can operate on the volume group but no logical volumes can be opened for input or o
245. gical partitions to allocate to the logical volume Physical volume names The physical volumes to use to allocate physical partitions If none are specified AIX will use physical volumes that best meet the specified allocation policies Logical volume type The type of logical volume jfs journaled file system jfslog log logical volume for a jfs boot contains AIX boot image paging system paging logical volume Position on physical volume Specifies the intra physical volume allocation policy SMIT uses the different terms to describe the regions on physical volume Range of physical volumes Specifies the inter physical volume allocation policy minimum tries to use minimum number of disks depending on number of mirror copies maximum will spread the logical volume over the available physical volumes Max number of physical volumes used Sets the number of physical volumes to be used in allocation Must be between one and the number of physical volumes in the volume group Number of copies Specifies the number of physical partitions to be allocated for each logical partition Must be between one no mirroring and three Mirror write consistency Whether mirror write consistency is turned on Allocate each logical partition copy on a separate physical volume Sets the strictness policy For availability one usually wants to allocate each copy of a logical volume on different physical volumes AIX Logical Volum
246. h level LVM commands 237 accomplished manually if you execute the mklvcopy command for each individual logical volume in a volume group As with mklvcopy the target physical drives to be mirrored with data must already be members of the volume group To add disks to a volume group run the extendvg command By default mirrorvg attempts to mirror the logical volumes onto any of the disks in a volume group If you wish to control which drives are used for mirroring you must include the list of disks in the input parameters PhysicalVolume Mirror strictness is enforced Additionally mirrorvg mirrors the logical volumes using the default settings of the logical volume being mirrored If you wish to violate mirror strictness or affect the policy by which the mirror is created you must execute the mirroring of all logical volumes manually with the mklvcopy command When mirrorvg is executed the default behavior of the command requires that the synchronization of the mirrors must complete before the command returns to the user If you wish to avoid the delay use the s or s option Additionally the default value of two copies is always used To specify a value other than 2 use the c option Note The mirrorvg command may take a significant amount of time before completing because of complex error checking the amount of logical volumes to mirror in a volume group and the time is takes to synchronize the new mirrored logical volumes
247. has not been applied Once striping is put onto a disk its importation into version 3 2 is prevented When you issue the importvg command to a previously defined volume group the QUORUM and AUTO ON values will be reset to volume group default values You should verify the parameters of the newly imported volume group with the 1svg command and change any values with the chvg command The importvg command makes the previously exported volume group known to the system The PhysicalVolume parameter specifies only one physical volume to identify the volume group any remaining physical volumes those belonging to the same volume group are found by the importvg command and AIX Logical Volume Manager from A to Z Troubleshooting and Commands included in the import An imported volume group is automatically varied unless the volume group is Concurrent Capable You must use the varyonvg command to activate Concurrent Capable volume groups before you access them When a volume group with file systems is imported the etc filesystems file is updated with values for the new logical volumes and mount points After importing the volume group and activating it with the varyonvg command you must run the fsck command before the file systems can be mounted However the mount point information would be missing from the LVCB logical volume control block if it is longer than 128 characters In this case the importvg command will not be able to update th
248. he cpio command The following summarizes the options for the cpio command cpio Copies files into and out of archive storage and directories Usage cpio o a c v B C Value lt FileName gt Output Usage cpio i b c d f m M r s t u v S 6 B C Value Pattern lt Input Usage cpio p a d I m M u v Directory lt FileName Directory Specifies the directory lt Filename Specifies a list of file names for the cpio command to use as input Appendix D Other related commands 331 gt Output Specifies the output device such as a diskette or file For more information on using tape devices see the rmt special file lt Input Specifies the input device where Input is the Output file created by the cpio o command For more information on using tape devices see the rmt special file Pattern Specifies the pattern as described in the ksh command to be used with the command The default for the Pattern parameter is an asterisk selecting all the files in the Input Resets the access times of the source files to their previous times Swaps both bytes and halfwords Note If there is an odd number of bytes or halfwords in the file being processed data can be lost Performs block input and output using 512 bytes toa record Note When using the B or C options to extract or create a tape archive the blocking factor must be
249. he logical volumes listed by Isvg I Volume_Group If a list of logical volumes is provided the highest priority goes to the first logical volume in the list down to the last logical volume with the lowest priority The volume group must be varied on and must have free partitions before use of the reorgvg command The relocatable flag of each logical volume must be set to y with the chlv r command or smit chlv for the reorganization to take effect otherwise the logical volume is ignored 14 AIX Logical Volume Manager from A to Z Troubleshooting and Commands m Note The reorgvg command does not reorganize the placement of allocated physical partitions for any striped logical volumes At least one free physical partition must exist on the specified volume group for the reorgvg command to run successfully In AIX Version 4 2 or later if you enter the reorgvg command with the volume group name and no other arguments the entire volume group is reorganized For lower levels of AIX if you enter the reorgvg command with the volume group name and no other arguments it will only reorganize the first logical volume in the volume group The first logical volume is the one listed by the 1svg 1 VolumeName command This command is not allowed if the volume group is varied on in concurrent mode For the command line options see reorgvg in Appendix A 26 The reorgvg command on page 265 or use the SMIT fastpath smit reo
250. he odmshow command odmshow An object file command that displays an object class definition on the screen Usage odmshow ObjectClass The odmshow command takes as input an object class name ObjectClass and displays the class description on the screen The class description is in the format taken as input to the odmcreate command Appendix C ODM commands 319 320 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Appendix D Other related commands This appendix summarizes the options for commands that are not as closely related to the Logical Volume Manager You will find here the commands to back up your data and to manipulate the journaled file systems or the paging space D 1 The backup command The following summarizes the options for the backup command backup Backs up files and file systems by name Usage backup i b Number p e RegularExpression f Device Number o q v backup Backs up files and file systems by i node Usage backup Level b Number c f Device L Length u FileSystem w W b Number For backups by name specifies the number of 512 byte blocks for backups by i node specifies the number of 1024 byte blocks to write in a single output operation When the backup command writes to tape devices the default is 100 for backups by name and 32 for backups by i node The write size is the number of block
251. heir normal places This command is typically used during system initialization and the corresponding mounts are referred to as automatic mounts CacheFS mount specific The CacheFS specific version of the mount command mounts a cached file system If necessary it NFS mounts its back file system It also provides a number of CacheFS specific options for controlling the caching process AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note If the mount command encounters a journaled file system that was not unmounted before reboot a replay of any JFS log records is attempted In order to move a compatible journaled file system to a system running an earlier release of the AIX operating system the file system must always be unmounted cleanly prior to its movement Failure to unmount first may result in an incompatible JFS log device If the movement results in an unknown log device the file system should be returned to the system running the latter operating system release and fsck should be run on the file system D 24 The ncheck command The following summarizes the options for the ncheck command ncheck Generates path names from i node numbers Usage ncheck a i InNumber s s FileSystem a Lists the dot and dot dot file names i INNumber Lists only the file or files specified by the InNumber parameter S Lists only special files and files w
252. hlv N Change a Logical Volume Type or select values in entry fields Press Enter AFTER making all desired changes TOP Entry Fields Logical volume NAME odin_lv Logical volume TYPE jfs POSITION on physical volume middle RANGE of physical volumes minimum MAXIMUM NUMBER of PHYSICAL VOLUMES 32 to use for allocation Allocate each logical partition copy yes on a SEPARATE physical volume RELOCATE the logical volume during yes reorganization Logical volume LABEL home odin MAXIMUM NUMBER of LOGICAL PARTITIONS 512 SCHEDULING POLICY for reading writing parallel logical partition copies PERMISSIONS read write Enable BAD BLOCK relocation yes Enable WRITE VERIFY no Mirror Write Consistency yes BOTTOM F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image _ F9 Shell F10 Exit Enter Do P 1 2 9 Set schedule policy The scheduling policy can be changed for a logical volume and will take effect immediately For the command line options see chivin Appendix A 1 The chlv command on page 199 or use the SMIT fastpath smit chlv SMIT provides the following fields Logical volume name The name of the logical volume that will have its scheduling policy changed F4 for list Scheduling policy for reading writing logical partition copies There are two options for the scheduling policy Parallel Writes to logical volume copies 34 AIX Logical Volume Manager from A
253. iate level commands 285 B 1 The allocp command 0 000 eee eee 285 B 2 The cfgvg command ies o erer kpa eaea ees 287 B 3 The chlvcopy command ssssaaaaa aaaea naaa 287 B 4 The copyrawlv command 0 ee 289 B 5 The getlvcb command 2 0 eee ee 289 B 6 The getlvname command 00 ees 290 B 7 The getlvodm command sssaaa 0 c eee ee 291 B 8 The getvgname command 0 0 cee es 293 B 9 The Ichangelv command 00 00 c eee ee 293 B 10 The Ichlvcopy command 000 eee ee 294 B 11 The Ichangepv command 00 ee 295 B 12 The Icreatelv command 0 0000 eee ees 296 B 13 The Icreatevg command 0 naaa anaana 297 B 14 The Ideletelv command 00 00 cee eee 297 B 15 The Ideletepv command 000 cee ees 298 B 16 The lextendlv command 000 cee eee 298 B 17 The linstallpv command 000 e eee ee 299 B 18 The Imigratelv command 00 ee 299 B 19 The Imigratepp command 0 00 cee eee 300 B 20 The Iquerylv command 0 00 0 cece ees 300 B 21 The Iquerypv command 2 0 cee ee 302 B 22 The Iqueryvg command 00 cece ee 302 B 23 The Iqueryvgs command 0 0c eee ee 303 B 24 The Ireducelv command 2 0 nannan 304 B 25 The Iresynclp command 2 00 c cee ee 305 B 26 The lresynclv command 0 0 00
254. ice variable specifies a range the restore command automatically goes from one device in the range to the next After exhausting all of the specified devices the restore command halts and requests that new volumes be mounted on the range of devices Restores only the actual directory not the files contained in it This flag can only be used when the archive is in file system format This flag is ignored when used with the r or R flags 368 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Allows you to interactively restore selected files from a file system archive The subcommands for the i flag are cd Directory add File delete File Is Directory extract pwd verbose setmodes quit help Changes the current directory to the specified directory Specifies that the File parameter is added to the list of files to extract If File is a directory that directory and all the files contained in it are added to the extraction list unless the h flag is used If File is not specified the current directory is added to the extraction list Specifies that the File parameter is to be removed from the list of files to be extracted If File is a directory that directory and all the files contained in it is removed from the extraction list unless the h flag is used Displays the directories and files contained within the Directory parameter Directory names are disp
255. ick a different set of drives that will satisfy an exact logical volume mapping of the entire volume group The designated disks must be equal to or exceed the size of the drives which are to be exactly mirrored regardless if the entire disk is used Also if any logical volume to be mirrored is already mirrored this command will fail Q By default in mirrorvg when a volume group s contents becomes mirrored volume group quorum is disabled If the user wishes to keep the volume group quorum requirement after mirroring is complete this option should be used in the command For later quorum changes refer to the chvg command S Returns the mirrorvg command immediately and starts a background syncvg of the volume group With this option it is not obvious when the mirrors have completely finished their synchronization However as portions of the mirrors become synchronized they are immediately used by the operating system in mirror usage S Returns the mirrorvg command immediately without performing any type of mirror synchronization If this option is used the mirror may exist for a logical volume but is not used by the operating system until it has been synchronized with the syncvg command Commands called lquerylv lqueryvg lquerypv and 1slv The mirrorvg command takes all the logical volumes on a given volume group and mirrors those logical volumes This same functionality may also be Appendix A Hig
256. icy flag upperbound flag strictness flag number of copies relocation flag d VGID Return the major number for this VGID e LVID Return the logical volume name for this LVID F Return the free configured physical volumes G LVID Return the strip size for this LVID g PVID Return the physical volume name for this PVID Appendix B Intermediate level commands 291 h Return all volume group names j PVID Return the VGID for this PVID L VGDescriptor Returns the logical volumes and the LVIDs for this volume group Name or VGID I LVDescriptor Returns the LVID for this Logical volume name or LVID m LVID Returns the mount point for this LVID N LVID Return the stripe width for this LVID p PVDescriptor Return the PVID for the physical volume name or PVID P Return the physical volume names PVIDs and volume groups of all the configured physical volumes Q VGID Returns the quorum attribute for this VGID R Returns the runtime attribute from the PdAt class r LVID Returns the relocatable flag for this LVID S Returns a list of all read write optical devices s VGDescriptor Returns the volume group state for this volume group name or VGID The value returned is as follows 0 Varied off 1 Varied on with all PVs 2 Varied on with missing PVs T VGDescriptor Returns the timestamp for this volume group name or VGI
257. ied Uses the time of extraction as the modification time The default is to preserve the modification time of the files N Blocks Allows the tar command to use very large clusters of blocks when it deals with streaming tape archives Note however that on input the tar command cannot automatically determine the block size of tapes with very long block sizes created with this flag In the absence of a N Blocks flag the largest block size that the tar command can automatically determine is 20 blocks Provides backwards compatibility with older versions non AlX of the tar command When this flag is used for reading it causes the extracted file to take on the User and Group ID UID and GID of the user running the program rather than those on the archive This is the default behavior for the ordinary user Says to restore fields to their original modes ignoring the present umask The setuid setgid and tacky bit permissions are also restored to the user with root user authority Writes the files specified by one or more File parameters to the end of the archive This flag is not valid for any tape devices because such devices do not support the addition of information at the end of a tape Appendix D Other related commands 387 SBlocksb S Feet S Feet Density Specifies the number of 512 KB blocks per volume first format independent of the tape blocking factor You can also specify the si
258. ied more recently than the file specified by the File parameter o Options Specifies file system implementation specific options p Prefix Adds the prefix specified by the Prefix parameter to each path name The default prefix is dot S Writes the file size in bytes after each path name u Writes the owner s login name after each path name V VFSName Instructs the command to assume the file system is of type VFSName thus overriding the value in the etc filesystems file The command reads the i nodes in the file system specified by the FileSystem parameter and then writes information about them to standard 344 AIX Logical Volume Manager from A to Z Troubleshooting and Commands output It assumes the FileSystem is a file system which is referenced in the etc filesystems file and saves i node data for files specified by flags The output from the command consists of the path name for each requested i node number in addition to other file information that you can request using the flags The output is listed in order by i node number with tabs between all fields The default line produced by the f command includes the path name and i node number fields With all flags enabled the output fields include path name i node number size and UID user ID The Number parameter is a decimal number that specifies a number of days It is prefixed by a or plus or minus sign Therefore 3 means
259. ifier which is generated by the mkvg command The 1spv command displays information about the physical volume if the specific physical volume name is specified If you do not add flags to the 1spv command the default is to print every known physical volume in the system along with its physical disk name physical volume identifiers PVIDs and which volume group if any it belongs to Appendix A High level LVM commands 227 228 m Note If the 1spv command cannot find information for a field in the Device Configuration Database it will insert a question mark in the value field As an example if there is no information for the PP RANGE field the following might be displayed PP RANGE The 1spv command attempts to obtain as much information as possible from the description area when it is given a logical volume identifier When no flags are used the following characteristics of the specified physical volume are displayed Physical volume Name of the physical volume Volume group Name of volume group Volume group names must be unique system wide names and can be from one to 15 characters long PV Identifier The physical volume identifier for this physical disk VG Identifier The volume group identifier of which this physical disk is a member PVstate State of the physical volume If the volume group that contains the physical volume is varied on with the varyonvg command the state is active m
260. ile system Appendix D Other related commands 349 Does not display messages about minor problems but fixes them automatically This flag does not grant the wholesale license that the y flag does and is useful for performing automatic checks when the system is started normally You should use this flag as part of the system startup procedures whenever the system is being run automatically It also allows parallel checks by group If the primary superblock is corrupt the secondary superblock is verified and copied to the primary superblock t File Specifies a File parameter as a scratch file on a file system other than the one being checked if the fsck command cannot obtain enough memory to keep its tables If you do not specify the t flag and the fsck command needs a scratch file it prompts you for the name of the scratch file However if you have specified the p flag the fsck command is unsuccessful If the scratch file is not a special file it is removed when the fsck command ends V VFSName Uses the description of the virtual file system specified by the VFSName variable for the file system instead of using the etc filesystems file to determine the description If the V VfsName flag is not specified on the command line the etc filesystems file is checked and the vfs Attribute of the matching stanza is assumed to be the correct file system type Assumes a yes response to all questions asked by
261. iles backed up by name using the i flag e Entire file system backed up by i node using the Level and FileSystem parameters If you issue the backup command without any parameters it defaults to a level 9 i node backup of the root file system to the dev rfdO device The default 9uf dev rfdo dev rhd4 324 AIX Logical Volume Manager from A to Z Troubleshooting and Commands The default backup device is dev rfdO If flags are specified that are not appropriate for the specified backup device the backup command displays an error message and continues with the backup A single backup can span multiple volumes m Note 1 Running the backup command results in the loss of all material previously stored on the selected output medium 2 Data integrity of the archive may be compromised if a file is modified during system backup Keep system activity at a minimum during the system backup procedure 3 If a backup is made to a tape device with the device block size set to 0 it might be difficult to restore data from the tape unless the default write size was used with the backup command The default write size for the backup Command can be read by the restore command when the tape device block size is 0 In other words the b flag should not be specified when the tape device block size is 0 If the b flag of the backup command is specified and is different from the default size the same size must be specified with the b flag o
262. ill be used If a PPsize is specified that is larger than appropriate for the disk sizes the specified larger PPsize will be used q Specifies that the usual prompt not be displayed before the restoration of the volume group image If this flag is not specified the prompt displays the volume group name and the target disk device names S Specifies that the logical volumes be created at the minimum size possible to accommodate the file systems This size is specified by the value of LV_MIN_LPS field of the lv_data stanza of the vgname data file where vgname is the name of the volume group The s flag overrides the values of the SHRINK and EXACT_FIT fields in the logical_volume_policy stanza of the vgname data file The s flag causes the same effect as values of SHRINK yes and EXACT_FIT no would cause PhysicalVolme s Specifies the names of physical volumes to be used instead of the physical volumes listed in the vgname data file Target physical volumes must be defined as empty that is they must contain a physical volume identifier and must not belong to a volume group If the target physical volumes are new they must be added to the system using the mkdev command If the target physical volumes belong to a volume group they must be removed from the volume group using the reducevg command Appendix A High level LVM commands 269 Commands called mkvg mklv and mkfs The restvg command restores the user vol
263. ince we no longer have quorum We also use the s flag to vary on in maintenance mode since we do not want the file systems to be mounted someone might start using them and we are going to have to take the volume group off line again soon We can safely run the low level commands now lreducelv 1 VGID 1 s 3 lowlv1 map raw lreducelv 1 VGID 2 s 2 lowlv2 map raw ldeletepv g VGID p 00017d377c3abace We have removed the reference to hdisk6 from hdisk7 s VGDA We must still be extremely careful If we bring hdisk6 back online with cfgmgr or more quickly cfgmgr 1 ssar if we are using SSA disks we still cannot run extendvg yet cfomgr 1 ssar extendvg lowvg hdisk6 0516 696 extendvg Physical Volume hdisk6 belongs to another volume group lspv grep lowvg hdisk7 00017d377b71d55c lowvg hdisk6 00017d377c3abace lowvg reducevg lowvg hdisk6 0516 022 ldeletepv Illegal parameter or structure value 0516 884 reducevg Unable to remove physical volume hdiske Ke J The issue here is that as far as the ODM is concerned the volume group still contains both disks Chapter 2 Problem determination and recovery 153 odmget q name lowvg and attribute pv CuAt grep value value 00017d377c3abace0000000000000000 value 00017d377b71d55c0000000000000000 If we now vary off export and re import to clear the ODM corruption we will spin the refresh wheel again and propagate the VGDA c
264. ion algorithm Large file enabled Provides support for files greater than 2 GB For the command line options see crfs in Appendix D 5 The crfs command on page 335 or use the SMIT fastpath smit crfs SMIT provides the following fields Logical Volume Name Name of the logical volume to use Mount Point The mount point for the file system will become the logical volume label Mount Automatically at system restart Mount the file system automatically when system starts Permissions Set the permissions for the file system read write or read only Mount Options Set the security related mount options nosuid prevents setuid and setgid from programs in this file system nodev no open system calls of devices from this mount Start disk accounting Turn on accounting for this file system Fragment size Fragment size can be 512 1024 2048 and 4096 Sizes below 4096 allow partial blocks to be written Number of bytes per inode Specifies the ration of the file system size to the number of inodes Allocation group size Determines the range of allowable nbpi number of bytes per i node Chapter 1 LVM commands 51 smit crfs ra N Add a Standard Journaled File System Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields LOGICAL VOLUME name MOUNT POINT B Mount AUTOMATICALLY at system restart no PERMISSIONS read write Mount OPTIONS Start Disk Accou
265. ion at the end of a tape 388 AIX Logical Volume Manager from A to Z Troubleshooting and Commands V Lists the name of each file as it is processed With the t flag v gives more information about the tape entries including file sizes times of last modification User Number UID Group Number GID and permissions W Displays the action to be taken followed by the file name and then waits for user confirmation If the response is affirmative the action is performed If the response is not affirmative the file is ignored X Extracts the files specified by one or more File parameters from the archive If the File parameter refers to a directory the tar command recursively extracts that directory from the archive If you do not specify the File parameter the tar command extracts all of the files from the archive When an archive contains multiple copies of the same file the last copy extracted overwrites all previously extracted copies If the file being extracted does not already exist on the system the file is created If you have the proper permissions the tar command restores all files and directories with the same owner and group IDs as they have on the tape If you do not have the proper permissions the files and directories are restored with your owner and group IDs It is not possible to ask for any occurrence of a file other than the last Number Uses the dev rmtNumber file instead of the default
266. ion of the logical partitions on the physical volume The Position variable is one of the following m Allocates logical partitions in the outer middle section of each physical volume This is the default position c Allocates logical partitions in the center section of each physical volume e Allocates logical partitions in the outer edge section of each physical volume ie Allocates logical partitions in the inner edge section of each physical volume im Allocates logical partitions in the inner middle section of each physical volume B Label Set the label field for the logical volume c Copies Set the number of copies for the logical volume e InterPolicy Sets the inter physical volume allocation policy the number of physical volumes to extend across using the volumes that provide the best allocation The value of the Range variable is limited by the Upperbound variable set with the u flag and is one of the following x Allocates logical partitions across the maximum number of physical volumes mAllocates logical partitions across the minimum number of physical volumes LVName Sets the logical volume name n NewLVName Set a new logical volume name for the logical volume ID r Relocatable Sets the bad block relocation policy The BadBlocks variable is one of the following y Allows bad block relocation to occur default n Prevents bad block relocation from
267. is created and then the data is copied For the command line options see cplv in Appendix A 4 The cplv command on page 209 or use the SMIT fastpath smit cplv SMIT provides the following fields Source logical volume name The name or logical volume ID of the source logical volume F4 for list Destination logical volume The name of the new logical volume Destination volume group The name of the volume group in which the destination logical volume will reside smit cplv Copy to a user created logical volume E N Copy to a user created logical volume Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields SOURCE logical volume name DESTINATION logical volume to create Destination VOLUME GROUP name F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Conmand F7 Edit F8 Image F9 Shel1 F10 Exit Enter Do 1 2 3 Copy a logical volume to same sized or bigger logical volume This example is similar to the previous example except that the logical volume will be copied to an already existing logical volume The type field in the destination logical volume must be set to copy before this command will work Chapter 1 LVM commands 27 Note The destination logical volume must be the same size or larger than the source logical volume as not all file system data will be copied This will lead to corruption of the destination file
268. is run The shell now loads the script s korn shell functions These will be displayed on the screen but are not actually executed at these point They will then be available for the main body of the script to execute AIX Logical Volume Manager from A to Z Troubleshooting and Commands test_retum NAVE test_retum DESCRIPTION Tests functia retum code Will exit and output error message if bed This section is mitted as we are concentrating cn lines as they are executed When a line actually runs it is preceded by a the shell s default PS4 variable If a line is not preceded by a plus it is merely being read by the shell it is however useful to see these lines as they allow us to see the names of the variables the shell is using cleanup NAVE clean DESCRIPTION Called from trap command to clean up environment and exit This is the cleamp function that all the high level commands have It deletes temporary files resets locks and runs savebase clearup set xv This set xv was added by trclvm trap 01215 C4 delete temporary files m f tmp pvid nares m f tmp rootdevicess m f tmp lvinfoss m f tmp pvinfos m tmp vginfoss These lines were commented out by trelvm s t flag exit EXIT COE H PATH usr bin ete usr sbin usr ucb usr bin X11 soin SPATH PATH usr bin etec usr sbin usr ucb usr bin X11 sbin
269. issing or removed If the physical volume is varied off with the varyoffvg command the state is varied off Allocatable Allocation permission for this physical volume Logical volumes Number of logical volumes using the physical volume Stale PPs Number of physical partitions on the physical volume that are not current VG descriptors Number of volume group descriptors on the physical volume PP size Size of physical partitions on the volume Total PPs Total number of physical partitions on the physical volume AIX Logical Volume Manager from A to Z Troubleshooting and Commands Free PPs Number of free physical partitions on the physical volume Used PPs Number of used physical partitions on the physical volume Free distribution Number of free partitions available in each intra physical volume section Used distribution Number of used partitions in each intra physical volume section A 11 The Isvg command The following summarizes the options for the 1svg command Isvg Displays information about volume groups Usage Isvg L o n DescriptorPhysicalVolume i I M p VolumeGroup s L Specifies no waiting to obtain a lock on the Volume group Note If the volume group is being changed using the L flag gives unreliable date p Lists the following information for each physical volume within the group specified by the VolumeGroup parameter Physical volumeA physical volume
270. ist Put volume group into system management mode This will allow logical volume to be operated on but not opened for input output smit varyoffvg A N Deactivate a Volume Group Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields VOLUME GROUP name Put volume group in SYSTEM no MANAGEMENT mode F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Conmand F7 Edit F8 Image F9 Shell1 F10 Exit Enter Do Chapter 1 LVM commands 7 1 1 4 Remove a physical volume from a volume group The reducevg command removes one or more physical volumes from the volume group When you remove all physical volumes in a volume group the volume group is also removed The volume group must be varied on that is active before it can be reduced All logical volumes residing on the physical volumes being removed must have been removed before starting the reducevg command For the command line options see reducevg in Appendix A 25 The reducevg command on page 263 or use the SMIT fastpath smit reducevg SMIT provides the following fields Physical volume names The names of the physical volumes to be removed from the volume group F4 for list Force deallocation of all partitions This will force the de allocation of any logical partitions on this physical volume This will destroy those logical volumes and the user will be asked to confirm 8 AIX Logical Volume
271. istinguish a binary data file from a text file that contains extended characters If the File parameter specifies an executable or object module file and the version number is greater than 0 the file command displays the version stamp The 1d command explains the use of a out files The file command uses the etc magic file to identify files that have some sort of a magic number that is any file containing a numeric or string constant that indicates type D 12 The fileplace command The following summarizes the options for the fileplace command fileplace Displays the placement of file blocks within logical or physical volumes Usage fileplace I p i v File Displays the indirect blocks for the file if any The indirect blocks are displayed in terms of either their logical or physical volume block addresses depending on whether the 1 or p flag is specified Displays file placement in terms of logical volume fragments for the logical volume containing the file The 1 and p flags are mutually exclusive Note If neither the 1 flag nor the p flag is specified the 1 flag is implied by default If both flags are specified the p flag is used pP Displays file placement in terms of underlying physical volume for the physical volumes that contain the file If the logical volume containing the file is mirrored the physical placement is displayed for each mirror copy The 1 and p flags ar
272. it lsvg output g N COMMAND STATUS Command OK stdout yes stderr no Before command completion additional instructions may appear below rootvg asgard_vg software F1 Help F2 Refresh F3 Cancel F6 Command F8 Image F9 Shel1 FLO Exit Find n Find Next Chapter 1 LVM commands 17 1 1 11 Listing the configuration of a volume group The 1svg command also displays more detailed information about volume groups When information from the Device Configuration Database is unavailable some of the fields will contain a question mark in place of the missing data The lsvg command attempts to obtain as much information as possible from the description area when the command is given a logical volume identifier For the command line options see 1svg in Appendix A 11 The Isvg command on page 229 or use the SMIT fastpath smit 1svg SMIT provides the following field List option status volume group configuration logical volumes logical volume details physical volumes physical volume usage smit lsvg List contents of a volume group C SN List Contents of a Volume Group Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields VOLUME GROUP name park_vg List OPTION status Fl Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image F9 Shell1 F10 Exit Enter Do a A 18 AIX Logical Volume Manager from A to Z Troubleshooting and
273. ith set user ID mode The ncheck command displays the i node number and path names for file system files It uses question marks displayed in the path to indicate a component that could not be found Path names displayed with ellipses at the beginning indicate either a loop or a path name of greater than 10 entries The ncheck command uses a simple hashing algorithm to reconstruct the path names that it displays Because of this it is restricted to file systems with less than 50 000 directory entries D 25 The restore command The following summarizes the options for the restore command Appendix D Other related commands 365 restore To restore files archived by file name Usage restore x d M v q b Number f Device s SeekBackup X VolumeNumber File s restore To list files archived by file name Usage restore T q v b Number f Device s SeekBackup restore To restore files archived by file system Usage restore r B q v y b Number f Device s SeekBackup restore To restore files archived by file system multi volume Usage restore R B v y b Number f Device s SeekBackup restore To interactively restore files archived by file system Usage restore i h m q v y b Number f Device s SeekBackup restore To individually restore files archived by file system Usage restore x B h mq v y b Number
274. ition map can be selected Chapter1 LVMcommands 41 42 smit 1lslv Showing the physical volume map a N COMMAND STATUS Command OK stdout yes stderr no Before command completion additional instructions may appear below grom 1v home odin PV COPIES IN BAND DISTRIBUTION hdisk11 012 000 000 100 000 012 000 000 000 F1 Help F2 Refresh F3 Cancel F6 Command F8 Image F9 Shel1 FLO Exit Find n Find Next Ne S Explanation of the fields displayed PV The physical volume name Copies The number of physical partitions used for each mirrored copy Shows if a logical volume is mirrored on the same physical volume In band The percentage of physical partitions on the physical volume belonging to the logical volume that have met their intra physical volume allocation policy Distribution The number of physical partitions laying in each region on the physical volume shown from outer edge to inner edge AIX Logical Volume Manager from A to Z Troubleshooting and Commands smit lslv Showing the logical partition map a N COMMAND STATUS Command OK stdout yes stderr no Before command completion additional instructions may appear below TOP grom 1v home odin LP PP1 PV1 PP2 PV2 PP3 PV3 0001 0109 hdisk11 0002 0110 hdisk11 0003 0111 hdisk11 0004 0112 hdisk11 0005 0113 hdisk11 0006 0114 hdisk11 0007 0115 hdisk11 0008 0116 hdisk11 0009 0117 hdisk11 0010 0118 hdisk11 0011 0119 hdisk11 0012 0
275. k for error retum Determine the major number of the volume group entered VGMAJOR getlvodm d SVGNAME 2 gt dev null if SVGMAJOR eq SROOIMAJOR then dspmsg s 1 omdlvm cat 762 lvmmsg 762 gt n SPROGNAME SVGNAME gt amp 2 exit fi else gt amp 2 gt amp 2 dspmsg s 1 cmdlvm cat 606 lvmmsg 606 n SPROGNAM dspmsg s 1 cmdlvm cat 760 lvmmsg 760 n SPROGNAM exit Ci 3 T N The above code sets up variables for the script s use and also performs sanity checking Now that the code block has been read by the shell we trace through the execution of the individual lines of code to see how this works 114 AIX Logical Volume Manager from A to Z Troubleshooting and Commands fs n victinvg Here is the parsing check for a nm mill SVG AM If this had failed we would have fallen through to the lvmmsg 606 volume group nawe not entered and lvmmsg 760 the usage statement for exportvg shown just above VQWEevictinvg The next three lines copy the 1s 1 output for the IPL device into a temporary file The order in which they are displayed may seam a little strange this is due to shell pipelining of the original 1s 1 dev TPL rootvg sed s g gt tmp rootdevicess comand sed s g 1s 1 dev IPL rootvg 1 gt tmp rootdevicel 7482 We now pull rootvg s major mmber out of this temporary file read skipfld skipfld skipfld skipfld ROOIMAJOR skipfld 0 lt tmp root
276. layed with a slash after the name Files and directories within the specified directory that are on the extraction list are displayed with an asterisk before the name If verbose mode is on the i node number of the files and directories is also displayed If the Directory parameter is not specified the current directory is used Restores all the directories and files on the extraction list Displays the full path name of the current directory Causes the 1s subcommand to display the i node number of files and directories Additional information about each file is also displayed as it is extracted from the archive Sets the owner mode and time for all directories added to the extraction list Causes restore to exit immediately Any files on the extraction list are not restored Displays a summary of the subcommands Appendix D Other related commands 369 Sets the access and modification times of restored files to the time of restoration If a restored file is an archive created by the ar command the modification times in all the member headers are also set to the time of restoration You can specify the m flag only when you are restoring individually named files and only if the x or x flags are also specified When the mflag is not specified the restore Command maintains the access and modification times as they appear on the backup medium The m flag is used when the data is in the AIX Version 2 backup by i
277. le should always be gathered about the nature of the problem Only then should attempts to fix the problem be made In fact the LVM scripts have code to detect and prevent this problem being caused by the high level commands themselves For example you cannot mkvg a volume group with non alphanumeric characters in its name However that does not mean that this situation could never occur whether through malicious user intervention or unpredictable circumstance 2 4 Gathering information about the problem As discussed in A methodology for problem determination on page 60 problem determination involves a information gathering process of forming deeper and deeper hypotheses For LVM this generally means starting with the high level commands and working down deeper towards the contents of the ODM and the low level LVM structures The VGDAs VGSAs and LVCBs Once again it is important to understand the problem before the attempt to fix it is made Some people just throw synclvodm exportvg and importvg at a problem making the common assumption that the ODM must be corrupt and leaving the scripts to sort out the damage Of course sometimes this is possible but when you are dealing with VGDA corruption these tactics can make things worse This is detailed in the following example another even stronger case of the need to be methodical about tackling LVM problems is presented in Corruption example 3 Low level VGDA co
278. les in the backup image are copied after the bosinst data and image data files X Expands tmp if needed Commands used lsvg lslv lsjfs df bootinf and bosboot The mkvgdata command creates a file containing information about a volume group for use by the savevg and restvg commands The information includes the list of logical volumes file systems and their sizes and the volume group name One of the following files is created depending on the type of volume group image data Created for information about the root volume group rootvg The savevg command uses this file to create a backup image that can be used by the bosinstall routine to reinstall the volume group to the current system or to a new system The mkvgdata command overwrites this file if it already exists The image data file is located in the directory Appendix A High level LVM commands 261 vgname data Created for information about a user volume group The vgname variable reflects the name of the volume group The savevg command uses this file to create a backup image that can be used by the restvg command to reinstall the user volume group The mkvgdata Command overwrites this file if it already exists The vgname data file is located in the tmp vgdata vgname directory where vgname is the volume group name The information in either of these files can be edited by the user before issuing the savevg command A 23 The readlv
279. level commands are designed to be called by high level commands and a large amount of checking is performed by the high level code that will not be called if low level commands are used For completeness note that in this example only the PVID was corrupt We did not in fact need to re create the VGDA and rebuild the maps The above procedure was introduced here as a valuable technique in LVM problem determination that should be understood as quickly as possible Appendix E 1 trclvm on page 393 gives a script that we can used to simply adjust the original PVID back on to disk Note that this is a highly 86 AIX Logical Volume Manager from A to Z Troubleshooting and Commands unsupported technique and is presented for explanatory purposes only Again use high level commands wherever possible Going back to the start of the rebuild process from the point where the varyonvg fails we see varyonvg workvg PV Status hdisk5 00017d372f1de834 PVACTIVE hdisk6 0000000000000000 INVPVID 00017d372 67da78 NAMIDMTCH varyonvg Volume group workvg is varied on chpvid 00017d372f67da78 hdisk6 varyoffvg workvg varyonvg workvg repair any damage if exportvg was run mount workfs ls workfs fs_ok lost found G J We now return to working through the methodological approach to LVM problem determination 2 6 Inspection commands As previously stated the general procedure in restoring LVM i
280. ll other copies of the logical partition whether or not they are stale Postpones writes for this volume group on other active concurrent cluster nodes until this sync operation is complete When using the u flag the r flag does not require that all the nodes on the cluster support the r flag This flag is ignored if the volume group is not varied on in concurrent mode Reads the names from standard input I LVName s The logical volume device name s to resynchronize P NumParallelLps Numbers of logical partitions to by synchronized in parallel The valid range for NumParallelLps is 1 to 32 NumParallelLps must be tailored to the machine disks in the volume group system resources and volume group mode p PVName s The physical volume device name s to resynchronize v VGName s The volume group device name s to resynchronize Commands called getlvodm lqueryvg lresyncpv and lresynclv A 35 The unmirrorvg command The following summarizes the options for the unmirrorvg command unmirrorvg Removes the mirrors that exist on volume groups or specified disks Appendix A High level LVM commands 277 Usage unmirrorvg c Copies VolumeGroup PhysicalVolume s c Copies Specifies the minimum number of copies that each logical volume must have after the unmirrorvg command has finished executing If you do not want all logical volumes to have the same numbe
281. llowing fields for each physical volume in the logical volume PV Physical volume name Copies The following three fields e The number of logical partitions containing at least one physical partition no copies on the physical volume e The number of logical partitions containing at least two physical partitions one copy on the physical volume e The number of logical partitions containing three physical partitions two copies on the physical volume In band The percentage of physical partitions on the physical volume that belong to the logical volume and were allocated within the physical volume region specified by Intra physical allocation policy DistributionThe number of physical partitions allocated within each section of the physical volume Outer edge outer middle center inner middle and inner edge of the physical volume 220 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Lists the following fields for each logical partition LPs PV1 PP1 PV2 PP2 PV3 PP3 Logical partition number Physical volume name where the logical partition s first physical partition is located First physical partition number allocated to the logical partition Physical volume name where the logical partition s second physical partition first copy is located Second physical partition number allocated to the logical partition Physical volume name where the logical partitio
282. llows A tmp All files are backed up relative to current working directory To exclude any file or directory for which it is important to have the search match the string at the beginning of the line use caret character as the first character in the search string followed by dot character followed by the filename or directory to be excluded If the filename or directory being excluded is a substring of another filename or directory use caret character followed by dot character to indicate that the search should begin at the beginning of the line and or use dollar sign character to indicate that the search should end at the end of the line f Device Specifies the device or file name on which the image is to be stored The default is the dev rmt0 device 272 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Calls the mkvgdata command which generates the vgname data file The vgname data file contains information on logical volumes file systems and physical volumes This information is included in the backup for future use by the installation process Calls the mkvgdata command with the m flag to generate map files Disables software packing of the files as they are backed up Some tape drives use their own packing or compression algorithms This flag only applies to AIX Version 4 2 or later Verbose mode Lists files as they are backed up This flag only
283. lminor 309 M migfix 309 migratepv 49 160 235 mini ODM 65 75 mirror write consistency 24 state 40 mirror write consistency cache 122 mirrorvg 236 MISSINGPV_VARYON 283 mkcd 238 mkfs 360 mklv 23 243 mklvcopy 29 250 mksysb 254 mkszfile 256 257 mkvg 3 60 259 mkvgdata 261 mount 83 362 mount option 51 mount point 22 41 51 55 N NBPI 51 ncheck 365 O od 121 412 AIX Logical Volume Manager from A to Z Troubleshooting and Commands ODM 62 63 ODM corruption 60 128 ODM lock 107 ODM stanzas 99 odmadd 86 315 odmchange 315 odmcreate 316 317 odmdelete 318 odmdrop 318 odmget 69 87 318 odmshow 319 P paging space 7 21 PdAt 65 physical partition 5 size 4 state 48 physical volume 8 name 4 8 11 state 22 46 problem determination 59 putlvcb 86 310 putlvodm 63 116 119 311 PV_MISSING 283 PVID corruption 75 Q quorum 5 11 153 283 R readivcopy 262 recovery techniques 59 redefinevg 129 263 reducevg 8 9 263 reorganization 25 reorgvg 14 31 32 265 replacepv 266 restore 365 restvg 268 rmfs 376 rmlv 82 267 rmlvcopy 268 S savebase 65 75 135 376 savevg 270 schedule policy 34 scheduling policy 25 34 40 parallel 34 sequential 35 SIGHUP 65 SIGINT 65 SIGTERM 65 snap 377 splitlvcopy 274 stale partition 6 41 strictness policy 24 30 32 strip size 41 stripe size 25 striping width 41 super strictness 32 sync 380 synchronization 6 30 synclvodm 66 129
284. lue 64k type R generic DU rep y nls_index 1101 CuAt name stripelv attribute upperbound value 2 type R generic DU rep y nls_index 646 CuAt name stripelv attribute size value 2 type R generic DU rep y nls_index 647 Chapter 2 Problem determination and recovery 105 Striped logical volume ODM objects CuDv name stripelv status 1 chgstatus 1 ddins location parent odmvg connwhere PdDvLn logical_volume lvsubclass lvtype CuDep name odmvg dependency stripelv CuDvDr resource devno valuel 41 value2 2 value3 stripelv 2 6 6 6 Logical volumes mirrored and striped Mirrored and striped logical volumes possesses the objects of both the mirrored logical volume and the striped logical volume This is not shown here for reasons of space The ODM should also be inspected for any strangeness as seen in Corruption example 1 Simple ODM corruption on page 66 The ODM lock is worthy of special consideration This is used as a serialization lock for a particular volume group one lock stanza for each lock that is held It will only be present if a lock is taken There is no entry if the lock is free If the lock is taken commands requiring the lock will hang until it is released A N lsvg mirrvg 0516 1201 lsvg Warning Volume group mirrvg is locked
285. lume group if a majority of the physical volumes are not accessible no Quorum This condition is true even if the quorum checking is disabled Disabling the quorum checking will only ensure that the volume group stays varied on even in the case of loss of quorum The volume group will not vary on if there are any physical volumes in PV_MISSING state and the quorum checking is disabled This condition is true even if there are a quorum of disks available To vary on in this situation either use the force option or set an environment variable MISSINGPV_VARYON to TRUE set this value in etc environment if the volume group needs to be varied with missing disks at the boot time In the above cases using the force varyon option and using MISSINGPV_VARYON variable you take full responsibility for the volume group integrity Appendix A High level LVM commands 283 284 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Appendix B Intermediate level commands These commands are usually called by the high level commands and are not aimed to the end user A careful system administrator can still use these commands Be aware that not all arguments and options are checked here and that you may end up destroying your system using these commands This appendix lists and explains the options for this set of commands B 1 The allocp command The following summarizes the options for the allocp command allocp An object file
286. lume s else syncphyvol then sycronize the physical volume s fi n n 1 M synclogvol J We switch to the synclogvol function The next few lines gather the information to pass to the low level action command CONC_STATE and CONC_SYNC_LOCK are used to manage concurrent volume groups We also see them as checkpoints in cleanup 176 AIX Logical Volume Manager from A to Z Troubleshooting and Commands CONC_SYNC_LOCK 0 getlvodm 1 stalelv LVID 00017d379 0b1047 1 0 0 getlvodm b 00017d379 0b1047 1 TMP_VGNAME stalevg getlvodm v stalevg VGID 00017d379 0b1047 lqueryvg g 00017d379 0b1047 C CONC_STATE 0 0 eq 1 J Now that information is gathered here we really just wanted an LVID we call the low level 1resynclv to act We then check for success and perform tidying up as we saw before in exportvg and redefinevg lresynclv 1 00017d379 0b1047 1 0 0 0 eq 1 a 0 eq 1 check exit_code to see if any unsuccessful attempts were made and if not set exit_code to successful if SEXIT_CODE 2 then EXIT _CODE 0 Reset exit code to indicate successful completion of syncvg fi 1 2 EXIT_CODE 0 exit trap will handle cleanup exit cleanup trap 01215 mm f tmp pvids17582 tmp pvinfol7582 0 eq 1 a 0 eq 1 exit 0 H t tt tt NS The syncvg is now complete Problems removing a mirrored copy
287. lv it would also not be on this disk In these kind of situations the copy hdisk should be taken off line and the original hdisk should be imported The copy hdisk s PVID can be reset with chdev a pv clear hdisk Chapter 2 Problem determination and recovery 161 allocp and ghost maps In this section and the section on VGSAs on page 170 techniques for recovering from the worst case scenario of lost or corrupt copies of all VGDAs are shown In this situation we have no good copies of the partition maps to work with The intermediate command allocp has predictable behavior in the sense that when it is run with a certain set of values for the following set of items a certain logical to physical partition map will be returned e Allocation policy and level of strictness e Map of used free partitions for a volume group made up of certain number of hdisks each having a certain number of physical partitions distributed among any pre existing logical volumes in a certain manner e Number of partitions requested for a given logical volume For example assume the new volume group allocpvg made up of one hdisk hdisk6 This hdisk has 1075 physical partitions When mklv asks allocp to provide it with a map for a new logical volume with the default outer middle allocation policy when it requests 10 partitions the following map should be returned mklv y allocplv allocpvg 10 allocplv 1lslv m allocplv allocplv N
288. m A to Z Troubleshooting and Commands Isvg rootvg lsvg o 0516 304 Unable to find device id 00017d3770963601 in the Device Configuration Database vgid 00017d3770963601 rootvg lspv M hdisk6 0516 320 Physical volume hdisk6 is not assigned to a volume group 1lslv m targetlv 0516 306 1slv Unable to find targetlv in the Device Configuration Database I p The high level inquiry commands that reference the LVM are unable to find any information for this volume group or its dependent logical volumes 1s targetfs data lost found umount targetfs mount targetfs 1s targetfs data lost found Note however that even though LVM is confused our data is still available We can even unmount and remount the file system Some people might try exporting or performing a vary on Exporting would never work since we can see by our data that we are already varied on seen as the missing device id 00017d3770963601 in the 1lsvg o output fo N exportvg targetvg 0516 306 getlvodm Unable to find targetvg in the Device Configuration Database 0516 772 exportvg Unable to export volume group targetvg varyonvg targetvg 0516 310 varyonvg Unable to find attribute lock in the Device Configuration Database Execute synclvodm to attempt to correct the database G J The recommended repair action synclvodm targetvg Will fail since it depends on having
289. m anymore Other non AIX code can also do this when working with raw logical volumes if its programmers are unaware of this issue Since the LVCB holds information such as the creation date of the logical volume information about mirrored copies and possible mount points in a journaled file system if its overwritten these values become undefined This will not be the case if we are using a bigVG since we run off the LVCBs in VGDA The following discussion mainly applies to non big VGs Certain LVM commands are required to update the LVCB as part of completeness of algorithms in LVM The old LVCB area is first read and analyzed to see if it is a valid LVCB If the information is verified as valid LVCB information then LVM writes an updated LVCB If the information is not valid then LVM does not perform an update due to the user data corruption possibility Instead LVM displays a warning such as 0516 622 synclvodm Warning cannot write lv control block data It is important to realize that loss of an LVCB does not prevent manipulation of a logical volume Once an LVCB is lost it is still possible to perform the following actions on the logical volume eextendlv emklvcopy ermlv AIX Logical Volume Manager from A to Z Troubleshooting and Commands e crfs d note this will destroy any information within the LVCB The problem with logical volumes with corrupted LVCBs is that they cannot be completely and reliably imported We sa
290. m backups and restores The File parameter is ignored when using either the r or the r flag File name archives File name archives are created by specifying a list of file names to archive to the backup command The restore command restores the files from a file name archive without any special understanding of the underlying Appendix D Other related commands 375 structure of the file system The restore command allows for metacharacters to be used when specifying files for archive extraction This provides the capability to extract files from an archive based on pattern matching A pattern file name should be enclosed in single quotations and patterns should be enclosed in brackets D 26 The rmfs command The following summarizes the options for the rmfs command rmfs Removes a file system Usage rmfs r FileSystem r Removes the mount point of the file system The rmfs command removes a file system If the file system is a journaled file system JFS the rmfs command removes both the logical volume on which the file system resides and the associated stanza in the etc filesystems file If the file system is not a JFS file system the command removes only the associated stanza in the etc filesystems file The FileSystem parameter specifies the file system to be removed You can use the Web Based System Manager File Systems application wsm fs fast path to run this command You can also
291. m in the volume group given with the v flag or in rootvg if that flag is not used m mksysb_image Specifies a previously created mksysb image If you do not give the mflag mkcd calls mksysb See the m flag for more information about where the mksysb image is placed 240 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Creates physical partition mapping during the mksysb or savevg creation You cannot use this flag with the mor s flags The p flag is not recommended for generic backup CDs p pkg_source_dir Names the directory or device that contains device and kernel package images The device can only be a CD device for example dev cd0 If you use the same CD R device that you gave with the dflag the product CD media must be inserted into the CD R drive first mkcd then prompts you to insert the writable CD before the actual CD creation You must use the p flag if using the c flag Prevents mkcd from removing the final CD images mkcd defaults by removing everything that it creates when it finishes executing The xr flag allows multiple CD image sets to be stored or for CD creation burn to occur on another system If multiple volumes are needed the final images are uniquely named using the process ID and volume suffixes Stops mkcd before writing to the CD R without removing the final CD images The s flag allows multiple CD sets to be created or for CDs to b
292. mands Retreive the quorum and auto on value from VGDA if bigvg else get them from ODM AUTO_ON getlvodm u VGNAME 2 gt dev null default value for new entry QUORUM getlvodm Q SVGNAME 2 gt dev null default value for new entry AUTO ON y AUTO ON y QUORUM y QUORUM y if n SdVAL then NqV lqueryvg NqVp dVAL if n SNqv then set SNqV if 1 gt 256 then BIGVG if 2 eq 0 then QUORUM n fi if 3 eq 0 then AUTO_ON n fi else from OM for small VGs if 0 then AUTO ON y fi if 0 then QUORUM y fi fi fi fi n hdisk6 lqueryvg NqVp hdisk6 NqV 256 n 256 set 256 256 gt 256 getlvodm u targetvg 2 gt dev null AUTO ON 3 0 AUTO ON y getlvodm Q targetvg 2 gt dev null QUORUM y L 0 t 0 J redefinevg now acts manipulating the ODM Chapter 2 Problem determination and recovery 143 VG in database or not putlvodm V SKEY VGID 2 gt dev null delete old entry putlvodm V 00017d3770963601 2 gt dev null putlvodm v VGNAME o SAUTO ON Q SQUORUM KEY VGID add the new VG putlvodm v targetvg o y Q y 00017d3770963601 The code shown above deletes any stale VGID information in the ODM It then rebuilds the VGID object Next redefinevg walks through the logical volumes for this volume group deleting stale logical volume information and rebuilding It is worth noting
293. mands The Availability variable can be either a Makes a physical volume available for logical input and output r Makes a physical volume unavailable removed for logical input and output If the physical volume is required in order to maintain a volume group quorum an error occurs and the physical volume remains open Commands called getlvodm and 1changepv m Note This command is not allowed if the volume group is varied on in concurrent mode The chpv command changes the state of the physical volume in a volume group by setting allocation permission to either allow or not allow allocation and by setting the availability to either available or removed This command can also be used to clear the boot record for the given physical volume Characteristics for a physical volume remain in effect unless explicitly changed with the corresponding flag A 3 The chvg command The following summarizes the options for the chvg command chvg A script that sets the characteristics of a volume group Usage chvg a AutoOn nly cl l Q nly u x nly t factor B VolumeGroup 204 AIX Logical Volume Manager from A to Z Troubleshooting and Commands a AutoOn Determines if the volume group is automatically activated during system startup The AutoOn variable can be either of the following n The volume group is not automatically activated during system startup
294. mber 304 AIX Logical Volume Manager from A to Z Troubleshooting and Commands B 25 The lresynclp command The following summarizes the options for the 1resynclp command Iresynclp An object file command that synchronizes all physical partitions that belonging to a particular logical partition Usage Iresynclp I LVID n LPNum I LVID The LVID of the logical volume that is to be synchronized n LPNum The partition number of the physical partition that is to be synchronized B 26 The Iresyncly command The following summarizes the options for the 1resynclv command lresynclv An object file command that synchronizes all the mirrored logical partition s in the logical volume Usage lresynclv I LVID I LVID The LVID of the logical volume to resynchronize B 27 The lresyncpv command The following summarizes the options for the 1resyncpv command Iresyncpv An object file command that synchronizes all mirrored partitions on a physical volume Usage Iresyncpv g VGID p PVID g VGID The VGID of the volume group that contains the physical volume to be synchronized p PVID The PVID of the physical volume to be synchronized Appendix B Intermediate level commands 305 B 28 The lvaryonvg command The following summarizes the options for the 1varyonvg command Ivaryonvg An object file that Varies a Volume Group online
295. mber Specifies the number of 512 byte blocks to write in a single output operation When the backup command writes to tape devices the default is 100 for backups by name The write size is the number of blocks multiplied by the block size The default write size for the backup command writing to tape devices is 51200 100 512 for backups by name The write size must be an even multiple of the tape s physical block size Appendix A High level LVM commands 271 Excludes files listed in the etc exclude rootvg file from being backed up The rules for exclusion follow the pattern matching rules of the grep command Note If you want to exclude certain files from the backup create the etc exclude rootvg file with an ASCII editor and enter the patterns of file names that you do not want included in your system backup image The patterns in this file are input to the pattern matching conventions of the grep command to determine which files will be excluded from the backup If you want to exclude files listed in the etc exclude rootvg file select the Exclude Files field and press the Tab key once to change the default value to yes For example to exclude all the contents of the directory called scratch edit the exclude file to read as follows scratch For example to exclude the contents of the directory called tmp and avoid excluding any other directories that have tmp in the pathname edit the exclude file to read as fo
296. mber of the owner and group Note The owner and group names for remote files are taken from the local etc passwd file e Number of links to the file e If the i node is for a normal file length of the file e If the i node is for a device major and minor device designations e Date of the last i node update e Date of the last file modification e Date of the last reference to the file If you specify the i nodeNumber and Device parameters the istat command also displays in hexadecimal values the block numbers recorded in the i node Note The Device parameter cannot refer to a remote device D 18 The logform command The following summarizes the options for the logform command logform Initializes a logical volume for use as a journal file system JFS log Usage logform LogName LogName The name of the device file in dev for example dev jfslog1 356 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note Executing the logform command on any journaled file system JFS log device can result in data loss and file system corruption for all JFS file systems logged by the log device The logform command initializes a logical volume for use as a JFS log device The logform command destroys all log records on existing log devices which may result in file system data loss The Logname parameter specifies the absolute path to the logical volume to be initiali
297. me Manager from A to Z Troubleshooting and Commands We now begin the parsing code Unlike exportvg there is a loop for parsing flags here fy Parse command line options set getoot d i V LF getoot d i V LF d hdisk6 targetvg get d hdisk6 targetvg if 0 if there is a syntax error then dsgomsg s 1 ardlvm cat 560 lumsg 560 gt n SPROG gt amp 2 exit fi 0 0 FAS WAE IFG IAE VUS WAL IFG HUG FUG QVAL 1FLAG IVe VELAG WALE IFAC FFIAG all FLAG values and associated values are reset so no extraneous flags are pulled in fron the shell environment outside the script while 1 While there is a comend line optim do case 1 in d GFIAG d GVAI S2 hift shift key device to amp fine VW i iFLAG i iVAl S2 shift shift id of VG to defire V VFLAG V WAL amp S 2 shift shift mjor mnber of G L LELAG L ghift importvg called with L qotion F FELAG F ghift importvg called with F qotion esac dane d dPIAG d AVALehdisk6 hift hift ae ates se The aly FLAG set is the d Parse command line arogurents shift skip past from getot hift if H 1 then dsomsg s 1 ardlvm cat 560 lumsg 560 gt n SPROG gt amp 2 exit fi 1 1 VQWVE S1 E VAPMERtargetyg J The command line option parsing has completed successfully We now b
298. me backup Once started the restore Command behavior is the same as with the x flag The x flag applies to file name archives only 372 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Restores individually named files specified by the File parameter If the File parameter is not specified all the archive members are restored If the File parameter is a directory and the archive is in file name format only the directory is restored If the File parameter is a directory and the archive is in file system format all the files contained in the directory are restored The file names specified by the File parameter must be the same as the names shown by the restore T command Files are restored with the same name they were archived with If the file name was archived using a relative path name filename the file is restored relative to the current directory If the archive is in file system format files are restored relative to the current directory The restore command automatically creates any needed directories When using this flag to restore file system backups you are prompted to enter the beginning volume number The restore command allows for shell style pattern matching metacharacters to be used when specifying files for archive extraction The rules for matching metacharacters are the same as those used in shell pathname globbing namely asterisk Matches zero or more characters bu
299. me group 1 Only displayed if volume group is concurrent capable AIX Logical Volume Manager from A to Z Troubleshooting and Commands 1 1 12 List contents of a volume group The 1svg command will also display information about the logical volumes or physical volumes in the volume group As seen in the previous section either logical volume or physical volume information can be specified in the list option smit lsvg Logical volume list option fe N COMAND STATUS Command OK stdout yes stderr no Before command completion additional instructions may appear below asgard vg LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT odin 1v jfs 12 12 1 closed syncd home odin thor 1v jfs 12 12 1 closed syncd home thor loglv00 jfslog 1 1 1 closed syncd N A loki 1v jfs 10 20 1 closed syncd N A F1 Help F2 Refresh F3 Cancel F6 Command F8 Image F9 Shel1 F1O Exit Find n Find Next Explanation of the fields displayed LV Name Name of the logical volume Type The logical volume type for example jfs journaled file system jfslog jfs log boot paging or dump LPs The number of logical partitions currently in the logical volume PPs The number of physical partitions currently in the logical volume PVs The number of physical volumes that the logical volume is spread over LV State The logical volume state open active closed inactive syncd mirror copies synchronized stale mirror copies not in sync Chapter
300. mmand F7 Edit F8 Image F9 Shel1 F10 Exit Enter Do N S 1 1 7 Add physical volume to an existing volume group The physical volume is checked to verify that it is not already in another volume group If the system believes the physical volume belongs to a volume group that is varied on it exits But if the system detects a description area from a volume group that is not varied on it prompts the user for confirmation in continuing with the command The previous contents of the physical volume will be lost so the user must be cautious when using the override function For the command line options see extendvg in Appendix A 7 The extendvg command on page 215 or use the SMIT fastpath smit extendvg SMIT provides the following fields Volume group name The unique name of the volume group to be extended F4 for list Physical volume name s The name s of the physical volume s to be added to the volume group F4 for list 12 AIX Logical Volume Manager from A to Z Troubleshooting and Commands smit extendvg gt N Add a Physical Volume to a Volume Group Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields VOLUME GROUP name PHYSICAL VOLUME names F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image F9 Shel1 FLO Exit Enter Do Ne S 1 1 8 Change name of volume group To change the name of a volume group
301. more than 3 days 3 means less than 3 days and 3 means 3 days where a day is defined as a 24 hour period The command lists only a single path name out of many possible ones for an i node with more than one link unless you specify the 1 flag With the 1 flag the command lists all links D 11 The file command The following summarizes the options for the file command file To classify the file type Usage file m MagicFile f FileList File file To Check the Magic File for format errors Usage file c m MagicFile C Checks the specified Magic File the etc magic file by default for format errors This validation is not normally done File typing is not done under this flag f FileList Reads the specified file list The file must list one file per line and must not contain leading or trailing spaces m MagicFile Specifies the file name of the Magic File the etc magic file by default The file command reads the files specified by the File parameter or the FileList variable performs a series of tests on each file and attempts to Appendix D Other related commands 345 classify them by type The command then writes the file types to standard output If a file appears to be in ASCII format the file command examines the first 1024 bytes and determines the file type If a file does not appear to be in ASCII format the file command further attempts to d
302. more than one step is required and familiarity with some of the commands used previously is needed To change the name of a volume group the volume group must first have all its logical volumes closed that is if they contain file systems then they must be unmounted The volume group must then be de activated then exported The volume group is then imported with the new name reactivated and any file systems remounted The only command not dealt with above is the command to export the volume group exportvg For the command line options see exportvg in Appendix A 5 The exportvg command on page 211 or use the SMIT fastpath smit exportvg SMIT provides the following fields Volume group name The unique name of the volume group to be exported F4 for list Chapter 1 LVM commands 13 smit exportvg gt N Export a Volume Group Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields VOLUME GROUP name F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image F9 Shel1 F10 Exit Enter Do N S 1 1 9 Reorganize a volume group The reorgvg command reorganizes the placement of allocated physical partitions within the volume group according to the allocation characteristics of each logical volume Specific logical volumes can also be reorganized within the volume group In the reorganization the order of priority goes in the order to t
303. ms file All check true file systems are in group 1 The fsck command attempts to check the root file system before any other file system regardless of the order specified on the command line or in the etc filesystems file The fsck command checks for the following inconsistencies Blocks or fragments allocated to multiple files i nodes containing block or fragment numbers that overlap i nodes containing block or fragment numbers out of range Discrepancies between the number of directory references to a file and the link count of the file Illegally allocated blocks or fragments i nodes containing block or fragment numbers that are marked free in the disk map i nodes containing corrupt block or fragment numbers A fragment that is not the last disk address in an i node This check does not apply to compressed file systems Files larger than 32 KB containing a fragment This check does not apply to compressed file systems Size checks e Incorrect number of blocks e Directory size not a multiple of 512 bytes These checks do not apply to compressed file systems Directory checks e Directory entry containing an i node number marked free in the i node map e i node number out of range e Dot link missing or not pointing to itself e Dot dot link missing or not pointing to the parent directory e Files that are not referenced or directories that are not reachable Inconsistent disk map Inconsistent i
304. n example 2 PVID corruption on page 75 Note that from this point on we are on the wrong track reducevg lowvg hdisk6 usr sbin reducevg 227 21524 Segmentation fault coredump 0516 882 reducevg Unable to reduce volume group We might try 1deletepv Since we have started working with low level commands we put the VGID into VGID to save typing lspv grep hdiske hdisk6 00017d377c3abace lowvg VGID 00017d3780467 5b ldeletepv g SVGID p 00017d377c3abace 0516 016 ldeletepv Cannot delete physical volume with allocated partitions Use either migratepv to move the partitions or reducevg with the d option to delete the partitions This should be simple enough We remember we cannot ldeletepv a logical volume with allocated partitions We set out to create maps in a format ldeletepv will understand This is done with lquerylv L lt lvid gt r Chapter 2 Problem determination and recovery 149 lquerylv L SVGID 1 r 00017d377c3abace 216 1 00017d377c3abace 217 2 00017d377c3abace 218 3 lquerylv L SVGID 2 r 00017d377b71d55c 216 00017d377c3abace 219 00017d377b71d55c 217 00017d377c3abace 220 lquerylv L SVGID 1 r gt lowlvl map raw lquerylv L SVGID 2 r gt lowlv2 map raw bothdisks L J ne AE a i aa We should edit lowlv2 map raw bothdisks since it currently contains all the logical partitions for both mirrors of the logical volume It should only contain the copies
305. n may fail since the partition number may no longer exist due to reduction It is recommended that backup is taken before the conversion and right after the conversion if the map option is utilized Since the VGDA space has been increased substantially every VGDA update operation creating an LV changing an LV adding a PV and so forth may have a considerably longer duration C Changes the volume group into a Concurrent Capable volume group However the volume group must be varied on in non concurrent mode for this command to take effect This flag only applies to AIX Version 4 2 or later 206 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Changes the volume group into a Non Concurrent Capable volume group The volume group must be varied on in non concurrent mode for this command to take effect This flag only applies to AIX Version 4 3 or later Determines if the volume group is automatically varied off after losing its quorum of physical volumes The default value is yes The change becomes effective the next time the volume group is activated n The volume group stays active until it loses all of its physical volumes y The volume group is automatically varied off after losing its quorum of physical volumes Note Run the bosboot Or savebase command after the chvg Q n Or chvg Q y command to update the boot image Appendix A High level LVM commands 207 t fac
306. n s third physical partition second copy is located Third physical partition number allocated to the logical partition n PhysicalVolume Accesses information from the specific descriptor area of PhysicalVolume variable The information may not be current since the information accessed with the n flag has not been validated for the logical volumes If you do not use the n flag the descriptor area from the physical volume that holds the validated information is accessed therefore the information that is displayed is current The volume group need not be active when you use this flag Appendix A High level LVM commands 221 p PhysicalVolume Displays the logical volume allocation map for the PhysicalVolume variable If you use the LogicalVolume parameter any partition allocated to that logical volume is listed by logical partition number Otherwise the state of the partition is listed as one of the following used Indicates that the partition is allocated to another logical volume free Indicates that the specified partition is not being used on the system stale Indicates that the specified partition is no longer consistent with other partitions The computer lists the logical partitions number with a question mark if the partition is stale The 1slv command displays the characteristics and status of the LogicalVolume or lists the logical volume allocation map for the physical partitions on the
307. n with a prefix already defined in the PdDv class in the Device Configuration Database for other devices or be a name already used by another device Commands called allocp getlvcb getlvname getlvodm lcreatelv lquerylv lquerypv lqueryvg lreducelv lvgenminor putlvcb lvrelminor rmlv lspv and putlvodm m Note Although the splitlvcopy command can split logical volumes that are open including logical volumes containing mounted file systems this is not recommended You may lose consistency between LogicalVolume and NewLogicalVolume if the logical volume is accessed by multiple processes simultaneously When splitting an open logical volume you implicitly accept the risk of potential data loss and data corruption associated with this action To avoid the potential corruption window close logical volumes before splitting and unmount file systems before splitting The splitlvcopy command removes copies from each logical partition in LogicalVolume and uses them to create NewLogicalVolume The Copies parameter determines the maximum number of physical partitions that remain in LogicalVolume after the split Therefore if LogicalVolume has three copies before the split and the Copies parameter is 2 LogicalVolume will have two copies after the split and NewLogicalVolume will have one copy You can not split a logical volume so that the total number of copies in LogicalVolume and NewLogicalVolume after the split i
308. nent dump devices The secondary device may be outside the root volume group unless it is a paging space Configuring remote dump devices with sysdumpdev The sysdumpdev command can also be used to configure remote dump devices The following conditions must be met before a remote dump device can be configured The local and the remote host must have Transmission Control Protocol Internet Protocol TCP IP installed and configured The local host must have Network File System NFS installed The remote host must support NFS The remote host must be operational and on the network This condition can be tested by issuing the ping command The remote host must have an NFS exported directory defined such that the local host has read and write permissions as well as root access to the dump file on the remote host The remote host cannot be the same as the local host The network device driver must support remote dump D 31 The tar command The following summarizes the options for the tar command tar Manipulates archives Usage tar c r t u x b Blocks B d F h i L InputList 1 m N Blocks o p s v w Number f Archive S Blocksb S Feet S Feet Density File s Directory s C Directory s 384 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Forces input and output blocking to 20 blocks per record With this option the
309. ng fields are taken from the VGDA itself Note that in this case the functioning of LVM itself will be unimpaired as LVM works off the LVCBs in the VGDA However some other parts of AIX code particularly IPL and RAS code still expect to find valid information in the on LV LVCB This should be remembered when debugging strange LVM RAS and LVM IPL problems Some applications may also expect to find intact LVCBs at the start of the LV as well AIX Logical Volume Manager from A to Z Troubleshooting and Commands 2 6 10 Watching system calls Commands such as trace may occasionally be useful in LVM problem determination and particularly for investigating performance problems However it is not within the scope of this document to discuss the trace command LVM does have its own trace hooks 105 and 10B These are described in usr include sys trchkid h JFS trace hook IDs are also listed there usr sys include Ilvm h can also be useful since it documents the errno return codes for the various LVM library functions 2 6 11 Examining internal kernel memory structures with crash In the event of deep LVM failures either the system may crash or it may be necessary to force a dump In both cases AlX support should be contacted as such dumps are really not customer serviceable An example script using kdb to examine the in kernel memory structures for LVM is given in Appendix E 7 pvsinvg on page 398 for those interested More inf
310. ng exceeded such as e Breaking the factor size PP limit e Running out of free space e Hitting the current maximum LPs for a logical volume 2 6 3 1 Ispv We have already seen 1spv M used in pulling out maps for physical volumes lspv is also very useful for obtaining the current state of a physical volume as of the last varyonvg active missing or removed It will go to the physical volume with the VGDA containing the latest timestamp lspv n descriptor physical volume physical volume will tell us about the standard Ispv information from the point of view of the VGDA on a particular physical volume This is useful if we suspect our VGDAs have become out of sync This kind of problem can occur in twin tailed configurations It is also a way to determine if a physical volume or associated VGDA has been lost if logical volumes start showing up as open stale in mirrored configurations As a further example of the use of the n flag consider that whenever the high level commands cannot extract information for a field this will be AIX Logical Volume Manager from A to Z Troubleshooting and Commands indicated with question marks For example if hdisk1 and hdisk2 are not in the same volume group lsvg p software N software PV_NAME PV STATE TOTAL PPs FREE PPs FREE DISTRIBUTION hdisk1 active 268 38 23455002002 2005515 hdisk2 active 542 0 00 00 00 00 00 hdisk3 active 542 0 00 00 00 00
311. ng with the undoanented L or F flags Neither of these will nomelly be supplied from a user command line generally they are used when redefinevg is driven by the importyg script Accordingly this sectia is mitted here else From list of all configured Pvs build list containing PV AVES and WI for all PVs which appear to kelang to the valme gap If a W s descriptor area has the key VED then the W is omsidered in the volure group for nme in SPVS do VGID lqueryvg p Snare v 2 gt dev mill if eg 0 a SVGID SKEY VGID then PVID getlvodm p Sname if q0 a n SPVID then echo S PVID Snare gt gt trp pvlistss fi fi done fi n o n lqeryw p higo v 2 gt amp wv all VGID 00017d37e1762ac7 0 eq 0 a 000170B37e1762ac7 0001703770963601 various other unsuccessful tests for physical volumes belonging to lqeryw p hdisk6 v 2 gt amp wv all VGID 00017d3770963601 0 eq 0 a 0001703770963601 0001703770963601 getlvodm p hdisk6 PVID 00017d37524304 02 0 eq 0 a n 000170B75243d402 echo 0001783752434402 hdisk6 1 gt gt onp pvlist22794 hdisk6 is a member of the target volume group other unsuccessful tests snipped if s tmp pvlistss zero length file then dsgomsg s 1 ardlvm cat 567 lymmsg 567 gt n SPROG gt amp 2 exit fi s tmp pvlist227 4 J We now have a file
312. nging to this B LOGICAL VOLUME F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Conmand F7 Edit F8 Image Q F9 Shell F10 Exit Enter Do 1 4 Journaled file system related commands The commands in this section relate to creating and modifying journaled file systems These commands can be run from the command line or from the SMIT System Storage Management gt File Systems menu See Figure 5 File Systems Journaled File Systems File Systems y Network File Systems NFS List All File Systems Cache FS Mount a Group of File Systems Unmount a File System Unmount a Group of File Systems Add a Journaled File System Verify a File System Backup a File System Restore a File System List Contents of a Backup Add a Journaled File System on a Previously Defined Logical Volume Change Show Characteristics of a Journaled File System Defragment a Journaled File System Figure 5 SMIT file systems options 50 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 1 4 1 Add a JFS to a previously defined logical volume menu The crfs command is used to create a file system on an already created logical volume or a new one can be created using the d flag The file system will not be mounted once it is created The file system can be Standard Standard journaled file system Compressed All data is compressed using the Lempev Zed compress
313. nnot be closed rootvg cannot be converted if it needs modification of the meta data as part of the chvg t operation Unlocks the volume group This option is provided if the volume group is left in a locked state by abnormal termination of another LVM operation such as the command core dumping or the system crashing Note Before using the u flag make sure that the volume group is not being used by another LVM command 208 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Changes the mode that the Concurrent Capable volume group is varied on The volume group must be varied on in non concurrent mode for this command to take effect This flag only applies to AIX Version 4 2 or later y auto varyon the volume group in concurrent mode n auto varyon the volume group in non concurrent mode Note If the volume group is not created Concurrent Capable this command has no effect on the volume group In order for this auto varyon into concurrency of the volume group to take effect you must enter the following line into the etc inittab file re_clvmv 2 wait usr sbin clvm_cfg 2 gt amp 1 Attention This entry must be added after the entry used to initiate srcmstr Commands called getlvodm lchangelv lqueryly mv and putlvodm If there is a volume group that is infrequently used you may not want it activated at system startup because it uses kernel resources memory The chvg c
314. ns this is discussed in Islv on page 97 lspv M hdisk5 hdisk5 1 215 hdisk5 216 loglv00 1 hdisk5 217 1075 lspv M hdisk6 hdisk6 1 215 hdisk6 216 1v02 1 hdisk6 217 1v02 2 hdisk6 218 1v02 3 hdisk6 219 1v02 4 hdisk6 220 1075 I 2 This shows we can see that the problem is accessing the file system itself on 1vo2 We know hdisk6 is the one with the problem since it is returning INVPVID at the time of the vary on shown above We can confirm lv02 is inaccessible by running od cx dev 1v02 This command will return 0000000 Chapter 2 Problem determination and recovery 77 N define LVM_INVPVID 5 physical volume is not a member of the volume group define LVM _DUPPVID 6 this PV id previously appeared in the list of input PVs define LVM LVMRECNMTCH 7 VGDA indicates this PV is a member of the VG but VG id in the PV s LVM record does not match this VG define LVM NONAME 8 name not given for physical volume id which is a member of the VG define LVM NAMIDNMTCH 9 the PV id was passed in but it was not the id of the named disk PV status values which may be returned from lvm varyonvg in the of LVM NOQUORUM or LVM NOVGDAS define LVM_PVNOTFND 10 physical volume could not be opened or its IPL record or LVM record could not be read define LVM PVNOTINVG 11 the PV s LVM reco
315. ns on the physical volume The Position variable is one of the following m Allocates logical partitions in the outer middle section of each physical volume This is the default position c Allocates logical partitions in the center section of each physical volume e Allocates logical partitions in the outer edge section of each physical volume ie Allocates logical partitions in the inner edge section of each physical volume im Allocates logical partitions in the inner middle section of each physical volume e Range Sets the inter physical volume allocation policy the number of physical volumes to extend across using the volumes that provide the best allocation The value of the Range variable is limited by the Upperbound variable set with the u flag and is one of the following x Allocates logical partitions across the maximum number of physical volumes m Allocates logical partitions across the minimum number of physical volumes m MapFile Specifies the exact physical partitions to allocate Partitions are used in the order given in the MapFile parameter Used partitions in the MapFile parameter are skipped All physical partitions belonging to a copy are allocated before allocating for the next copy of the logical volume The MapFile parameter format is PVname PPnum1 PPnum2 In this example PVname is a physical volume name for example hdiskO It is one record per physical partition ora ran
316. nt for AIX Version 4 3 3 for more information about what publications are considered to be product documentation References in this publication to IBM products programs or services do not imply that IBM intends to make these available in all countries in which IBM operates Any reference to an IBM product program or service is not intended to state or imply that only IBM s product program or service may be used Any functionally equivalent program that does not infringe any of IBM s intellectual property rights may be used instead of the IBM product program or service Information in this book was developed in conjunction with use of the equipment specified and is limited in application to those specific hardware and software products and levels IBM may have patents or pending patent applications covering subject matter in this document The furnishing of this document does not give you any license to these patents You can send license inquiries in writing to the IBM Director of Licensing IBM Corporation North Castle Drive Armonk NY 10504 1785 Licensees of this program who wish to have information about it for the purpose of enabling i the exchange of information between independently created programs and other programs including this one and ii the mutual use of the information which has been exchanged should contact IBM Corporation Dept 600A Mail Drop 1329 Somers NY 10589 USA Such information may be available
317. nter again To list the supported tasks for the resource highlighted press List Once all selections have been made press Commit To exit without selecting a resource press the Exit key All Resources This selection will select all the resources currently displayed hdisk3 30 58 00 8 0 16 Bit SCSI Disk Drive 9100 MB hdisk4 30 58 00 9 0 16 Bit LVD SCSI Disk Drive 4500 MB hdisk5 30 58 00 10 0 16 Bit LVD SCSI Disk Drive 4500 MB fdo 01 D1 00 00 Diskette Drive hdisko 10 60 00 8 0 16 Bit SCSI Disk Drive 9100 MB hdisk1 10 60 00 9 0 16 Bit LVD SCSI Disk Drive 4500 MB hdisk2 10 60 00 10 0 16 Bit LVD SCSI Disk Drive 4500 MB F1 Help F4 List F7 Commit NO A Select the disk s to certify by moving the cursor to them and pressing return They will be marked with a sign When all disks are selected press F7 or escape 7 if the terminal does not support function keys You cannot certify a disk that is in use and a disk will be in PvMIssInG status while certification is being run Chapter 2 Problem determination and recovery 183 CERTIFY MEDIA TASK 802584 Device hdisk4 in location 30 58 00 9 0 The certify operation is in progress Please stand by 8 completed Disk Drive Capacity 4512 MB Data Errors Recovered 0 Data Errors Not Recovered 0 Equipment Check Errors Recovered 0 Equipment Check Errors Not Recovered 0 F3 Cancel Esc 0 Exit
318. nternational Technical Support Organization Austin Center Before joining the ITSO three years ago Laurent worked in the french Risc System 6000 Technical Center in Paris where he conducted benchmarks and presentations for the AIX and RS 6000 solutions Ronald van der Knaap is a Senior T Specialist in the Netherlands He has 11 years of experience in the Unix AIX field His areas of expertise include a wide range of system and network management products AIX related products HACMP performance and tuning and RS 6000 SP systems He has written extensively on LVM performance and journaled file systems Dugald Foreman is an AIX support specialist in England He has two years of experience in AIX both spent working for IBM His areas of expertise Copyright IBM Corp 2000 ix include problem determination in software development and the AIX base operating system He has written extensively on LVM recovery procedures Keigo Matsubara is an Advisory I T Specialist in Japan He has seven years of experience in the AIX filed His areas of expertise include a wide range of AIX related products particularly RS 6000 SP and high end storage systems He has written extensively on mirroring striping and concurrent access volume groups This is his second redbook Antony Steel is an Advisory IT Specialist in Australia He has eight years of experience in the field of Unix He holds a degree in Theoretical Chemistry from the University of Sydne
319. nting no Fragment Size bytes 4096 Number of bytes per inode 4096 Allocation Group Size MBytes 8 F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Conmand F7 Edit F8 Image F9 Shel1 FLO Exit Enter Do Ne S More information about fragment size nbpi and allocation group size can be found in the redbook AIX Logical Volume Manager from A to Z Introduction and Concepts SG24 5432 1 4 2 Change show details of a file system 52 The chfs command is used to change the attributes of a file system The new mount point automatic mounts permissions and file system size can be set or changed Some file system attributes are set at the time the file system is created and cannot be changed For the journaled file system these attributes include the fragment size block size number of bytes per i node compression and the minimum file system size For the command line options see chfs in Appendix D 2 The chfs command on page 327 or use the SMIT fastpath smit chjfs SMIT provides the following fields File system name The current mount point of the file system New mount point The new mount point for the file system AIX Logical Volume Manager from A to Z Troubleshooting and Commands Size of file system in 512 byte blocks Mount Group The new size of the file system It will be rounded up to the nearest physical partition File system can be the member of a mount group File systems that can be mounted as a
320. null od x 000000 7 a amp 206026 R210 0001 7 R Q20 016 37e5___ 2686 16a9 5288 0001 7d37 5251 Be 0000020 O 0 O O O O W W 0 004 001 0 0 026 0 003 o00c0 0000 0000 0000 00 o 0016 0003 0000040 0003 b 2 O0 O 0 001 0 001 O O Y Y Y 0 0003 0832 0000 0001 0001 0000 0000 000 000060 O 10 Y Y Y O Y Y Y Y Y Y Y Y Y Y o 00 0000 o o00 0000 00 0000 00 000 0001000 S S we see that the first eight bytes run 37e5 2686 16a9 5288 Consider that the VGDA begins with a 64 bit timestamp If we run 1queryvg on this physical volume we can see first that we have found the VGDA since we match the lqueryvg timestamp for hdisk9 If we then run lqueryvg targeting the VGID we can see that this VGDA is up to date since it matches the latest timestamp for the volume group lqueryvg p hdisk9 Tt Time Stamp 3 7e5268616a95288 lqueryvg g getlvodm v odmvg Tt Time Stamp 3 7e5268616a95288 We have also confirmed that at least this part of our LVM information record is good as well This is as far as can be done in this document in describing the VGDA and VGSA If there is a problem requiring knowledge of the finer details of the VGDAs and VGSAs AlX support should be contacted as these structures Chapter 2 Problem determination and recovery 125 126 are really not customer serviceable It should be clear that a lot of damage can be done by attempting to manipulate them manually
321. o command preceded by five special tokens D 23 The mount command The following summarizes the options for the mount command mount Makes a file system available for use Usage mount f n Node o Options p r v VFSName t Type Device Node Directory Directory all a V generic_options special_mount_points a Mounts all file systems in the etc filesystems file with stanzas that contain the true mount attribute all Same as the a flag f Requests a forced mount during system initialization to enable mounting over the root file system n Node Specifies the remote node that holds the directory to be mounted 362 AIX Logical Volume Manager from A to Z Troubleshooting and Commands o Options Specifies options Options you enter on the command line should be separated only by a comma not a comma and a space The following file system specific options apply to JFS bsy Prevents the mount operation if the directory to be mounted over is the current working directory of a process log LVName Specifies the full path name of the file system logging logical volume name where the following file system operations are logged nodev Specifies that you cannot open devices from this mount This option returns a value of ENXIO if a failure occurs nosuid Specifies that execution of setuid and setgid programs by way of this mount is not allowed This option returns a
322. o the logical volume U Userid Specifies user ID for logical volume special file u UpperBound Sets the maximum number of physical volumes for new allocation The value of the Upperbound variable should be between one and the total number of physical volumes When using striped logical volumes or super strictness the upper bound indicates the maximum number of physical volumes allowed for each mirror copy Note When creating a superstrict logical volume you must specify physical volumes or use the u flag v Verify Sets the write verify state for the logical volume Causes y all writes to the logical volume to either be verified with a follow up read or prevents n the verification of all writes to the logical volume The Verify parameter is represented by one of the following n Prevents the verification of all write operations to the logical volume This is the default for the v flag y Causes the verification of all write operations to the logical volume Appendix A High level LVM commands 247 w MWCflag Mirror write consistency flag can be one of the following y Turns on mirror write consistency which insures data consistency among mirrored copies of a logical volume during normal I O processing n No mirror write consistency See the f flag of the syncvg command x Maximum Sets the maximum number of logical partitions that can be allocated to the logical volume The default
323. o use in place of a system generated name for the new logical volume Logical volume names must be unique system wide names and can range from 1 to 15 characters Y Prefix Specifies a prefix to use in building a system generated name for the new logical volume The prefix must be less than or equal to 13 characters A name cannot begin with a prefix already defined in the PdDv class in the Device Configuration Database for other devices or a name already used by another device Commands called copyrawlv getlvodm lquerylv lqueryvg mklv putlvodm and rmlv 210 m Note Do not copy from a larger logical volume containing data to a smaller one Doing so results in a corrupted file system because some data including the superblock is not copied It should be noted that this command will fail if cp1v creates a new logical volume and the volume group is varied on in concurrent mode The cplv command copies the contents of SourceLogicalVolume to a new or existing DestinationLogicalVolume The SourceLogicalVolume parameter can be a AIX Logical Volume Manager from A to Z Troubleshooting and Commands logical volume name or a logical volume ID The cplv command creates a new logical volume with a system generated name by using the default syntax The system generated name is displayed m Note 1 If you are copying a striped logical volume and the destination logical volume does not exist an identic
324. occurring 312 AIX Logical Volume Manager from A to Z Troubleshooting and Commands s StrictState Determines the strict allocation policy Copies of a logical partition can be allocated to share or not to share the same physical volume The Strict variable is one of the following y Sets a strict allocation policy so that copies of a logical partition cannot share the same physical volume default n Does not set a strict allocation policy so that copies of a logical partition can share the same physical volume S Sets a super strict allocation policy so that the partitions allocated for one mirror cannot share a physical volume with other partitions from another mirror t Type Sets the logical volume type jfs boot jfslog or paging u UpperBound Set the upperbound between 1 and 32 y CopyFlag Sets the copy flag z Size Sets the number of partitions LVID The logical volume ID putivodm An object file command that puts volume group values into the Configuration Database Usage o AutoOn k K q State v VGName VGID K For volume group VGID unlock the volume group k For volume group VGID lock the volume group o AutoOn For volume group VGID set the AutoOn flag y or n q State For volume group vgid set the State 0 Varied off 1 Varied on with all PVs 2 Varied on with missing PVs v VGName Add a new volume group VGID VGID The
325. ode Copyright IBM Corp 2000 ODM Object Database Manager PVID Physical Volume Identifier RAID Redundant Array of Independent Disks RDBMS Relational Database Management System RSCT Risc System Cluster Technology ROS Read Only Storage RVSD Recoverable Virtual Shared Disk SCSI Small Computer Systems Interface SSA Serial Storage Architecture VGDA Volume Group Descriptor Area VGID Volume Group Identifier VGSA volume group status area VMM Virtual Memory Manager VSD Virtual Shared Disk 409 410 AIX LVM System Management Volume 2 Index Symbols ODMDIR 112 etc environment 283 etc exclude rootvg 255 etc filesystems 10 78 83 154 337 345 etc magic 346 image data 256 ust include lvm h 78 ust include sys obdir h 124 usr include sys hd_psn h 121 usr include sys lvmrec h 122 A accounting 51 allocation group size 51 allocation map 25 30 allocation policy 24 30 31 42 inter physical 41 intra physical 41 strict 32 allocp 285 B backup 321 324 bad block relocation 25 41 boot 21 Cc C language 59 cfgvg 287 chfs 52 327 chlv 33 199 chlvcopy 267 287 chps 330 chpv 203 chvg 65 107 204 concurrent mode 6 copyrawly 289 cp 28 cpio 28 331 cplv 27 209 crash 87 crfs 51 335 Copyright IBM Corp 2000 CuDvDr 61 D deallocation 8 defragfs 338 Device Configuration Database 5 18 df 87 339 dfsck 341 disk failure 89 dspmsg 63 dump d
326. of commands in earlier versions of AIX 3 1 Replace a failed physical volume This first section describes how to replace a failing disk in AIX Version 4 3 2 and lower Assume that we have a system with one physical volume hdiskn which is part of a volume group asgard_vg and contains mirrored copies of logical volumes as well as parts of un mirrored logical volumes 3 1 1 Step 1 We will remove any copies of logical volumes that reside on the failed physical volume For example we have a logical volume called mirrorlv which had two copies one on hdiskn and one on hdiskm Therefore the new maximum number of logical partition copies is one rmlvcopy mirrorlv 1 hdiskn or use the following Copyright IBM Corp 2000 189 smit rmlvcopy da N Remove Copies from a Logical Volume Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields LOGICAL VOLUME name mirrorlv NEW maximum number of logical partition 1 copies PHYSICAL VOLUME names hdiskn F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Conmand F7 Edit F8 Image C F9 Shell FLO Exit Enter Do z Repeat step 1 for every logical volume that had copies on hdiskn 3 1 2 Step 2 We will now remove any un mirrored logical volumes that used partitions from the failed disk if any existed For example we will assume that we have a logical volume dudlv that was not mirrored and used partitions on hdiskn rmlv
327. of false or partial mirrors These were discussed in Sixteen zeros errors and other corrupt PVIDs in the VGDA on page 159 hdisks with greater than 1016 physical partitions and factor size 1 The other problem worth mentioning is that of physical volumes containing more than 1016 physical partitions or the equivalent multiple if factor size is used as the default VGSA on each disk can only track 1016 physical partitions Since AIX Version 4 2 the LVM commands were changed to prevent the creation of volume groups with hdisks exceeding this limit This will only be a problem if the system is running below this level or there is a volume group that was created at an earlier level of AIX If you have a volume group in this state since the VGSA is used to track the staleness of mirrors there may be a false and nonsensical report that a AIX Logical Volume Manager from A to Z Troubleshooting and Commands non mirrored logical volume is stale Alternatively there may be a false indication that a mirror copy has gone stale As well as providing a confusing view of the state of the LVM from the high level commands this symptom may cause some commands to fail migratepv may fail because migratepv briefly uses a form of mirroring to move a logical volume from one disk to another If the target logical partition is incorrectly considered stale then migratepv cannot remove the source logical partition so the command will fail in the middle of
328. ogical volume or a tape device For a network dump the device can be a host name and a path name q Suppresses all messages to standard output If this flag is used with the 1 r zor L flag the q command will be ignored r Host Path Frees space used by the remote dump file on server Host The location of the dump file is specified by the Path s Device Temporarily changes the secondary dump device to the specified device The device can be a logical volume or a tape device For a network dump the device can be a host name and a path name 382 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Determines if a new system dump is present If one is present a string containing the size of the dump in bytes and the name of the dump device will be written to standard output If a new system dump does not exist nothing is returned After the sysdumpdev z command is run on an existing system dump the dump will no longer be considered recent If no flags are used with the sysdumpdev command the default dump devices are used The sysdumpdev command changes the primary or secondary dump device designation in a system that is running The primary and secondary dump devices are designated in a system configuration object The new device designations are in effect until the sysdumpdev command is run again or the system is restarted If no flags are used with the sysd
329. olume 000008 32 1 2 8 Set write verify for a logical volume 2 00 33 1 2 9 Set schedule policy 0 eee ee 34 1 2 10 Increase the maximum size of a logical volume 35 1 2 11 Increase the size of a logical volume 4 37 1 2 12 List the configuration of a logical volume 38 1 2 13 List a logical volumes detailed mapping 41 1 3 Physical volume related commands 22 00005 43 1 3 1 List all physical volumes in system 0000005 44 1 3 2 List configuration of a physical volume 4 45 1 3 3 List contents of a physical volume 00000 47 1 3 4 Move data from a physical volume 0000000 49 1 4 Journaled file system related commands 00 05 50 1 4 1 Add a JFS to a previously defined logical volume menu 51 1 4 2 Change show details of a file system 00 52 1 4 3 Defrag a file system 0 2 0 cee ee 54 Copyright IBM Corp 2000 iii iv 1 4 4 Mount a file system 0 0 0 2 ee 55 1 4 5 Unmounta file system 2 0000000 e eee eee 57 Chapter 2 Problem determination and recovery 59 2 1 A methodology for problem determination 5 60 2 1 1 Breaking up high level commands 0 05 60 2 2 Producing debug output for LVM Scripts 0 5 62 2 3 Corruption example
330. olume group that is activated is available for use When a volume group is activated physical partitions are synchronized if they are not current A list of all physical volumes with their status is displayed to standard output whenever there is some discrepancy between the Device Configuration Database and the information stored in the Logical Volume Manager The volume group may or may not be varied on You must carefully examine the list and take proper action depending on each reported status to preserve your system s integrity A list of every status and its meanings can be found in the lvm_varyonvg subroutine While varying on in concurrent mode if the vary on process detects that there are logical volumes which are not previously known to the system their definitions are imported The permissions and ownership of the new device special files are duplicated to those of the volume group special file If you have changed the permissions and or ownership of the device special files of the logical volume on the node it was created you will need to perform the same changes on this node If the volume group cannot be varied on due to a loss of the majority of physical volumes a list of all physical volumes with their status is displayed To vary on the volume group in this situation you will need to use the force option AIX Logical Volume Manager from A to Z Troubleshooting and Commands The varyonvg command will fail to vary on the vo
331. olumes known volume groups and so on stored in the ODM It may be useful as an alternative to odmget and is very useful in passing parameters into the Iquery commands with constructs of the form lquerylv L getlvodm 1 mirrorstripelv p hdisk9 At lquerylv L getlvodm 1 mirrorstripelv p hdisk9 At D Iname mirrorstripelv VGid 170375251870e MexLP 512 MPolicy 5 MartConsist Ivstate 1 Csize 2 PPsize 22 Penmissicns Relocation 1 Wrverify 2 q n close 2 stripe exp 16 stripingwid 2 Nmtpies 2 HMiraCyy 0 IWAP 00091974e4f3d56e 218 1 Mtype 000174375251870e 4 1 00017d375243d402 219 IMAP 00017d37e671 51b 109 1 Mtype 000170375251870e 4 2 00017d0374dd8C762 109 na J It may also be useful in situations where extensive damage has occurred to large numbers of LVM objects and time and availability considerations make it desirable to script a solution rather than work through repair commands by hand A full list of its capabilities is documented in Appendix C ODM commands on page 315 108 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note Essentially the low level commands prefixed with 1 simply act as parsers and front ends onto the LVM API calls To gain a good understanding of these commands capabilities one may look up the relevant API call for example 1vm queryvgs for lqueryvgs These calls are listed in the AIX Version 4 3 Technical Reference Kernel an
332. ommand has options such as increasing the size of the VGDA and changing the physical partition factor that are not supported by SMIT A 4 The cplv command The following summarizes the options for the cplv command cplv Copies the contents of a logical volume to a new logical volume Usage cplv v VolumeGroup y NewLogicalVolume Y Prefix SourceLogicalVolume cplv copies the contents of a logical volume to an existing logical volume Usage cplv e DestinationLogicalVolume f SourceLogicalVolume Appendix A High level LVM commands 209 e Specifies that the DestinationLogicalVolume exists and that a new logical volume should not be created If the DestinationLogicalVolume is smaller than the SourceLogicalVolume the extra logical partitions are not copied When you use this flag any data already in the DestinationLogicalVolume is destroyed For this reason user confirmation is required unless the f flag is added The Type characteristic of the DestinationLogicalVolume must be copied to prevent inadvertently overwriting data To change the Type characteristic use the chlv command f Copies to an existing logical volume without requesting user confirmation v VolumeGroup Specifies the volume group where the new logical volume resides If this is not specified the new logical volume resides in the same volume group as the SourceLogicalVolume y NewLogicalVolume Specifies the name t
333. on people occasionally have trouble typing long strings of numbers when they are afraid that they have lost 132 AIX Logical Volume Manager from A to Z Troubleshooting and Commands their company s data Therefore we could also have run redefinevg d hdisk6 targetvg 3 redefinevg d hdisk6 targetvg We check the state of our volume groups and logical volumes lsvg rootvg targetvg lsvg o rootvg targetvg lsvg l targetvg targetvg LV NAME TYPE LPs PPs PVs LV STATE MOUNT POINT targetlog 2 2 1 open syncd N A targetlv 2 1 T 1 open syncd targetfs Gi J The damage is not completely repaired It is advised to review the current state of the ODM Chapter 2 Problem determination and recovery 133 CuAt name targetvg attribute vgserial_ id value 00017d3770963601 type R generic D rep n nls index 637 CuAt name targetvg attribute pv value 00017d375243d4020000000000000000 type R generic rep Wop nls _ index 0 CuDv name targetvg status 1 chgstatus 1 ddins un location parent connwhere PdDvin logical _volume vgsubclass vgtype CuDvDr resource ddins valuel targetvo value2 41 value CuDvDr resource devno valuel 41 value2 0 C value3 targetvg J We have re created the entries for the VGID PVID objects for physical volumes contain
334. on 4 2 or later o frag 512 1024 2048 4096 Specifies the JFS fragment size in bytes A file system fragment is the smallest unit of disk storage that can be allocated to a file The default fragment size is 4096 bytes 360 AIX Logical Volume Manager from A to Z Troubleshooting and Commands o frag 512 1024 2048 4096 Specifies the JFS fragment size in bytes A file system fragment is the smallest unit of disk storage that can be allocated to a file The default fragment size is 4096 bytes o compress no LZ specifies data compression If you do not want data to be compressed set this option to no Selecting compression requires a fragment size of 2048 or less o nbpi 512 1024 2048 4096 8192 16384 32768 65536 131072 Specifies the number of bytes per i node nbpi The nbpi is the ratio of file system size in bytes to the total number of i nodes The default nbpi value is 4096 bytes The values 32768 65536 and 131072 only apply to AIX Version 4 2 or later p Prototype Specifies the name of the prototype file Options specified on the command line override attributes in the prototype file s Size Specifies the size of the file system in 512 byte blocks The mkfs command makes a new file system on a specified device The mkfs command initializes the volume label file system label and startup block The Device parameter specifies a block devic
335. on one CD mkca gives instructions for CD replacement and removal until all the volumes have been created A 17 The mklv command The following summarizes the options for the mklv command Appendix A High level LVM commands 243 mklv Creates a logical volume Usage mklv a Position b BadBlocks c Copies d Schedule e Range i L Label m MapFile r Relocate s Strict t Type u UpperBound v Verify w MirrorWriteConsistency x Maximum y NewLogicalVolume Y Prefix S StripeSize U Userid G Groupid P Modes VolumeGroup NumLP PhysicalVolume s a Position Sets the intra physical volume allocation policy the position of the logical partitions on the physical volume The Position variable can be one of the following m Allocates logical partitions in the outer middle section of each physical volume This is the default position Allocates logical partitions in the center section of each physical volume e Allocates logical partitions in the outer edge section of each physical volume ie Allocates logical partitions in the inner edge section of each physical volume im Allocates logical partitions in the inner middle section of each physical volume b BadBlocks Sets the bad block relocation policy The Relocation variable can be one of the following y Causes bad block relocation to occur This is the default n Prevents bad block relocation from occurring
336. opy on a separate PV PV distribution Stripe width Stripe size Current state of allocation strict non strict or super strict A strict allocation states that no copies for a logical partition are allocated on the same physical volume If the allocation does not follow the strict criteria is called non strict A non strict allocation states that at least one occurrence of two physical partitions belong to the same logical partition A super strict allocation states that no partition from one mirror copy may reside the same disk as another mirror copy The distribution of the logical volume within the volume group The physical volumes used the number of logical partitions on each physical volume and the number of physical partitions on each physical volume are shown The number of physical volumes being striped across The number of bytes per stripe A 10 The Ispv command The following summarizes the options for the 1spv command 224 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Ispv Displays information about a physical volume within a volume group Usage Ispv L Il pl M n DescriptorPhysicalVolume v VGID PhysicalVolume L Specifies no waiting to obtain a lock on the Volume group Note If the volume group is being changed using the L flag gives an unreliable date Lists the following fields for each logical volume on the physical volume LVname
337. or in rootvg if that flag is not used If mkcd creates the file system it is removed upon command completion unless either the Ror s flag is used If the R or s flag is used consideration must be made for adequate file system directory or disk space especially when creating multi volume CDs i image data Specifies the user supplied image data file This data file takes precedence over the image data file in the mksysb image If you do not give the i flag then mkcd restores the image data from the given mksysb image or generates a new image data file during the creation of mksysb Note The i flag cannot be used to specify a user supplied vgname data file for use with a savevg image package_list Specifies the file containing a list of additional packages you want copied to the usr Ipp inst images directory of the CD file system The images are copied from the location named with the p flag If you use the 1 flag you must also use the p flag M mksysb_target States the directory or file system where the mksysb or savevg image is stored if a previously created backup is not given with the mor s flags If the m flag is not used anda mksysb or savevg image is not provided mkcd verifies that mkcd mksysb_image exists If the directory does not exist then mkcd creates a separate file system mkcd mksysb_image where the mksysb or savevg images are temporarily stored The command creates the file syste
338. ore we suspect VGDA corruption We can use 1svg n to target the different VGDAs in the volume group to find which are damaged and which are intact i N lsvg n hdisk6 Segmentation fault coredump lsvg n hdisk7 VOLUME GROUP lowvg VG IDENTIFIER 00017d3780467f 5b VG STATE active PP SIZE 4 megabyte s VG PERMISSION read write TOTAL PPs 2150 8600 megabytes MAX LVs 256 FREE PPs 2142 8568 megabytes LVs 3 USED PPs 8 32 megabytes OPEN LVs 0 QUORUM 2 TOTAL PVs 2 VG DESCRIPTORS 3 STALE PVs 0 STALE PPs 0 ACTIVE PVs 0 AUTO ON yes MAX PPs per PV 2032 MAX PVs 16 w J We have identified the VGDA s on hdiske as corrupt and the one on hdisk7 as good Our first priority is to get maps for our logical volumes out of the good VGDA lspv M n hdisk7 hdisk6 hdisk6 1 215 hdisk6 216 lowlv1 1 hdisk6 217 lowlv1 2 hdisk6 218 lowlv1 3 hdisk6 219 lowlv2 1 2 hdisk6 220 lowlv2 2 2 hdisk6 221 1075 We can turn these into maps suitable for mklv m with the following commands 148 AIX Logical Volume Manager from A to Z Troubleshooting and Commands E lspv M n hdisk7 hdisk6 grep lowlvl awk print 1 gt lowlv1 map lspv M n hdisk7 hdisk6 grep lowlv2 awk print 1 gt lowlv2 map cat lowlv1 map hdisk6 216 hdisk6 217 hdisk6 218 cat lowlv2 map hdisk6 219 hdisk6 220 I J We could aim to reduce and re create the VGDA as we did in Corruptio
339. ormation on LVM kernel internals may be found in the AIX Version 4 3 Kernel Extensions and Device Support Programming Concepts SC23 4125 in Chapter 10 and LVM and information on investigating these structures is found in Chapter 17 Debugging Tools of the same document 2 6 11 1 Journaled file system enhancements Before the file system enhancements of AIX 4 3 3 it was sometimes necessary to use crash to isolate file system corruption These enhancements are available for lower levels of AIX with the following APARs AIX level APAR 4 1 5 1X83878 IX87162 4 2 1 1X82819 1X84402 IX89392 4 3 2 IX86362 IX89598 When the file system enhancements are in place an entry will be generated in the error log when file system corruption is detected This entry will be of the form Chapter 2 Problem determination and recovery 127 Label JFS_FSCK REQUIRED Class OTHER Type INFO Description FILE SYSTEM RECOVERY REQUIRED Recommended Actions PERFORM FULL FILE SYSTEM RECOVERY USING FSCK UTILITY Detail Data MAJOR MINOR DEVICE NUMBER FILE SYSTEM DEVICE AND MOUNT POINT x S This informational entry will usually be paired with a diagnostic entry such as KERNHEAP LOW LOG WRITE ERR META CORRUPTION META EXCEPTION Of META WRITE ERR As a final note on inspection commands the investigator should take care to focus on the problem and no
340. orrect we begin searching the remaining physical partitions for a candidate for the next logical partition position Again this depends on understanding data well enough to perform the integrity checks and this technique is not possible for some types of data organization In the case of JFS this cannot be accomplished with fsck since fsck treats the file system as atomic All partitions must be present before integrity checks are performed Thus we cannot tell how many allocation groups or inodes a fsck has advanced or failed to advance for each candidate The data organization issue we face here is standard to BSD file system implementations JFS does however allow mounting a file system containing one LP While many files will not be accessible you may be able to recover at least some data A final suggestion is that if a disk has been in a volume group in the past and then reducevg was run a stale copy of the VGDA will still be on that disk as reducevg only clears the LVM information record and MWCC areas However given the complexity of rewalking techniques it is strongly recommended that backups are used instead This will provide much faster availability of data and also guarantees integrity when the backup is good 2 7 2 1 VGSAs The main problem area with mirrored logical volumes whether mirrored explicitly by the user or internally by LVM commands such as migratepv which are only noticeable when they fail is the case
341. orruption we ran cfgmgr too early varyoffvg lowvg exportvg lowvg importvg y lowvg hdisk7 usr sbin redefinevg 215 14446 Segmentation fault coredump lowvg After running the 1deletepv the correct procedure is varyoffvg lowvg cp etc filesystems tmp filesystems bak exportvg lowvg importvg y lowvg hdisk7 lowvg cfomgr 1l ssar extendvg lowvg hdisk6 Note that we also make a copy of etc filesystems before we run an exportvg since we will lose the entries for lowfs1 and lowfs2 and we do not have any copies of lowlv1 s logical control block on hdisk7 this is not a bigVG As a brief aside if you tried to script the above commands you may see a strange albeit harmless error when you run the exportvg after the 1deletepv and varyoffvg ldeletepv g 00017d3785f9cc2d p 00017d377c3abace varyoffvg lowg varyoffvg lowvg exportvg lowvg exportvg lowvg 0516 008 lqueryvg LVM system call returned an unknown error code 97 0516 932 usr sbin syncvg Unable to synchronize volume group lowvg _ importyg y lowvg hdisk7 E 154 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note that LVM sometimes does perform control actions asynchronously This is not a problem in normal LVM operation However it is a good reason not to script using low level commands Now after the extendvg is run the LVM is in the following state
342. ort this logical volume in earlier versions of AIX A 19 The mksysb command The following summarizes the options for the mksysb command mksysb Creates an installable image of the root volume group either in a file or onto a bootable tape Usage mksysb b Number e p v X i m Device File b Number Specifies the number of 512 byte blocks to write in a single output operation When the backup command writes to tape devices the default is 100 for backups by name The write size is the number of blocks multiplied by the block size The default write size for the backup command writing to tape devices is 51200 100 512 for backups by name The write size must be an even multiple of the tape s physical block size 254 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Excludes files listed in the etc exclude rootvg file from being backed up The rules for exclusion follow the pattern matching rules of the grep command Note If you want to exclude certain files from the backup create the etc exclude rootvg file with an ASCII editor and enter the patterns of file names that you do not want included in your system backup image The patterns in this file are input to the pattern matching conventions of the grep command to determine which files will be excluded from the backup If you want to exclude files listed in the etc exclude rootvg file select the Excl
343. orted for all file systems mountable Causes the fsck command to exit with success returning a value of 0 if the file system in question is mountable clean If the file system is not mountable the fsck command exits returning with a value of 8 mytype Causes the fsck command to exit with success 0 if the file system in question is of the same type as either specified in the etc filesystems file or by the v flag on the command line Otherwise 8 is returned For example fsck o mytype V jfs exits with a value of 0 if the root file system is a journaled file system Does not display messages about minor problems but fixes them automatically This flag does not grant the wholesale license that the y flag does and is useful for performing automatic checks when the system is started normally You should use this flag as part of the system startup procedures whenever the system is being run automatically Also allows parallel checks by group t File Specifies a File parameter as a scratch file on a file system other than the one being checked if the fsck command cannot obtain enough memory to keep its tables If you do not specify the t flag and the fsck command needs a scratch file it prompts you for the name of the scratch file However if you have specified the p flag the fsck command is unsuccessful If the scratch file is not a special file it is removed when the fsck command ends 342 AIX Logic
344. oth mirrors on the same disk is used in which case a simple rmlvcopy should be run The 0516 304 error may also be caused by passing an incomplete map to reducevg df Another cause is rmlvcopy removing some partitions from one drive and then discovering the remaining partitions for a logical volume are off line when no drive was specified These can both be fixed simply by removing the remaining partitions In the past some LVM defects have also caused the 16 zeros symptom Two disks having the same PVID There have been occasions when people have used dd as an unsupported method to back up a disk This is not recommended Sometimes this has led to people running two disks with the same PVID on a machine This can obviously confuse the LVM As an example we create a volume group ddvg and then back it up with dd no logical volumes in the volume group We now run rmdev cfgmgr and 1spv E N mkvg ft2 y ddvg hdisk6 ddvg lspv grep ddvg hdisk6 0001703 78bb9 828 ddvg dd if dev hdisk6 bs 32768 of dev hdisk7 rmdev dl hdisk7 hdisk7 deleted cfomgr 1 ssar lspv grep ddvg hdisk7 0001703 78bb9 828 ddvg hdisk6 0001703 78bb9 828 ddvg mklv y ddlv ddvg 1 ddlv varyoffvg ddvg exportvg ddvg Ne D Assume the original hdisk hdisk6 goes off line and we import the volume group using the VGDA on hdisk7 We have now lost touch with our control data If any data had been stored in dd
345. oubleshooting and Commands We have found something but it is probably a good idea to go on checking in case something else is wrong We check cubvpr next ls 1 dev bad if we had found anything with the last comand we might try using the major and minor mirbers to narrow down our search of GiDvDr omt q resource devno GuDvDr many entries of the fam QDvDr resource devno valuel 11 value2 0 value3 pcil but nothing relevant cmt q resource ddins GiDvDr again many entries but nothing is particularly unusual Dvdr resource dins valuel sysram value2 o value3 Li J Note that we also looked for entries in the dev directory It seems that we have some corruption in CuDv We try odmdelete o CuAt q name badvg However this returns 0518 307 odmdelete 0 objects deleted There is some strangeness here We re run the odmget redirecting the output to a file odmget q PdDvLn logical _volume vgsubclass vgtype CuDvl gt PdDvLn CuDv and now edit the file with vi Scrolling down to the entry for badvg Chapter 2 Problem determination and recovery 73 ddins location parent connwhere PdDvin logical_volumme vgsubclass vgtype name bad1 Hvg status 0 chgstatus 1 ddins un location parent connwhere PdDvin logical _volume vgsubclass vgtype AS S the source of the corru
346. ount 4399 of dev hdisk6 dd if dev zero count 4399 of dev hdisk7 lsvg M allocpvg Physical volume is not a volume group member Check the physical volume name specified 0516 066 Once problem determination has been performed and we can confirm there are indeed no good copies of the VGDA we turn as a last resort to the SMIT logs 7 Sep Sep klv y al Sep mklv y al Sep klv y al cat smit script 30 1999 16 19 33 30 1999 16 20 20 locplv1 allocpvg 10 30 1999 16 20 42 locplv2 allocpvg 15 30 1999 16 21 13 locplv3 e x allocpvg 5 mkvg f y allocpvg s 8 hdisk6 hdisk7 In a real life situation the SMIT logs will probably not be this neat and may need to be pre processed either by hand or with a UNIX text processing tool In this case however we can simply clean up the stale ODM entries for allocpvg then chmod the smit script so that it is executable and then run it exportvg chmod x allocpvg allocplv1 allocplv2 allocplv3 allocplv4 Ne Va varyoffvg allocpvg allocpvg smit script smit script mlv Logical volume allocplv1 is removed Chapter 2 Problem determination and recovery 165 166 In this case we can verify that our test data is recovered dd skip 1 if dev allocplv2 head 1 allocplv2_data dd skip 1 if dev allocplv3
347. out The name is formed by concatenating the volume group prefix vg to the next two integer sequence number available from the CuDv class If the n option is used then the CuDv class will be checked to make sure the name does not already exist B 9 The Ichangelv command The following summarizes the options for the 1changelv command Ichangelv An object file command that changes the attributes of a logical volume Usage Ichangelv I LVID s MaxLVsize n LVname M MirrorPolicy p Permissions r Relocation v WriteVerify w MWConsistency Appendix B Intermediate level commands 293 I LVID The LVID of the logical volume to change s MaxLVsize New maximum size of logical volume n NewLVname New logical volume name The size of the string is between 1 and 64 characters M MirrorPolicy New mirror policy The value is as follows 1 Sequential mirroring 2 Parallel mirroring p Permission Set whether the logical volume is read write or read only The value is as follows 1 Read write 2 Read only r Relocation Sets whether there will be bad block reallocation The value is as follows 1 Bad block reallocation 2 No bad block reallocation v WriteVerify Sets the write verify flag The value is as follows 1 Write verify on 2 Write verify off w MWConsistency Sets the mirror write consistency flag The value is as follow
348. owed if the volume group is varied on in concurrent mode chlv This script changes the characteristics of the logical volume s in the list logical volume IDs can also be used Usage chlv a Position b BadBlocks d Schedule e Range G Groupid L label P Modes p Permission r Relocate s Strict t Type U Userid u Upperbound v Verify w MirrorWriteConsistency x Maximum LogicalVolume s Copyright IBM Corp 2000 199 a Position Sets the intraphysical volume allocation policy the position of the logical partitions on the physical volume The Position variable is one of the following m Allocates logical partitions in the outer middle section of each physical volume This is the default position c Allocates logical partitions in the center section of each physical volume e Allocates logical partitions in the outer edge section of each physical volume ie Allocates logical partitions in the inner edge section of each physical volume im Allocates logical partitions in the inner middle section of each physical volume b BadBlocks Sets the bad block relocation policy The BadBlocks variable is one of the following y Allows bad block relocation to occur default n Prevents bad block relocation from occurring d Schedule Sets the scheduling policy when more than one logical partition is written The scheduling policy must be either parallel o
349. p on separate lines when used with the d flag A Returns all static attributes and the partition map Appendix B Intermediate level commands 301 Returns the partition map in lreducelv format Includes the appropriate tags in the output B 21 The Iquerypv command The following summarizes the options for the 1querypv command Iquerypv An object file command that queries the attributes of a physical volume Usage Iquerypv p PVID g VGID N PVName s c P n a D d A t p PVID The PVID of the physical volume to query g VGID The VGID of the volume group that contains the physical volume to query N PVName The physical volume name If both the physical name and VGID are provided the VGID is ignored S Returns the physical volume partition size C Returns the current state of the physical volume 1 Active 2 Missing P Returns the total number of physical partitions n Returns the number of allocated physical partitions a Returns all static attributes of the physical volume PPSize PPState TotalPPs AllocPPs VGDAs D Returns the number of VGDAs on the physical volume d Returns the physical partition map of the physical volume A Returns all attributes of the physical volume that is the static attributes and the physical partition map t Adds a short description when use with the above options B 22
350. physical volume s that will make up the volume group These physical volumes cannot be part of another volume group Other physical volumes can be added at a later date Activate volume group automatically at system restart Some administrators may not want to automatically start some volume groups particularly in HACMP setups Volume Group major number The system will assign the next available major number if left blank Concurrent Capable Concurrent Mode Used only for concurrent volume groups and requires further software installed See the concurrent access volume group description in AlX Logical Volume Manager from A to Z Introduction and Concepts SG24 5432 AIX Logical Volume Manager from A to Z Troubleshooting and Commands SMIT mkvg N Add a Volume Group Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields VOLUME GROUP name keovg Physical partition SIZE in megabytes 16 PHYSICAL VOLUME names hdisk11 Activate volume group AUTOMATICALLY yes at system restart Volume Group MAJOR NUMBER Create VG Concurrent Capable no Auto varyon in Concurrent Mode no F1l Help F2 Refresh F3 Cancel F4 List Esc 5 Reset Esc 6 Command Esc 7 Edit Esc 8 Image Esc 9 Shel1 Esc 0 Exit Enter Do Ne S 1 1 2 Activate a volume group The varyonvg command activates the volume group specified by the VolumeGroup parameter and all associated logical volumes A
351. physical volume was hot swappable run cfgmgr Chapter 3 Replacing a drive in a mirrored configuration 191 3 1 6 Step 6 3 1 7 Step 7 192 to bring the new disk into the configuration database and then lspv to confirm that there is a new hdiskn Because we deleted the old hdiskn from the configuration database the name hdiskn is now available and it will be reassigned to the new disk If another disk with a lower hdisk number had been previously removed then this number will be assigned to the new disk This can be got around by creating a dummy disk on the lower number For example mkdev d c disk s ssar p ssar t hdisk w 11111111 will create a dummy disk to fill the vacant hdisk number Now add the new hdisk2 to the volume group asgard_vg extendvg asgard vg hdiskn or refer to 1 1 7 Add physical volume to an existing volume group on page 12 Add the mirror copies back for each logical volume that originally used physical partitions on the failed physical volume For example we had logical volume mirrorlv which used to have two copies one was on hdiskn mklvcopy mirrorlv 2 hdiskn or refer to 1 2 5 Add a mirrored copy to a logical volume on page 29 Repeat step 7 for each mirrored logical volume m Note To retain the original placement use map files when creating the new mirror or add the logical volumes onto the replacement physical volume in the order in which they were originall
352. point for the file system Displays statistics in units of 1024 byte blocks Displays the mount point information for the file system in the second column Appendix D Other related commands 339 Displays information on the file system in POSIX portable format When the r flag is specified the header line appears similar to Filesystem 512 blocks Used Available Capacity Mounted on n If the k flag is specified in addition to the r flag the column heading 512 blocks is replaced by the heading 1024 blocks File system statistics are displayed on one line in the following order FileSystem TotalSpace UsedSpace FreeSpace UsedPercentage MountPoint Gets file system statistics from the VFS specific file system helper instead of the statfs system call Any arguments given when using the s flag must be a journaled file system mount point or device The file system must also be listed in etc filesystem Includes figures for total allocated space in the output Displays all information for the specified file system 340 The df command displays information about total space and available space on a file system The FileSystem parameter specifies the name of the device on which the file system resides the directory on which the file system is mounted or the relative path name of a file system The File parameter specifies a file or a directory that is not a mount point If the File paramet
353. port e Copy of the customized Object Data Manager ODM database Appendix D Other related commands 379 Trace file e User environment Amount of physical memory and paging space Device and attribute information Security user information The output of the snap g command is written to the tmp ibmsupt general general snap file The snap command checks for available space in the tmp ibmsupt directory the default directory for snap command output You can write the output to another directory by using the d flag If there is not enough space to hold the snap Command output you must expand the file system Each execution of the snap command appends information to previously created files Use the r flag to remove previously gathered and saved information Note Press the Ctrl C key sequence to interrupt the snap command A prompt will return with the following options Press the Enter key to return to current operation press the S key to stop the current operation press the Q key to quit the snap command completely D 29 The sync command The following summarizes the options for the sync command sync Updates the i node table and writes buffered files to the hard disk Usage sync The sync command runs the sync subroutine If the system must be stopped run the sync command to ensure file system integrity The sync command writes all unwritten system buffers to disk including modified i
354. possible to locate the areas within the scripts where these actions occur m Note The use of portions of LVM scripts in this document does not imply that the contents of LVM scripts is fixed The LVM scripts are subject to modification at different levels of AIX or even of PTF 2 2 Producing debug output for LVM Scripts Since we are dealing with standard shell scripts here it is possible to trace these scripts in the normal fashion for korn shell scripts We may sh xv importvg y prod datavg hdisk2 from a shell prompt v will display each line as the shell reads it therefore displaying comments functions definitions and so on The x flag will display each line as it actually executes This will give the input values of variables passed in to the script line followed by its 62 AIX Logical Volume Manager from A to Z Troubleshooting and Commands evaluation This allows tracking the execution of the target script but will not descend into sub shells created when other scripts are started from within the script The behavior of korn shell debug output is also such that it will not produce this output for functions within a script It is necessary to insert set xv into the body of the function to produce this output set xv could also be used to display debugging information for sub shells of other high level commands However it is generally sufficient to run sh xv on each high level command to produce the required out
355. ption is immediately apparent We have an embedded control character in the ODM This ODM object is named badi Hvg Note that looking at this file with pg would not have shown us the problem pg does not display control characters like this There are alternative ways to check for this type of corruption odmget output may be piped through commands such as od followed by a grep to check for backslashes to check for control characters Alternatively odmget q name LIKE ba CuDv would have pulled out the corrupt entry however this is clearly not as reliable as the detailed procedure above So to delete the above entry we can run odmget q name LIKE ba CuDv CuDv name badvg status 0 chgstatus 1 ddins n location parent connwhere PdDvin logical _volume vgsubclass vgtype that was to check we would only delete this one entry odmdelete o CuDv q name LIKE ba 0518 307 odmdelete 1 objects deleted lsvg rootvg datavg N S 74 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 1svg now runs clean savebase should now be run in case the mini ODM is also corrupt It would also be a good idea to check that we don t have any corrupted file names in the dev directory The above procedure may seem somewhat long winded and certainly with experience could be performed more quickly Note When tackling an LVM problem as much data as possib
356. put Since editing the high level commands is an undesirable method of problem determination or LVM investigation this book provides the unsupported trclvm script Appendix E 1 trclvm on page 393 which makes a copy of a high level command instruments it with set xv lines and runs it The script command can be used to capture the large amounts of data that are generated when running high level commands in debug mode to a log file for later examination By default this will produce a typescript file containing input and output standard and error to the terminal running the command The embedded control M s in the typescript file displayed by vi may be removed with the vi command s m use control V control M to produce m in vi command mode We can look at debug output for the code fragment shown above to add a new volume group to the ODM by running sh xv mkvg y testvg hdisk9 and capturing the output with script The relevant portion of the typescript file is shown below Add new volume group to odm putlvodm CFLAG C VAL XFLAG X_VAL v VGNAME o AUTO ON q 0 VGID putlvodm v testvg o y q 0 00017d3700bfbf 5c if 0 If putlvodm failed output warning and continue then dspmsg s 1 cmdlvm cat 1024 1024 n mkvg gt amp 2 fi 01 0 Here putlvodm runs and places the volume group identifier VGID 00017d3700bf bf 5c into the ODM We can see this operation has succeeded with the
357. r 175 180 used outer middle mirror lv jfs N A 181 186 stale outer middle mirror lv jfs N A 187 215 free outer middle 216 322 free center 323 429 free inner middle 430 537 free inner edge F1 Help F2 Refresh F3 Cancel F6 Command F8 Image F9 Shel1 F1O Exit Find n Find Next 1 3 4 Move data from a physical volume The data in allocated physical partitions can be moved from one physical volume to one or more physical volumes as long as they are in the same volume group This process can be limited to the physical partitions belonging to one logical volume otherwise all physical partitions will be moved For the command line options see migratepv in Appendix A 14 The migratepv command on page 235 or use the SMIT fastpath smit migratepv SMIT provides the following fields Source physical volume The name of the physical volume from which to move the data Destination physical volumes The destination physical volume or volumes Move only data belonging to this logical volume If a logical volume is specified only physical partitions used by this logical Chapter 1 LVM commands 49 volume will be moved else all physical partitions will be moved smit migratepv a Move Contents of a Physical Volume Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields SOURCE physical volume name hdisk9 DESTINATION physical volumes Move only data belo
358. r 3 Replacing a drive in a mirrored configuration 189 3 1 Replace a failed physical volume 0 00 00 ee eee 189 Bide 1 Step Ve acne teed He Bee bee eee abe Aa tee Shi bat 189 Bil Step 2s 8 eee ha OE deal ee aa etna hs 190 313 Siep 3 oire dee Hea ea Pate ae ea ela ea ea 191 Bid Step 4s ona ra ae eae ined i A a 191 31 5 Step neeaae A ba ee hohe dete eae dk eee ane de 191 Se Step Onc og ee ne BaP he Se de deeded aan es 192 SLZ SUP 0225 teed Genk toa ea oa ened eh eave ala eee aa ees 192 3 128 Step 8 20 need eee ee a aed EG ee aaa Be 193 S19 Step Oia end ete vas bee ae ba ee aba eee an a 193 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 3 2 Using the replacepv command 0 000 e eee eee 193 3 2 1 Description of the test environment 0 193 Appendix A High level LVM commands 199 A 1 The chlv command 0 00 cee eee 199 A 2 The chpv command 000 c eee ete eee 203 A 3 The chvg command 00 cee tee 204 A 4 The cplv command 00 eee 209 A 5 The exportvg command 0 00 eee 211 A 6 The extendlv command 0 00 ees 212 A 7 The extendvg command 0 eee eee 215 A 8 The importvg command 0 2 ee ee 216 A 9 The Islvcommand 0 eee 219 A 10 The Ispv command 0 0 00 eects 224 A 11 The Isvg command 00 0c eee tee 229 A 12 The Isvgfs commande ser
359. r Do J 36 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 1 2 11 Increase the size of a logical volume The extendlv command increases the number of logical partitions allocated to the logical volume by allocating an additional number of logical partitions This can be limited to specific physical volumes or all physical volumes can be considered to be available m Note The e m s and u flags are not valid with a striped logical volume By default the logical volume is expanded using the existing characteristics These can be temporarily overridden by specifying different allocation policies However the characteristics of the logical volume do not change For the command line options see extendlv in Appendix A 6 The extendlv command on page 212 or use the SMIT fastpath smit extendlv SMIT provides the following fields Logical volume name The name of the logical volume to be extended F4 for list Number of additional logical partitions The number of logical partitions to add to the logical volume Physical volume names This is used to specify which physical volumes will be used else all physical volumes will be considered to be available Position on physical volume Sets the intra physical volume allocation policy for this allocation alone The policy of the logical volume is not changed Range of physical volumes Sets the inter physical volume allocation policy for
360. r from A to Z Troubleshooting and Commands Reports the current state of the file system and the state that would result if the defragfs command is run without either qor r flag The defragfs command increases a file system s contiguous free space by reorganizing allocations to be contiguous rather than scattered across the disk You can specify the file system to be defragmented with the Device variable the path name of the logical volume for example dev hd4 You can also specify it with the FileSystem variable which is the mount point in the etc filesystems file The defragfs command is intended for fragmented and compressed file systems However you can use the defragfs command to increase contiguous free space in non fragmented file systems You must mount the file system read write for this command to run successfully Using the qflag or the r flag generates a fragmentation report These flags do not alter the file system D 7 The df command The following summarizes the options for the df command df Reports information about space on file systems Usage df P l M i t v k s FileSystem s File s Displays the number of free and used i nodes for the file system this output is the default when the specified file system is mounted Displays information on the total number of blocks the used space the free space the percentage of used space and the mount
361. r logging thresholds were reset Chapter 2 Problem determination and recovery 91 92 LABEL LVM_MISSPVADDED ENTIFIER 26120107 Date Time Sat Sep 11 17 09 46 Sequence Number 50 Machine Id 00017D374C00 Node Id itsosrv1 Class S Type UNKN Resource Name LIBLVM Description PHYSICAL VOLUME DEFINED AS MISSING Probable Causes POWER DRIVE ADAPTER OR CABLE FAILURE Detail Data MAJOR MINOR DEVICE NUMBER 001E 2008 SENSE DATA 0000 0000 0000 0000 0000 0000 0000 0000 LABEL SSA LINK OPEN IDENTIFIER 625E6B9A LABEL REBOOT _ID ENTIFIER 2BFA76F6 Neni J We can identify the disk associated with the Lvm MISPVADDED by taking the major and minor numbers from the errpt output and converting from hexadecimal to decimal a N 2008 8200 quit 1s 1 dev grep 8200 brw 1 root system 30 8200 Sep 23 16 16 This gives us a major number 30 and minor number 8200 hdisk9 We can check if a disk is actually available with the 1sdev command as discussed in Checking device availability on page 98 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Now that we understand the situation we can regain access to our data with a forced varyonvg An unforced varyonvg will fail since we have quorum turned off The fact that we have a majority of VGDAs will not be considered varyonvg workvg varyonvg workvg
362. r of copies then reduce the mirrors manually with the rmlvcopy command If this option is not used the copies will default to 1 Commands called allocp getlvname getlvodm lcreatelv lextendlv lquerylv lquerypy lqueryvg putlvcb putlvodm and mlv 278 The unmirrorvg command unmirrors all the logical volumes detected ona given volume group This same functionality may also be accomplished manually if you execute the mlvcopy command for each individual logical volume in a volume group By default unmirrorvg will pick the set of mirrors to remove from a mirrored volume group If you wish to control which drives no longer are to contain mirrors you must include the list of disks in the input parameters PhysicalVolume When the PhysicalVolume parameter is listed in the command this indicates that only the mirrors that exist on this disk should be unmirrored Mirrors that exist on other drives in the volume group but not listed in a user provided disk list are left alone and remain mirrored Note that if a logical volume copy spans more than one disk the portion of the logical volume copy that resides on a disk not listed by the user is also removed When unmirrorvg is executed the default COPIES value for each logical volume becomes 1 If you wish to convert your volume group from triply mirrored to doubly mirrored use the c option Note that to use this command you must either have root user authority or be
363. r sequential as follows p Establishes a parallel scheduling policy default s Establishes a sequential scheduling policy e Range Sets the interphysical volume allocation policy the number of physical volumes to extend across using the volumes that provide the best allocation The value of the Range variable is limited by the Upperbound variable set with the u flag and is one of the following x Allocates logical partitions across the maximum number of physical volumes m Allocates logical partitions across the minimum number of physical volumes 200 AIX Logical Volume Manager from A to Z Troubleshooting and Commands G Groupid Specifies group ID for the logical volume special file L Label Sets the logical volume label The maximum size of the Label variable is 127 characters p Permission Sets the access permission to read write or read only The Permission variable is represented by one of the following w Sets the access permission to read write default r Sets the access permission to read only P Modes Specifies permissions file modes for the logical volume special file r Relocate Sets the reorganization flag to allow or prevent the relocation of the logical volume during reorganization The Relocate variable is one of the following y Allows the logical volume to be relocated during reorganization If the logical volume is striped the chlv command will not let you ch
364. rd indicates it is not a member of the specified VG define LVM_PVINVG 12 the PV s LVM record indicates it is PV status values which can be returned from lvm varyonvg in addition to the state values of LVM_PVMISSING LVM_PVREMOVED LVM_PVACTIVE in the varyonvg output structure These are returned only if a quorum is obtained or if a request has been made to override the no quorum error they may be returned if any volume group descriptor area copy has been obtained varyonvg output structure if a quorum is not obtained Error return a member of the specified VG Figure 8 Return values from varyonvg Figure 8 shows the other potential return codes for varyonvg These are taken from usr include lvm h This file serves as a useful reference in LVM problem determination At this point assume that the exportvg importvg combination is wrongly executed perhaps because the investigator knows that the ODM contains PVIDs but not that they shadow the values owned by LVM on disk Note that before we do this an entry is present in etc filesystems for workfs 78 AIX Logical Volume Manager from A to Z Troubleshooting and Commands grep p workfs etc filesystems workfs dev dev 1v02 vfs jfs log dev loglv00 mount false account false a J Now we run the importvg exportvg combination N varyoffvg workvg exportvg workvg importvg workvg
365. re This is standard We now perform the usual parsing of flags and of command line input 172 AIX Logical Volume Manager from A to Z Troubleshooting and Commands N Parse command line options PROG S0 set up program name for error messages PROG tmp 1vmtrace16090 syncvg set getopt P Hfivlp getopt P Hfivlp 1 stalelv set l stalelv if 0 Determine if there is a syntax error then dspmsg s 1 cmdlvm cat 1145 lvmmsg 1145 n PROG gt amp 2 exit fi 0 0 FFLAG IFLAG VFLAG LFLAG PFLAG CASE PFLAG HFLAG FFLAG IFLAG VFLAG LFLAG PFLAG CASE PFLAG Lot HFLAG Dy All flags are reset We now fill the flag variables Chapter 2 Problem determination and recovery 173 while 1 While there is a command line option do case 1 in specifies the names are read from standard in i IFLAG i shift synchronizes even the non stale partitions f FFLAG f shift synchronizes the indicated volume group v VFLAG v shift CASE y synchronizes the indicated logical volume 1 LFLAG 1 shift CASE y synchronizes the indicated physical volume p PFLAG p shift CASE y NUMBER of P if 2 lt 1 o 2 gt 32 then dspmsg s 1 cmdlvm cat 1148 lvmmsg 1148 gt n PROG gt amp 2 dspmsg s 1 cmdlvm cat 1145 lvmmsg 1145 gt n PROG
366. repair efforts and tries to avoid actions that might result in the loss of valid data In certain cases however the fsck command recommends the destruction of a damaged file If you do not allow the fsck command to perform the necessary repairs an inconsistent file system may result Mounting an inconsistent file system may result in a system crash If you do not specify a file system with the FileSystem parameter the fsck command checks all file systems listed in the etc filesystems file for which the check attribute is set to true You can enable this type of checking by adding a line in the stanza as follows check true The fsck command can perform simultaneous checks on multiple file systems This procedure can reduce the time required to check a large number of file systems Use a minus sign to separate the file systems when specified as part of the argument You can also perform simultaneous checks on multiple file systems by grouping the file systems in the etc filesystems file To do so change the check attribute in the etc filesystems file as follows check Number The Number parameter tells the fsck command which group contains a particular file system File systems that use a common log device should be placed in the same group Each group is checked in a separate parallel Appendix D Other related commands 351 process File systems are checked one at a time in the order that they are listed in the etc filesyste
367. rgvg SMIT provides the following fields Volume group name Name of the volume group to be reorganized F4 for list Logical volume names The names of the logical volumes to be included if only these ones are to be reorganized The order of priority will be in the order in which they are listed F4 for list Chapter 1 LVM commands 15 smit reorgvg Va Reorganize a Volume Group Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields VOLUME GROUP name park vg LOGICAL VOLUME names F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Conmand F7 Edit F8 Image g F9 Shell F10 Exit Enter Do 1 1 10 Listing the volume groups on the system The 1svg command displays information about volume groups If you use no parameters a list of the names of all defined volume groups is displayed For the command line options see 1svg in Appendix A 11 The Isvg command on page 229 or use the SMIT fastpath smit 1svg 16 AIX Logical Volume Manager from A to Z Troubleshooting and Commands smit lsvg gt N List Volume Groups Move cursor to desired item and press Enter List All Volume Groups List Contents of a Volume Group List All Logical Volumes by Volume Group F1 Help F2 Refresh F3 Cancel F8 Image F9 Shel1 F10 Exit Enter Do SMIT provides the following field List only the active volume groups The default is to list all volume groups sm
368. rict allocation policy Copies of a logical partition can be allocated to share or not to share the same physical volume The Strict variable is one of the following y Sets a strict allocation policy so that copies of a logical partition cannot share the same physical volume default n Does not set a strict allocation policy so that copies of a logical partition can share the same physical volume S Sets a super strict allocation policy so that the partitions allocated for one mirror cannot share a physical volume with other partitions from another mirror Note When changing a non super strict logical volume to a superstrict logical volume you must use the u flag S StripeSize Specifies the number of bytes per striped Must be a power of two between 4 K and 128 K for example 4 K 8 K 16 K 32 K 64 K or 128 K Note The d e m and s flags are not valid when creating a striped logical volume using the s flag s Size Number of logical partitions t Type The logical volume type jfs jfslog and so on 286 AIX Logical Volume Manager from A to Z Troubleshooting and Commands u UpperBound Sets the maximum number of physical volumes for new allocation The value must be between one and the total number of physical volumes The default is the total number of physical volumes in the volume group There is a known problem with allocp which causes this error 0516 423 allocp Not
369. roup can be exported not individual physical volumes Appendix A High level LVM commands 211 Using the exportvg command and the importvg command you can also switch ownership of data on physical volumes shared between two processors m Note 1 A volume group that has a paging space volume on it cannot be exported while the paging space is active Before exporting a volume group with an active paging space volume ensure that the paging space is not activated automatically at system initialization and then reboot the system 2 The mount point information of a logical volume would be missing from the LVCB logical volume control block if it is longer than 128 characters Please make a note of the mount points that are longer than 128 characters as you will need to edit the etc filesystems file manually upon executing importvg command to import this volume group completely A 6 The extendlv command The following summarizes the options for the extendlv command extendlv Increases the size of a logical volume by adding unallocated physical partitions from within the volume group Usage extendlv a Position e Range u Upperbound s Strict LPartitions PhysicalVolume s Usage extendlv mMapFile LPartitions 212 AIX Logical Volume Manager from A to Z Troubleshooting and Commands a Position Sets the intra physical volume allocation policy the position of the logical partitio
370. rruption on page 146 2 5 Corruption example 2 PVID corruption The following example shows techniques for recreating VGDAs and rebuilding logical volumes from maps Chapter 2 Problem determination and recovery 75 For an example of what not to do with the exportvg command we will make a volume group workvg consisting of two disks hdisk5 and hdisk6 mkvg ft2 y workvg hdisk5 hdisk6 workvg 1svg workvg VOLUME GROUP workvg VG IDENTIFIER 00017d372f40c3 5 VG STATE active PP SIZE 4 megabyte s VG PERMISSION read write TOTAL PPs 2150 8600 megabytes MAX LVs 256 FREE PPs 2145 8580 megabytes IVs 2 USED PPs 5 20 megabytes OPEN LVs 2 QUORUM 2 TOTAL PVs 2 VG DESCRIPTORS 3 STALE PVs 0 STALE PPs 0 ACTIVE PVs 2 AUTO ON yes MAX PPs per PV 2032 MAX PVs 16 crfs v jfs a size 32768 m workfs g workvg Based on the parameters chosen the new workfs JFS file system is limited to a maximum size of 134217728 512 byte blocks New File System size is 32768 mount workfs cd workfs Hecho data gt fs ok 1s fs _ok lost found Bi J We have also created a reference file system workfs containing some sample data to recover Let us assume we experience PVID corruption We can simulate this by running dd if dev zero of dev hdisk6 count 1 Note that in this example only the PVID is corrupted If there is any suspicion that VGDAs are corrupt there are some very import
371. run the rmdev at this point we will have to take our data off line Note Whenever repair of VGDA corruption is performed with low level modification commands and we have a mix of good and bad VGDAs all devices containing bad copies of the VGDA should be removed before repair is attempted This should be done so that the corruption does not propagate across the VGDAs We then run the commands shown below Note that before we do the vary off as far as the logical volume device driver is concerned lowvg contains both hdisk6 and hdisk7 After we vary on again only hdisk7 is in the volume group from the drivers point of view since hdisk6 has been removed If you want to follow the driver s view of which physical volumes are in a volume group you can use a script such as pvsinvg provided in Appendix E 7 pvsinvg on page 398 we can t use 1svg p Or lqueryvg g until we have fixed the core dump problem Note that this script only displays physical volumes that have an open logical volume so it will not work when volume groups are varied on in maintenance mode 152 AIX Logical Volume Manager from A to Z Troubleshooting and Commands umount lowfs1 umount lowfs2 varyoffvg lowvg rmdev dl hdisk6 hdisk6 deleted varyonvg f lowvg PV Status 00017d377c3abace PVREMOVED hdisk7 00017d377b71d55c PVACTIVE varyonvg Volume group lowvg is varied on G J Note that the forced vary on s
372. s 1 Mirror write consistency is on 2 Mirror write consistency is off B 10 The Ichlvcopy command The following summarizes the options for the 1chlvcopy command IchIlvcopy An object file that changes the logical volume in question for online backup Usage Ichlvcopy I LVID c Copy f P n new_minor_num m child_minor_num I LVID The logical volume ID c Copy Mirror copy to mark as split mirror copy The allowed values of copy are 1 2 or 3 If this option is not specified the default for copy is the last mirror copy of the logical volume 294 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Forces split mirror copy even if there are stale partitions If used with the B option the child backup logical volume if one was created with the I option will be removed with the force option Maintains information about the existence of an online split mirror copy across a reboot and also allows other nodes in a concurrent mode environment to be aware of the existence of the online split mirror copy n new_minor_num m child_minor_num The minor number of the new logical volume B 11 The Ichangepv command The following summarizes the options for the 1changepv command Ichangepv An object files that changes the attributes of a physical volume Usage Ichangepv g VGID p PVID r PVState a AllocState
373. s command sequence checks all file systems in which the check attribute is set to true check true If the fsck command executed from the etc rc file is unable to guarantee the consistency of any file system system initialization continues However the mount of any inconsistent file systems may fail A mount failure may cause incomplete system initialization Note that by default the usr var and tmp file systems have the check attribute set to false check false in their etc filesystem stanzas The attribute is set to false for the following reasons 1 The boot process explicitly runs the fsck command on the usr var and tmp file systems 2 The usr var and tmp file systems are mounted when the etc rc file is executed The fsck command will not modify a mounted file system Furthermore the fsck command run on a mounted file system produces unreliable results Appendix D Other related commands 353 D 14 The fsdb command The following summarizes the options for the fsdb command fsdb Debugs file systems Usage fsdb FileSystem Disables the error checking routines used to verify i nodes and block addresses The o subcommand switches these routines on and off When these routines are running the fsdb command reads critical file system data from the superblock The obtained information allows the fsdb command to access the various file system objects successfully and to perform various
374. s attribute is required when creating a JFS file system The maximum size of a JFS file system is a function of its fragment size and the nbpi value These values yield the size restrictions shown in Table 1 on page 330 328 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Specifies the attributes for auto mount yes File system is automatically mounted at system start no File system is not automatically mounted at system start d Attribute Deletes the specified attribute from the etc filesystems file for the specified file system m NewMountPoint Specifies the new mount point n NodeName Specifies a node name for the specified file system The node name attribute in the etc filesystems file is updated with the new name The node name attribute is specific to certain remote virtual file system types such as the NFS Network File System virtual file system type Sets the permissions for the file system ro Specifies read only permissions rw Specifies read write permissions Sets the accounting attribute for the specified file system yes File system accounting is to be processed by the accounting subsystem no File system accounting is not to be processed by the accounting subsystem this is the default u MountGroup Specifies the mount group Mount groups are used to group related mounts so that they can be mounted as one instead of mounting each
375. s greater than the number of copies in LogicalVolume before the split The NewLogicalVolume will have all the same logical volume characteristics as LogicalVolume If LogicalVolume does not have a logical volume control block the command will succeed with a warning message and create NewLogicalVolume without a logical volume control block There are additional considerations to take when splitting a logical volume containing a file system After the split there will be two logical volumes but there will only be one entry in the etc filesystems file that refers to LogicalVolume To access NewLogicalVolume as a file system you must create an additional entry in etc filesystems with a different mount point that refers to NewLogicalVolume If the mount point does not already exist you Appendix A High level LVM commands 275 have to create it before the new file system can be mounted In addition if NewLogicalVolume was created while LogicalVolume was open you have to run the command fsck dev NewLogicalVolume before the new file system can be mounted You can not use the System Management Interface Tool SMIT to run this command Message catalogs are not supported for this command therefore the error messages are provided in English only with no message catalog numbers A 33 The synclvodm command The following summarizes the options for the synclvodm command synclvodm Synchronizes or rebuilds the logical volume control
376. s made to the logical volume are not reflected in the file system They must be done separately A 2 The chpv command The following summarizes the options for the chpv command chpv Changes the characteristics of a physical volume in a volume group Usage chpv a Allocation v Availability c PhysicalVolume s a Allocation Sets the allocation permission for additional physical partitions on the physical volume specified by the PhysicalVolume parameter Either allows yes the allocation of additional physical partitions on the physical volume or prohibits no the allocation of additional physical partitions on the physical volume The Allocation variable can be either y Allows the allocation of additional physical partitions on the physical volume n Prohibits the allocation of additional physical partitions on the physical volume The logical volumes that reside on the physical volume can still be accessed C Clears the boot record of the given physical volume Appendix A High level LVM commands 203 v Availability Sets the availability of the physical volume If you set the availability to closed logical input and output to the physical volume are stopped You should close a physical volume when the physical volume is removed from operation Access to physical volume data by the file system or the virtual memory manager is stopped but you can continue to use the system management com
377. s multiplied by the block size The default write size for the backup command writing to tape devices is 51200 100 512 for backups by name and 32768 32 1024 for backups by i node The write size must be an even multiple of the tape s physical block size The value of the b flag is always ignored when the backup command writes to diskette In this case the command always writes in clusters that occupy a complete track Specifies that the tape is a cartridge not a nine track Copyright IBM Corp 2000 321 e RegularExpression Specifies that the files with names matching the regular expression are not to be packed A regular expression is a set of characters meta characters and operators that define a string or group of strings in a search pattern It can also be a string containing wildcard characters and operations that define a set of one or more possible strings The e flag is applied only when the p flag is specified f Device Specifies the output device To send output to a named device specify the Device variable as a path name such as dev rmt0 To send output to the standard output device specify a minus sign The minus feature enables you to pipe the output of the backup command to the dd command You can also specify a range of archive devices The range specification must be in the following format dev deviceXXX YYY where XXX and YYY are whole numbers and XXX must alwa
378. s the write verify state for the logical volume Sets whether all writes to the logical volume should be verified with a follow up read or not The Verify variable is one of the following y All writes to the logical volume are to be verified with a follow up read n All writes to the logical volume are not to be verified with a follow up read default w MWCFlag Mirror write consistency ensures data consistency among mirrored copies of a logical volume during normal I O processing The MWC flag sets mirror write consistency to on or off y Turns on mirror write consistency default n No mirror write consistency See the f flag of the syncvg command x Maximum Sets the maximum number of logical partitions that can be allocated to the logical volume The maximum number of logical partitions per logical volume is 32 512 Commands called getlvodm lchangelv lqueryly mv and putlvodm The flags to change the position of the logical partitions the interphysical volume allocation policy the strictness or the upperbound will only take effect when new partitions are added or removed The other flags take effect immediately If any of the flags are used that change the permissions of the special file and a big VGDA is used for the volume group then these permissions will be maintained during an importvgif the R flag is used 202 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Change
379. s to allocate for extending the logical volume Note the ExtendSize the CurrentSize of the logical volume should not exceed LVM_MAXLPS 65535 In the current implementation of the LVM a logical volume cannot go beyond 2 gigabytes 298 AIX Logical Volume Manager from A to Z Troubleshooting and Commands filename The name of the file that contains the map This map comprises of a set of triplets one for each partition to be allocated Each triplet contains the physical volume ID the physical partition number and the logical partition number respectively thus specifying the exact location of each physical partition B 17 The linstallpv command The following summarizes the options for the linstallpy command a VGDA linstallpv An object file command that adds a physical volume to a volume group and adds Usage installpv N PVName g VGID f N PVName The name of the physical volume to add g VGID The volume group ID of the volume group to which the physical volume will be added Forces the physical volume to be added to the specified volume group unless it is a member of another volume group in the Device Configuration Database or of a volume group that is active B 18 The Imigratelv command The following summarizes the options for the Imigratelv command Imigratelv An object file command that moves the partitions for a particular logical volume following
380. s to inspect and then repair In the previous example situations we have already seen a variety of inspection commands odmget the high level commands the low level query commands and so on In fact the toolbox of inspection commands for LVM recovery and problem determination is very large including in an approximate attempt to order the commands by depth of enquiry into the internal structures of AIX e Checking the errorlog errpt e Checking free file system space df e The high level commands 1spv 1slv svg e Checking fileset levels 1slpp e Checking device availability 1sdev lsattr Checking the ODM odmget korn shell debug sh xv trclvm The low level commands getlvcb lqueryvg lquerypv lquerylv Examining raw physical volumes VGDAs and logical volumes dd e Examining in kernel memory structures crash kdb Chapter 2 Problem determination and recovery 87 In addition we have already seen that we can identify the point of failure in the high level commands by shelling out high level commands and tools such aS trclvm There are often several ways to obtain a particular piece of information within AIX This list is not meant to be exhaustive The value of many of the tools should be clear however a discussion of the utility of these tools is of value 2 6 1 Checking the errorlog 88 The error report should indicate both hardware and software LVM errors and identifying if there is a hardware or softw
381. se this is the default Specifying bf true requires a fragment size of 4096 and compress no This attribute only applies to AIX Version 4 2 or later a compress no LZ Specifies data compression If you do not want data to be compressed set this option to no The default compress value is no Selecting compression requires a fragment size of 2048 or less a frag 512 1024 2048 4096 Specifies the JFS fragment size in bytes A file system fragment is the smallest unit of disk storage that can be allocated to a file The default fragment size is 4096 bytes a logname LVName Specifies the log logical volume name The specified logical volume will be the logging device for the new JFS The LVName logical volume must already exist The default action is to use an existing logging device in the target volume group a nbpi 512 1024 2048 4096 8192 16384 32768 65536 131072 Specifies the number of bytes per i node nbpi The nbpi affects the total number of i nodes on the file system The nbpi value is inversely proportional to the number of i nodes on the file system The default nbpi value is 4096 bytes The values 32768 65536 and 131072 only apply to AIX Version 4 2 or later a size Value Specifies the size of the JFS in 512 byte blocks If the specified size is not evenly divisible by the physical partition size it is rounded up to the closest number that is evenly divisible Thi
382. sicalVolume parameter otherwise all the physical volumes in the volume group are available for the Appendix A High level LVM commands 235 transfer All physical volumes must be within the same volume group The specified source physical volume cannot be included in the list of DestnPV parameters This command is not allowed if the volume group is varied on in concurrent mode The allocation of the new physical partitions follows the policies defined for the logical volumes that contain the physical partitions being moved The migratepv command only when the source and target physical volumes are specified fails when a boot logical volume is found on the source physical volume When you migrate a physical volume the boot logical volume must remain intact Two contiguous physical partitions and the new boot image must be built on the new boot logical volume If you specify a logical volume that contains the boot image the migratepv 1 command attempts to find enough contiguous partitions on one of the target physical volumes If the migration is successful the migratepv command prints a message that recommends the user run the bosboot command to indicate a change in the boot device The attempted migration fails if the migratepv 1 command is unable to find enough contiguous space to satisfy the request Note All Logical Volume Manager migrate functions work by creating a mirror of the logical volumes involved and then re
383. size of less than 2 gigabytes are recognized on an AIX Version 3 2 system Furthermore file systems created with an ag value greater than 8 are not recognized on an AIX Version 4 1 system much less an AIX Version 3 2 system The ag bf compress frag and nbpi attributes are set at file system creation and cannot be changed after the file system is successfully created The size attribute defines the minimum file system size and you cannot decrease it once the file system is created 5 The root file system cannot be compressed 6 Some nbpi values and allocation group sizes are mutually exclusive If the bs flag is specified by itself and no conversions other than sync noerror or notrunc are specified then the data from each input block will be written as a separate output block if the read returns less than a full block and sync is not specified then the resulting output block will be the same size as the input block If the bs flag is not specified or a conversion other than sync noerror or notrunc is specified then the input will be processed and collected into full sized output blocks until the end of input is reached D 6 The defragfs command The following summarizes the options for the defragfs command defragfs Increases a file system s contiguous free space Usage defragfs q r Device FileSystem q Reports the current state of the file system 338 AIX Logical Volume Manage
384. sk5 0018 0036 hdisk1 0036 hdisk4 0019 0065 hdisk3 0037 hdisk5 0020 0037 hdisk1 0037 hdisk4 Ne J hdisk5 has completely replaced hdisk2 The failed disk can then be removed from the system rmdev d 1 hdisk2 It should be noted that this method doesn t keep the same hdisk number Note Any logical volumes on the failed physical volume that were not mirrored will not be re created on the new physical volume Therefore they must be re created manually and restored from backup 198 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Appendix A High level LVM commands The high level commands are the ones aimed at administrators and users They check the validity of the arguments that are passed and prevent non valid values from destroying the system Wise users should stay with these commands as much as possible This appendix provides the options for these commands A 1 The chlv command The following summarizes the options for the chlv command chlv This script changes the name of a logical volume in a volume group Usage chlv n NewLVName LogicalVolume n Changes the name of the logical volume to that specified by the NewLVName variable Logical volume names must be unique system wide and can range from 1 to 15 characters If the logical volume that you change the name for is a log then each file system that uses that log must be pointed to the new name This option is not all
385. slog open syncd N A lsvg p lowg lowvg PV_NAME PV STATE TOTAL PPs FREE PPs FREE DISTRIBUTION hdisk6 active 1075 1070 275 52103 215 3 62156 215 hdisk7 active 1075 1072 215 2212 4215 5215 215 Li P eat HNW HA w HNE We will corrupt the VGDA with dd Since lowvg is on hdisk6 we run echo c dd bs 1 seek echo ib 8 n200000 11000 bc gt of dev hdisk6 LVM commands will now begin to dump core lsvg M lowg Segmentation fault coredump varyonvg lowvg usr sbin syncvg 16 16184 Segmentation fault coredump usr sbin syncvg 18 16186 Segmentation fault coredump 0516 932 usr sbin syncvg Unable to synchronize volume group lowvg Note that at this stage our data is still available a N 1s l lowfs1 total 16 Yw r r 1 root sys 8 Sep 27 12 56 datal drwxrwx 2 root system 512 Sep 27 12 56 lost found 1s 1 lowfs2 total 16 rYw r r 1 root sys 8 Sep 27 12 56 data2 drwxrwx 2 root system 512 Sep 27 12 56 lost found p Let us now approach this as a support call We begin with problem determination Chapter 2 Problem determination and recovery 147 m Note if the following steps get us nowhere we may need debug versions of the LVM libraries or commands to produce a core that contains the current state of LVM to be analyzed by AlX support This would be the case where we suspect a defect Since LVM commands are dumping c
386. smit 1slv SMIT provides the following fields Logical volume name The name of the logical volume to report F4 for list List options Status The configuration of the logical volume 38 Physical volume map Usage of physical partitions by physical volume and region AIX Logical Volume Manager from A to Z Troubleshooting and Commands smit 1lslv Show the characteristics of a logical volume Logical partition map The physical partition to logical partition mapping i N Show Characteristics of a Logical Volume Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields LOGICAL VOLUME name odin _1v List OPTION status F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image F9 Shel1 F10 Exit Enter Do x S smit 1slv can also be used to list the configuration of a logical volume N COMMAND STATUS Command OK stdout yes stderr no Before command completion additional instructions may appear below LOGICAL VOLUME odin 1v VOLUME GROUP asgard vg LV IDENTIFIER 00017d378ed 7c9d 1 PERMISSION read write VG STATE active complete LV STATE closed syned TYPE jfs WRITE VERIFY off MAX LPs 512 PP SIZE 4 megabyte s COPIES 1 SCHED POLICY parallel LPs 12 PPs 12 STALE PPs 0 BB POLICY relocatable INTER POLICY minimum RELOCATABLE yes INTRA POLICY middle UPPER BOUND 32 MOUNT POINT home odin LABEL home odin MIRR
387. some of the fields will contain a question mark in place of the missing data The lsvg command attempts to obtain as much information as possible Appendix A High level LVM commands 231 from the description area when the command is given a logical volume identifier Note that to determine a volume group s major number use the 1s al dev VGName Command This command lists the special device file that represents the volume group The volume group major number is the same as the major device number of the special device file For example for a volume group named hatvg enter the following command ls al dev halvg This command returns the following crw rw 1 root system 52 0 Aug 27 19 57 dev halvg In this example the volume group major number is 52 When no flags are used the following characteristics of the specified volume group are displayed Information displayed if you do not specify any flags Volume group Name of the volume group Volume group names must be unique system wide and can range from one to 15 characters Volume group state State of the volume group If the volume group is activated with the varyonvg comman4d the state is either active complete indicating all physical volumes are active or active partial indicating some physical volumes are not active If the volume group is not activated with the varyonvg command the state is inactive Permission Access permission read only or read writ
388. spare disk and create a volume group with similar characteristics to the lost volume group AIX Logical Volume Manager from A to Z Troubleshooting and Commands mkvg ft2 y pplvg hdisk6 pplvg mklv m map y ppllv pplvg 1 pollv od tc dev hdisk6 sed s g grep AIXLVCB 10400000ATXLVCB 0 0j s 0 0 0 cH be ib 8 10400000 2228224 10400000 1000 4352 Converting octal 10400000 bytes to decimal 512 byte octal 1000 sectors gives us 4352 For a bigVG we can see PP1 is located in sector 17408 mkvg B ft2 y pplbigvg hdisks pplbigvg cat bigmap hdisk8 1 mklv m bigmap y pplbiglv pplbigvg 1 pplbiglv od tc dev hdisk8 sed s g grep AIXLVCB 42000000ATXLVCB 0 j fs 0_ cH be ib 8 42000000 1000 17408 quit eu J Now that we know the offset PP1 for a given size of VGDA we can grep along the disk in PPSIZE chunks and find where our logical volumes started We can also obtain the logical size from the LVCB a four byte integer at offset 208 within the LVCB If we make the assumption that we are using the default minimum inter policy and our LV is contiguous less likely but this is another reason apart from performance that you should run reorgvg we can assume a map and lay down a contiguous logical volume Obviously this will not work every time Again we are in a last resort situation We are also in trouble if our logical volume
389. stdout yes stderr no Before comand carpletio additional instructions may appear below PHYSICAL VOLLME hdisk9 VOLLME GROUP park vg PV IDENTIFIER 00017d37628265fc VG IDENTIFIER 00017d378ed 7c90 PV SIATE active STALE PARTITIONS 6 ALLOCATABLE yes PP SIZE 4 megabyte s LOGICAL VOLUMES 5 TIAL PPs 537 2148 megabytes VG DESCRIPICRS 1 FREE PPs 459 1836 megabytes USED PPs 78 312 megabytes FREE DISTRIBUTION 108 29 107 107 108 USED DISTRIBUTION 00 78 00 00 00 F1l Help F2 Refresh F3 Cancel F Comend F8 Image F9 Shel1 F1l0 xit Find n Find Next W J Explanation of the fields displayed Physical volume Volume group PV Identifier VG Identifier PVstate Stale Partitions PP size Total PPs 46 Name of the physical volume Name of volume group The physical volume identifier for this physical disk The volume group identifier of the volume group that this physical disk is a member State of the physical volume If the volume group that contains the physical volume is varied on with the varyonvg commana the state is active missing or removed If the physical volume is varied off with the varyoffvg command the state is varied off Number of physical partitions on the physical volume that are not current Size of physical partitions on the volume Total number of physical partitions on the physical volume AIX Logical Volume Manager from A to Z Troubleshooting and
390. stem boot If the copy fails at boot time the a flag ignores the system dump e Estimates the size of the dump in bytes for the current running system If the dump will be compressed then the size shown is the estimate of the size after compression Appendix D Other related commands 381 The reset button or the dump key sequences will force a dump with the key in the normal position or on a machine without a key mode switch Note On a machine without a key mode switch a dump can not be forced with the reset button or the key switch without this value set Requires the key mode switch to be in the service position before a dump can be forced with the reset button or the dump key sequences This is the default setting Displays statistical information about the most recent system dump This includes date and time of last dump number of bytes written and completion status If the dump was compressed then this flag shows both the original uncompressed size and the compressed size of the dump The compressed size is the size of what was actually written to the dump device Lists the current value of the primary and secondary dump devices copy directory and forcecopy attribute Makes permanent the dump device specified by the p or s flags The p flag can only be used with the por s flags p Device Temporarily changes the primary dump device to the specified device The device can be a l
391. subject to appropriate terms and conditions including in some cases payment of a fee The information contained in this document has not been submitted to any formal IBM test and is distributed AS IS The information about non IBM vendor products in this manual has been supplied by the vendor and IBM assumes no responsibility for its accuracy or completeness The use of this information or the implementation of any of these techniques is a customer responsibility and depends on the customer s ability to evaluate and integrate them into the customer s operational environment While each item may have been reviewed by IBM for accuracy in a specific situation there is no Copyright IBM Corp 2000 401 guarantee that the same or similar results will be obtained elsewhere Customers attempting to adapt these techniques to their own environments do so at their own risk Any pointers in this publication to external Web sites are provided for convenience only and do not in any manner serve as an endorsement of these Web sites Any performance data contained in this document was determined in a controlled environment and therefore the results that may be obtained in other operating environments may vary significantly Users of this document should verify the applicable data for their specific environment This document contains examples of data and reports used in daily business operations To illustrate them as completely as possible the e
392. synchronizing the logical volumes The original logical volume is then removed Therefore the migratepv command alone should not be used to move a logical volume containing the primary dump device The command will execute but any subsequent system dump will fail In addition the physical volume cannot be removed from the volume group You must first reassign the dump device using the sysdumpdev command A 15 The mirrorvg command The following summarizes the options for the mirrorvg command mirrorvg mirrors all the logical volumes that exist on a given volume group Usage mirrorvg S s Q c Copies m VolumeGroup PhysicalVolume s 236 AIX Logical Volume Manager from A to Z Troubleshooting and Commands c Copies Specifies the minimum number of copies that each logical volume must have after the mirrorvg command has finished executing It may be possible through the independent use of mklvcopy that some logical volumes may have more than the minimum number specified after the mirrorvg command has executed Minimum value is 2 and 3 is the maximum value A value of 1 is ignored m Allows mirroring of logical volumes in the exact physical partition order that the original copy is ordered This option requires you to specify a PhysicalVolume s where the exact map copy should be placed If the space is insufficient for an exact mapping then the command will fail You should add new drives or p
393. sysb command issues a warning indicating that the tape will not be bootable You can install a mksysb image from a system that does not support booting from tape by booting from a CD and entering maintenance mode In maintenance mode you will be able to install the system backup from tape A 20 The mkszfile command The following summarizes the options for the mkszfile command mkszfile Creates a file containing information about the rootvg volume group for use by the mksysb command Usage mkszfile X m Appendix A High level LVM commands 257 Creates map files that specify the mapping of the logical to physical partitions for each logical volume in the volume group This mapping can be used to allocate the same logical to physical mapping when the image is restored The map file locations are stored in the MAPFILE field in the image data file for each logical volume Sufficient space would exist in the tmp file system for map creation because the installation routines place the maps in the tmp file system before issuing the mklv command For example for the hd7 logical volume the location of the map file is tmp vgdata rootvg hd7 map The MAPFILE field in the image data file for the hd7 logical volume is under the entry MAPFILE tmp vgdta rootvg hd7 map The map files in the backup image are copied after the bosinst data and image data files X Expands tmp if needed Commands
394. system For the command line options see cplvin Appendix A 4 The cplv command on page 209 or use the SMIT fastpath smit cplv SMIT provides the following fields Source logical volume name The name of the logical volume to be copied F4 for list Destination logical volume The name of an existing logical volume to which the source logical volume will be copied smit cplv Copy over an existing logical volume S Copy over an existing logical volume Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields SOURCE logical volume name DESTINATION logical volume F1l Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image F9 Shel1 F10 Exit Enter Do B 1 2 4 Copy a logical volume to a smaller logical volume 28 This cannot be performed by simply copying the logical volume The smit cplv command will allow this but will result in a corrupted file system To do this properly the user must create the new logical volume at the required size create a file system on the logical volume and mount it The information can then be copied using cp r or cpio p for details on the cpio command see Appendix D 4 The cpio command on page 331 AIX Logical Volume Manager from A to Z Troubleshooting and Commands 1 2 5 Adda mirrored copy to a logical volume The mklvcopy command increases the number of copies of each logical parti
395. system dump device This should not be done with a usr client diskless client or dataless client First we check which existing paging spaces are in rootvg This can be done by running lsvg p rootvg If hd6 is the only paging space in rootvg then it will be necessary to create an alternate paging space with mkps s lt number gt a rootvg where number is an appropriate value for paging space size for this system 1sps p rootvg is then run again to get the name of the new paging space hd6 is now deactivated for the next reboot with chps n hd6 It is now very important to change the default paging space used by the sbin rc boot script To do this we carefully alter the line swapon dev hde to swapon dev pagingnn where the new paging space is either the new paging space created with mkps or another pre existing paging space This will allow the last phase of the boot process to complete To make sure the earlier phases of boot run successfully as well we need to re create the boot logical volume with the new copy of rc boot We check which hdisk the boot logical volume is on with 1slv m hd5 and run bosboot ad dev hdiskn A reboot is necessary to deactivate hd6 After doing this if either the primary or secondary dump device point to hd6 this can be checked with sysdumpdev 1 then that pointer should be reset by running sysdumpdev P p dev sysdumpnull It is now possible to delete and re create hd6 it is advisable to re
396. t not a period or slash question mark Matches any single character but nota period or slash brackets Matches any one of the characters enclosed within the brackets If a pair of characters separated by a dash are contained within the brackets the pattern matches any character that lexically falls between the two characters in the current local Additionally a period or a slash within the brackets will not match a a period or a slash in a file name backslash Matches the immediately following character preventing its possible interpretation as a metacharacter Appendix D Other related commands 373 Continues restoring when tape errors are encountered y Normally the restore command asks you whether or not to continue In either case all data in the read buffer is replaced with zeroes The y flag applies only when the archive is in file system format Displays a usage message The restore command reads archives created by the backup command and extracts the files stored on them These archives can be in either file name or file system format An archive can be stored on disk diskette or tape Files must be restored using the same method by which they were archived This requires that you know the format of the archive The archive format can be determined by examining the archive volume header information that is displayed when using the
397. t access smit varyonvg a gt Activate a Volume Group Type or select values in entry fields Press Enter AFTER making all desired changes Entry Fields VOLUME GROUP name RESYNCHRONIZE stale physical partitions yes Activate volume group in SYSTEM no MANAGEMENT mode FORCE activation of the volume group no Warning this may cause loss of data integrity Varyon VG in Concurrent Mode no F1 Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image F9 Shel1 F10 Exit Enter Do 1 1 3 Deactivate a volume group The varyoffvg command deactivates the volume group specified by the VolumeGroup parameter along with its associated logical volumes The logical volumes must first be closed For example if the logical volume contains a file system it must be unmounted 6 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Note A volume group that has a paging space volume on it cannot be varied off while the paging space is active Before de activating a volume group with an active paging space volume ensure that the paging space is not activated automatically at system initialization and then reboot the system For the command line options see varyoffvg in Appendix A 38 The varyoffvg command on page 280 or use the SMIT fastpath smit varyoffvg SMIT provides the following fields Volume group name The name of the volume group to be de activated F4 for l
398. t allocation policy for a logical volume The allocation policy can be changed for a logical volume but it will not affect the allocation of existing physical partitions It will only affect the allocation of 30 AIX Logical Volume Manager from A to Z Troubleshooting and Commands new partitions or if the logical volume is reorganized using the reorgvg command For the command line options see chlv in Appendix A 1 The chlv command on page 199 or use the SMIT fastpath smit chlv SMIT provides the following fields Logical volume name The name of the logical volume that will have its allocation policy changed F4 for list Position on physical volume Sets the intra physical volume allocation policy Range of physical volumes Sets the inter physical volume allocation policy smit chlv aa N Change a Logical Volume Type or select values in entry fields Press Enter AFTER making all desired changes TOP Entry Fields Logical volume NAME odin 1v Logical volume TYPE jfs POSITION on physical volume middle RANGE of physical volumes minimum MAXIMUM NUMBER of PHYSICAL VOLUMES 32 to use for allocation Allocate each logical partition copy yes on a SEPARATE physical volume RELOCATE the logical volume during yes reorganization Logical volume LABEL home odin MAXIMUM NUMBER of LOGICAL PARTITIONS 512 SCHEDULING POLICY for reading writing parallel logical partition copies PERMISSI
399. t become confused about problems that are not there An investigation into why a logical volume has gone stale will not get far if strenuous efforts are not made to determine why an Isdev on the logical volume device shows it as defined in fact this is normal behavior 2 7 Rebuilding and repair We have already seen some examples of rebuilding problem situations The following section expands on this by introducing a range of typical problems The reader must be aware that sometimes situations simply are not recoverable The data just may not be around any more No amount of recovery procedures are a substitute for a single good backup One should not overestimate the power of recovery procedures There is also the case where a disk is failing and recovery procedures may be performed to get data back on line easily so that it can be copied to a good media There is the further case where a logical volume may be brought back online but LVM control data integrity can no longer be guaranteed This may be a case of salvaging as much as possible and then re creating the logical volume or volume group Hopefully the previous sections will have given us some idea of what the failing component is in the recovery situation Depending on what is damaged the reader should refer to the following sections which cover e ODM corruption 128 AIX Logical Volume Manager from A to Z Troubleshooting and Commands LVM control data corruption VGDA
400. t if HACMP clvmis not running y VolumeGroup Specifies the volume group name which must be unique system wide and can be from one to 15 characters Can only contain A Z a z 0 9 ee wou or Meal Commands called getlvodm getvgname lcreatevg linstallpv putlvodn lvgenminor lvgenmajor and varyonvg 260 AIX Logical Volume Manager from A to Z Troubleshooting and Commands A 22 The mkvgdata command The following summarizes the options for the mkvgdata command mkvgdata Creates a file containing information about a volume group for use by the savevg and restvg commands This file is a symbolic link to the mkszfile script Usage mkvgdata X m VG Name M Creates map files that specify the mapping of the logical to physical partitions for each logical volume in the volume group This mapping can be used to allocate the same logical to physical mapping when the image is restored The map file locations are stored in the MAPFILE field in the image data file for each logical volume Sufficient space should exist in the tmp file system for map creation because the installation routines place the maps in the tmp file system before issuing the mklv command For example for the hd7 logical volume the location of the map file is tmp vgdata rootvg hd7 map The MAPFILE field in the image data file for the hd7 logical volume is under the entry MAPFILE tmp vgdta rootvg hd7 map The map fi
401. t of migrations in source destination format MigrateMoves The list of legal moves to accomplish the migration B 38 The putlvcb command The following summarizes the options for the putlvcb command putlvcb An object file command that writes the logical volume control block Usage putlvcb a IntraPolicy c Copies e InterPolicy f FSLabel L Label i Identifier n NumLPs r Reloc s Strict t Type u UpperBound N v V Name x VGAutoOn LVName a IntraPolicy Writes the intra physical volume allocation policy to the logical volume LVName s control block c Copies Writes the copy allocation values to the logical volume LVName s control block e InterPolicy Writes the inter physical volume allocation policy to the logical volume LVName s control block f FSLabel Writes the file system label field to the logical volume LVName s control block L Label Writes the logical volume label field to the logical volume LVName s control block i Identifier Writes the logical volume identifier to the logical volume LVName s control block n NumLPs Writes the number of logical partitions for lvname to the logical volume LVName s control block 310 AIX Logical Volume Manager from A to Z Troubleshooting and Commands r Reloc Writes the relocation policy to the logical volume LVName s control block s Strict Writes the strictness allocation polic
402. t2 y workvg hdiske hdisk9 workvg chvg Qn workvg lspv grep workvg hdisk6 00091974de731316 workvg hdisk9 00017d37297a96c8 workvg lsvg workvg VOLUME GROUP workvg VG IDENTIFIER 00017d3729050708 VG STATE active PP SIZE 4 megabyte s VG PERMISSION read write TOTAL PPs 1612 6448 megabytes MAX LVs 256 FREE PPs 1612 6448 megabytes LVs 0 USED PPs 0 0 megabytes OPEN LVs 0 QUORUM 1 TOTAL PVs 2 VG DESCRIPTORS 3 STALE PVs 0 STALE PPs 0 ACTIVE PVs 2 AUTO ON yes MAX PPs per PV 2032 MAX PVs 16 Ne S We now create a file system on a new logical volume and check the partition maps on each hdisk with 1spv Chapter 2 Problem determination and recovery 89 crfs v jfs a size 16384 g workvg m workfs Based on the parameters chosen the new workfs JFS file system is limited to a maximum size of 134217728 512 byte blocks New File System size is 16384 mount workfs cd workfs Hecho data gt fs ok ls fs_ok lost found lspv M hdisk6 hdisk6 1 215 hdisk 6 216 loglv00 1 hdisk6 217 1v02 1 hdisk 218 1v02 2 hdisk 6 219 1075 lspv M hdisk9 hdisk9 1 537 y Note that in this case there is nothing on hdisk9 Partitions 1 through 537 are empty Let s now suppose that hdisk9 dies This with leave us in a situation where all of our data is still intact However we have lost at least one of our VGDAs Actually in this situation we have lost only one VGDA
403. tall image and an empty table of contents followed by the system backup root volume group image The root volume group image 256 AIX Logical Volume Manager from A to Z Troubleshooting and Commands is in backup file format starting with the data files and then any optional map files One of the data files mksysb uses is the bosinst data file If a bosinst data file doesn t exist var adm ras bosinst data is copied to root In AIX Version 4 3 3 and later versions mksysb always updates the target_disk_data stanzas in bosinst data to match the disks currently in the root volume group of the system where the mksysb command is running If you are using a customized bosinst data file and do not want the target_disk_data stanzas updated you must create the file save_bosinst data_file The mksysb command does not update bosinst data if the save_bosinst data_file exists m Note 1 The image the mksysb command creates does not include data on raw devices or in user defined paging spaces 2 If you are using a system with a remote mounted usr file system you cannot reinstall your system from a backup image 3 The mksysb command may not restore all device configurations for special features such as dev netbios and some device drivers not shipped with the product 4 Some rspc systems do not support booting from tape When you make a bootable mksysb image on an rspc system that does not support booting from tape the mk
404. tar command can work across communications channels where blocking may not be maintained b Blocks Specifies the number of 512 bytes blocks per record Both the default and the maximum is 20 which is appropriate for tape records Due to the size of inter record gaps tapes written with large blocking factors can hold much more data than tapes with only one block per record The block size is determined automatically when tapes are read the xor t function flags When archives are updated with the u and r functions the existing record size is used The tar command writes archives using the specified value of the Blocks parameter only when creating new archives with the c flag For output to ordinary files with the flag you can save disk space by using a blocking factor that matches the size of disk blocks for example the b4 flag for 2048 byte disk blocks C Directory Causes the tar command to perform a chdir subroutine to the directory specified by the Directory variable Using the C flag allows multiple directories not related by a close common parent to be archived using short relative path names For example to archive files from the usr include and etc directories you might use the following command tar c C usr include Filel File2 C etc File3 File4 The cDirectory flag must appear after all other flags and can appear in the list of file names given Creates a new archive and writes the files spe
405. te this redbook 0 00 c eee eee ix Comments welcome se raker edee s enee ea na ene teens X Chapter 1 LVM commands 00 eee eee 1 1 1 Volume group related commands 0 000 ee eee eee 2 1 1 1 Add a volume group 2 cee ee 3 1 1 2 Activate a volume group 2 cee ee 5 1 1 3 Deactivate a volume group sssaaa aana aaraa 6 1 1 4 Remove a physical volume from a volume group 8 1 1 5 Remove a volume group 0 0 ee 9 1 1 6 Bring a volume group into the system 10 1 1 7 Add physical volume to an existing volume group 12 1 1 8 Change name of volume group 00 00 eee eee 13 1 1 9 Reorganize a volume group 0 00 e eee eee 14 1 1 10 Listing the volume groups on the system 16 1 1 11 Listing the configuration of a volume group 18 1 1 12 List contents of a volume group 0 eee eee 21 1 2 Logical volume related commands 00000 eee eee 22 1 2 1 Adda logical volume herren eaa erea ee 23 1 2 2 Copy a logical volume to a new logical volume 26 1 2 3 Copy a logical volume to same sized or bigger logical volume 27 1 2 4 Copy a logical volume to a smaller logical volume 28 1 2 5 Adda mirrored copy to a logical volume 29 1 2 6 Set allocation policy for a logical volume 30 1 2 7 Set strictness for a logical v
406. ted Appendix A High level LVM commands 279 A 38 The varyoffvg command The following summarizes the options for the varyoffvg command varyoffvg Deactivates a volume group Usage varyoffvg s VolumeGroup S Puts the volume group into System Management mode so that only logical volume commands can be used on the volume group In this mode no logical volume can be opened or accessed by users Commands called getlvodm lvaryoffvg and putlvodm The varyoffvg command deactivates the volume group specified by the VolumeGroup parameter along with its associated logical volumes The logical volumes first must be closed For example if the logical volume contains a file system it must be unmounted To activate the volume group use the varyonvg command Note 1 To use this command you must either have root user authority or be a member of the system group A volume group that has a paging space volume on it cannot be varied off while the paging space is active Before deactivating a volume group with an active paging space volume ensure that the paging space is not activated automatically at system initialization and then reboot the system A 39 The varyonvg command The following summarizes the options for the varyonvg command varyonvg Activates a volume group Usage varyonvg b c f n p s u VolumeGroup 280 AIX Logical Volume Mana
407. ted on dev hd3 81920 0 100 105 2 tmp Se J Chapter 2 Problem determination and recovery 93 Here is another example that shows the incorrect running of exportvg on a volume group that is still varied on when tmp is full exportvg fullvg sed There is not enough space in the file system Isvg rootvg lsvg o rootvg fullvg Here we see the exportvg command was partially successful It removed the volume group from the ODM but did not vary it off You may also experience other forms of corruption In another test of this we saw the following lsvg o 0516 304 Unable to find device id 00017d374c013e1f in the Device Configuration Database vgid 00017d374c013e1f rootvg This second symptom is actually more common The point is that behavior may be undefined in this kind of situation To recover from the first situation it is sufficient to vary off the volume group manually we still have to fix the full tmp condition as well of course This will rebuild the ODM so that we can then run exportvg To recover from the second situation we have to think about what we are seeing Since we cannot pass a VGID to varyoffvg we have a stale entry in the entry in the kernel tables for this VGID Since this is what we are seeing we can attack it with lvaryoffvg g 00017d374c013e1f We will not need to run exportvg in this second case since the ODM entries were not recreated by the low level lvaryoffvg
408. tempt restore y to do n AIX LVCB found at PP216 offset 1765632 Type jfs LVname scrapelv NUMLPS 5 Attempt restore y to do y scrapelv ATX LVCB found at PP221 offset 1806592 Type jfslog LVname scrapelog NUMLPS 1 Attempt restore y to do y scrapelog qA J Note that we see the remnant LVCBs of other logical volumes we are not interested in Since the script has no real error checking it may also ask if you wish to restore logical volumes that already exist although a mklv on these logical volumes will of course fail mount scrapefs 1s 1 scrapefs total 16 Yw r r 1 root sys 5 Oct 01 13 58 data drwxrwx 2 root system 512 Oct 01 13 56 lost found Laying down contiguous logical volumes is not the last word in data recovery when only the raw data on disks is available Depending on how well we understand our data it may be possible to start with the first logical partition and then consider remaining logical partitions as candidate logical partitions Each candidate logical partition may be joined onto the logical volume with a mapped extend1v and data integrity checks performed If more data blocks are checked successfully we know we have a good candidate and permanently Chapter 2 Problem determination and recovery 169 170 extend the logical volume Otherwise we back out our extendvg with lreducelv and try another candidate partition Once we have confirmed the candidate is c
409. the form from replacepv Update the timestamp in odm for VG disk _timestamp g VGID T if eq 0 then putlvodm T disk_timestamp VGID fi 2 6 6 3 Logical volumes simple This is the base set of classes required for a logical volume Mirroring or striping result in the generation of additional objects Simple logical volume ODM classes Simple logical volume ODM classes name simplelv attribute lvserial_iq value 00017d375251870e 1 type R generic D rep n nls_index 648 CuAt name simplelv attribute stripe width value 0 type R generic DU rep y nls_index 1100 name simplelv status 1 chgstatus 1 ddins location parent odmvg connwhere PdDvin logical_volume lvsubclass lvtype CuDep name odmvg dependency simplelv 102 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Simple logical volume ODM classes CuDvDr resource devno valuel 41 value2 1 value3 simplelv 2 6 6 4 Logical volumes mirrored Mirrored logical volume ODM objects Mirrored logical volume ODM objects CuAt name mirrorlv attribute Ivserial_id value 00017d375251870e 3 type R generic D rep n nls index 648 CuAt name mirrorlv attribute copies value 2 type R generic DU rep y nls in
410. then dspmsg s 1 cmdlvm cat 764 lvmmsg 764 gt n S PROGNAM exit fi done lt tmp vginfos s fi fi gt amp 2 GW n 00017d376 b7ee896 putlvodm k 00017d376b7ee896 2 gt dev null test_return 0 0 0 DA N We have successfully taken the volume group lock Next shown below we set the LOCKED variable This will be used by the cleanup routine to decide whether or not to unlock the volume group We wouldn t want to always run a putlvodm K to unlock in cleanup since we might not have reached the section of code where we took the lock To always release the lock could result in accidental deletion of a lock taken by another process running another high level command If another high level command was running this could result in multiple critical sections of LVM code running simultaneously If this was allowed it could cause corruption LOCKED y lqueryvgs 1 gt tmp vginfol7482 2 gt dev null 0 0 0 lt tmp vginfo17482 116 AIX Logical Volume Manager from A to Z Troubleshooting and Commands We make a list of every varied on volume group on the system into a temporary file exportvg then loops through this file reading a VGID and a major number each time and tests to see if our VGID is in this list If it is we must fail since you cannot export a volume group that is varied on C D read VGID_ON MAJOR NUM 4 00017d376b7ee896 read VGID ON MAJOR NUM 0001743
411. thout having to acknowledge the fact that you will not be able to import this logical volume in earlier versions of AIX A 18 The mklvcopy command The following summarizes the options for the mklvcopy command mklvcopy Provides copies of data with the logical volume Usage mklvcopy a Position e Range k m MapFile s Strict u UpperBound LogicalVolume Copies PhysicalVolume s 250 AIX Logical Volume Manager from A to Z Troubleshooting and Commands a Position Sets the intra physical volume allocation policy the position of the logical partitions on the physical volume The Position variable can be one of the following m Allocates logical partitions in the outer middle section of each physical volume This is the default position Allocates logical partitions in the center section of each physical volume Allocates logical partitions in the outer edge section of each physical volume Allocates logical partitions in the inner edge section of each physical volume Allocates logical partitions in the inner middle section of each physical volume e Range Sets the inter physical volume allocation policy the number of physical volumes to extend across using the volumes that provide the best allocation The Range value is limited by the UpperBound variable set with the u flag and can be one of the following x Allocates across the maximum number of physical volumes
412. till use reducevg but with the Physical Volume ID PVID instead of the disk name reducevg VolumeGroup PVID A 26 The reorgvg command The following summarizes the options for the reorgvg command reorgvg Reorganizes the physical partition allocation for a volume group Usage reorgvg i VolumeGroup LogicalVolume s i Specifies physical volume names read from standard input Only the partitions on these physical volumes are organized Commands called allocp getlvodm lmigratepp lquerylv lquerypv lqueryvg and putlvodm Note This command is not allowed if the volume group is varied on in concurrent mode The reorgvg command reorganizes the placement of allocated physical partitions within the VolumeGroup according to the allocation characteristics of each logical volume Use the LogicalVolume parameter to reorganize specific logical volumes In the reorganization the order of priority goes in the order the logical volumes are listed by Isvg I Volume_Group If a list of logical volumes is provided the highest priority goes to the first logical volume in the list down to the last logical volume with the lowest The volume group must be varied on and must have free partitions before you can use the reorgvg command Appendix A High level LVM commands 265 The relocatable flag of each logical volume must be set to y with the chlv r command for the reorganization to take eff
413. tion in the logical volume This is accomplished by increasing the total number of physical partitions for each logical partition to the number represented by Copies You can request that the physical partitions for the new copies be allocated on specific physical volumes within the volume group otherwise all the physical volumes within the volume group are available for allocation The logical volume modified with this command uses the copies parameter as its new copy characteristic The data in the new copies will not be synchronized until one of the following occurs e The k option is used e The volume group is activated by the varyonvg command e The volume group or logical volume is synchronized explicitly by the syncvg command Individual logical partitions are always updated as they are written to Note To create a copy of a striped logical volume all systems that access the volume group must be at least at AIX Version 4 3 3 For the command line options see mklvcopy in Appendix A 18 The mklvcopy command on page 250 or use the SMIT fastpath smit mklvcopy SMIT provides the following fields Logical volume name The name of the logical volume to add the copy to F4 for list New total number of logical partition copies The new total number of copies for each logical partition after this copy has been added Physical volume names The names of the physical volumes that will be used to create th
414. tive paging spaces and journaling logs that are in current use In general before these logical volumes can be edited they must be unmounted or deactivated Furthermore in the case of mounted file systems any valuable data they contain should be archived for example to tape prior to unmounting and editing m Note The 1lvedit command cannot be used to reduce the size of an existing logical volume However the reduce command of the logical volume editor can be used to reduce the size of a logical volume that has been extended but not yet committed A 14 The migratepv command The following summarizes the options for the migratepv command migratepv Moves allocated physical partitions from one physical volume to one or more other physical volumes Usage migratepv i LogicalVolume SrcPhysicalVolume DestnPV s Reads the DestinationPhysicalVolume parameter from standard input LogicalVolume Moves only the physical partitions allocated to the specified logical volume and located on the specified source physical volume Commands called lmigratelv getlvodm dspmsg lquerylv lquerypv lqueryvg sort grep and awk The migratepv command moves allocated physical partitions and the data they contain from the SourcePhysicalVolume to one or more other physical volumes To limit the transfer to specific physical volumes use the names of one or more physical volumes in the DestinationPhy
415. to Z Troubleshooting and Commands performed at once reads from the most accessible copy Sequential Writes to logical volume copies performed sequentially and wait for completion before proceeding to the next Reads are from the primary copy smit chlv C N Change a Logical Volume Type or select values in entry fields Press Enter AFTER making all desired changes TOP Entry Fields Logical volume NAME odin 1v Logical volume TYPE jfs POSITION on physical volume middle RANGE of physical volumes minimum MAXIMUM NUMBER of PHYSICAL VOLUMES 32 to use for allocation Allocate each logical partition copy yes on a SEPARATE physical volume RELOCATE the logical volume during yes reorganization Logical volume LABEL home odin MAXIMUM NUMBER of LOGICAL PARTITIONS 512 SCHEDULING POLICY for reading writing parallel logical partition copies PERMISSIONS read write Enable BAD BLOCK relocation yes Enable WRITE VERIFY no Mirror Write Consistency yes BOTTOM F1l Help F2 Refresh F3 Cancel F4 List Esc 5 Reset F6 Command F7 Edit F8 Image _ F9 Shel1 F10 Exit Enter Do 1 2 10 Increase the maximum size of a logical volume The maximum size stops users increasing the size of logical volumes beyond a pre determined size While users can still increase the size of logical volumes by first changing the maximum size it serves as a warning The maximum size may be set as Chapter 1 LVM commands 35
416. tom of full dev is the inability to create new LVM objects such as logical volumes These may fail with errors such as the following Chapter 2 Problem determination and recovery 95 mklv y nospacelv1l devvg 1 0516 576 getlvname Unable to update device configuration database 0516 362 getlvname Unknown Object Data Manager error 0 0516 822 mklv Unable to create logical volume This kind of problem can be simulated by artificially filling dev then repeatedly running mklv mk1v will not fail immediately failure will only occur when the directory file for dev itself needs to be expanded Remember this is contained within the file system and we are creating special files here Interestingly when testing this mk1v would occasionally fail and take an ODM lock Procedures for clearing this are discussed in Checking the ODM on page 98 A useful test for finding out what is stealing space in a file system when space is taken as a file system is expanded is to run find file system xdev exec ls ld gt file1 then increase the size of the file system and run find filesystem xdev exec ls ld gt file2 Any changes to file sizes can be seen with diff file1 file2 Note that this method will not show files that are still open by applications but have been deleted 2 6 3 The high level commands 96 The high level commands are good for getting an initial feel for a problem and for checking if limits are bei
417. tomer number Credit card number Credit card expiration date Card issued to Signature We accept American Express Diners Eurocard Master Card and Visa Payment by credit card not available in all countries Signature mandatory for credit card payment 408 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Glossary Your glossary term acronym or abbreviation Term definition AIX Advanced Interactive eXecutive BSD Berkeley Software Distribution C SPOC Cluster Single Point Of Control CD ROM Compact Disc Read Only Memory CLVM Concurrent Logical Volume Manager DASD Direct Access Storage Devices FC AL Channel arbitrated loop GPFS General Parallel File System HACMP High Availability Cluster Multi Processing HACMP CRM High Availability Cluster Multi Processing Concurrent Resource Manager HACMP ES High Availability Cluster Multi Processing Enhanced Scalability HiPPI High Performance Parallel Interface Adapter I O Input Output IBM International Business Machines IPL Initial Program Load ITSO International Technical Support Organization JFS Journaled File System LPSN Last Physical Sector Number LTG Logical Track Group LVCB Logical Volume Control Block LVDD Logical Volume Device Driver LVID Logical Volume Identifier LVM Logical Volume Manager MWC Mirror Write Consistency MWCC Mirror Write Consistency Checking NBPI Number of Bytes Per In
418. tor Changes the limit of the number of physical partitions per physical volume specified by factor factor should be between 1 and 16 for 32 disk volume groups and 1 and 64 for 128 disk volume groups If factor is not supplied it is set to the lowest value such that the number of physical partitions of the largest disk in volume group is less than factor x 1016 If factor is specified the maximum number of physical partitions per physical volume for this volume group changes to factor x 1016 Notes 1 If the volume group is created in AIX 3 2 4 1 2 in violation of 1016 physical partitions per physical volume limit this flag can be used to convert the volume group to a supported state This will ensure proper stale fresh marking of partitions 2 factor cannot be changed if there are any stale physical partitions in the volume group 3 Once volume group is converted it cannot be imported into AIX Version 4 3 or lower versions 4 This flag cannot be used if the volume group is varied on in concurrent mode 5 The maximum number of physical volumes that can be included in this volume group will be reduced to MAXPVS factor 6 Change of the volume group may require the modification of the LVM meta data In this situation the volume group will be varied off in management mode to ensure the integrity of the Volume group needing the closure of all open logical volumes in this volume group Since logical volumes in rootvg ca
419. u can specify the Vgid of any physical volume belonging to the volume group that you are redefining i Vgid The volume group identification number of the volume group to be redefined Commands called odmget lqueryvg lvgenmajor lvgenminor getlvodm putlvodm and mknod During normal operations the device configuration database remains consistent with the Logical Volume Manager LVM information in the reserved area on the physical volumes If inconsistencies occur between the device configuration database and the LVM the redefinevg command determines which physical volumes belong to the specified volume group and re enters this information in the device configuration database The redefinevg command checks for inconsistencies by reading the reserved areas of all the configured physical volumes attached to the system A 25 The reducevg command The following summarizes the options for the reducevg command reducevg Removes physical volumes from a volume group Usage reducevg d f VolumeGroup PhysicalVolume s Appendix A High level LVM commands 263 d Deallocates the existing logical volume partitions and then deletes resultant empty logical volumes from the specified physical volumes User confirmation is required unless the f flag is added Attention The reducevg command with the d flag automatically deletes all logical volume data on the physical volume before removing th
420. u have another copy of it somewhere these steps will depend on what caused the corruption 2 7 3 JFS problems Some problems experienced with JFS are mentioned in the following 2 7 3 1 Super block corruption Attempting to mount or fsck file systems with a corrupt superblock may generate messages as such fsck Not an AIX3 file system fsck Not an AIXV3 file system fsck Not an AIX4 file system Chapter 2 Problem determination and recovery 179 180 fsck Not an AIXV4 file system fsck Not a recognized file system type mount invalid argument If this occurs sometimes we can recover the superblock from a secondary backup copy stored further on down the file system To copy the backup superblock over the primary superblock run fsck p dev 1v00 On versions of AIX before version 4 this is done by running dd count 1 bs 4k skip 31 seek 1 if dev lv00 of dev 1v00 Rechecking the file system is then done with fsck dev 1voo If this procedure fails the file system will have to be re created and user data restored from backup 2 7 3 2 File system corruption If the file system itself is corrupt this can sometimes be fixed by running fsck y Note that you cannot reliably fsck a mounted file system To run fsck on the AIX system file systems in rootvg will therefore require booting into maintenance mode and importing the root volume group without mounting file systems from the maintenance menus 2 7 3 3 Corrupt log devic
421. u2 odmvg 2 e Mirrored and striped logical volume a sample may be created for examination with mklv y mirrorstripelv S64K u2 c2 odmvg 2 This section is intended as a checklist of the ODM objects required for the dependent LVM objects to function correctly If these are missing or corrupt it may be necessary to repair them with high or low level LVM commands or as a last resort with the low level ODM commands 2 6 6 1 Physical volumes Some objects present in the ODM for a particular volume are dependent on the particular kind of disk being used for example SCSI or SSA Since these attributes are not LVM specific they are not shown here If such values are required they should be obtained by comparison with a known good disk of the same kind Chapter 2 Problem determination and recovery 99 Physical volume ODM classes Physical volume ODM object CuAt name hdisk11 attribute size _in mb value 2255 type R generic D rep nr nls index 60 CuAt name hdisk11 attribute pvid value 00017d37e671 e4b0000000000000000 type R generic D rep Wot nls index 15 CuDvDr resource devno valuel 30 value2 8199 value3 hdisk11 Note that PVIDs stored in ODM should be 32 characters long with the last 16 characters set to zero 2 6 6 2 Volume groups Volume group ODM objects are shown below Note that there should be one entry for each physical
422. ude Files field and press the Tab key once to change the default value to yes For example to exclude all the contents of the directory called scratch edit the exclude file to read as follows scratch For example to exclude the contents of the directory called tmp and avoid excluding any other directories that have tmp in the path name edit the exclude file to read as follows A tmp All files are backed up relative to current working directory To exclude any file or directory for which it is important to have the search match the string at the beginning of the line use caret character as the first character in the search string followed by dot character followed by the filename or directory to be excluded If the filename or directory being excluded is a substring of another filename or directory use caret character followed by dot character to indicate that the search should begin at the beginning of the line and or use dollar sign character to indicate that the search should end at the end of the line Appendix A High level LVM commands 255 Calls the mkszfile command which generates the image data file The image data file contains information on volume groups logical volumes file systems paging space and physical volumes This information is included in the backup for future use by the installation process Note Before running the mkszfile command ensure that enough space is availa
423. ume for future reference m MapFile Specifies the exact physical partitions to allocate Partitions are used in the order given in the MapFile parameter Used partitions in the MapFile parameter are not legal since the new logical volume cannot occupy the same physical space as a previously allocated logical volume All physical partitions belonging to a copy are allocated before allocating for the next copy of the logical volume The MapFile parameter format is PVname PPnum1 PPnum2 In this example PVname is a physical volume name for example hdiskO as specified by the system It is one record per physical partition or a range of consecutive physical partitions PPnum is the physical partition number PVname Name of the physical volume as specified by the system PPnum Physical partition number P Modes Specifies permissions file modes for the logical volume special file Appendix A High level LVM commands 245 r Relocate Sets the reorganization relocation flag For striped logical volumes the Relocate parameter must be set to n the default for striped logical volumes The Relocate parameter can be one of the following y Allows the logical volume to be relocated during reorganization This is the default for relocation Prevents the logical volume from being relocated during reorganization s Strict Determines the strict allocation policy Copies of a logical partition can
424. ume group and all its containers and files as specified in the tmp vgdata vgname vgname data file where vgname is the name of the volume group contained within the backup image created by the savevg command The restvg command restores a user volume group The bosinstall routine reinstalls the root volume group rootvg If the restvg command encounters a rootvg volume group in the backup image the restvg command exits with an error If a yes value has been specified in the EXACT_FIT field of the logical_volume_policy stanza of the tmp vgdata vgname vgname data file the restvg command uses the map files to preserve the placement of the physical partitions for each logical volume The target disks must be of the same size or larger then the source disks specified in the source_disk_data stanzas of the vgname data file m Note To view the files in the backup image or to restore individual files from the backup image the user must use the restore command with the Tor x flag respectively Refer to the restore command for more information A 31 The savevg command The following summarizes the options for the savevg command savevg Finds and backs up all file belonging to a specified volume group The file is a symbolic link to the mksysb script Usage savevg b Blocks e f Device i m p v X VGName 270 AIX Logical Volume Manager from A to Z Troubleshooting and Commands b Nu
425. umpdev command the dump devices defined in the SWservAt ODM object class are used The default primary dump device is dev hd6 The default secondary dump device is dev sysdumpnull Note 1 A mirrored paging space may be used as a dump device 2 Do not use a diskette drive as your dump device 3 If you use a paging device only use hd6 the primary paging device AIX Version 4 2 1 or later supports using any paging device in the root volume group rootvg as the secondary dump device You can also use the sysdumpdev command to specify whether or not dumps should be compressed before writing them to the dump device Compressing dumps reduces the size needed for dump devices but may cause the dump process to take longer m Note 1 The savecore command should be used to copy a compressed dump from the dump device to a file 2 The dump compression feature only applies to AIX Version 4 3 2 and later versions Running sysdumpdev in non rootvg volume groups Appendix D Other related commands 383 You can use a dump logical volume outside the root volume group if it is nota permanent dump device For example if the p flag is not specified However if you choose a paging space you cannot copy the dump device unless it is in rootvg During the time you must copy the dump device Only rootvg is active before paging is started The primary dump devices must always be in the root volume group for perma
426. urred for example if a one disk volume group has both copies of the VGDA accidently wiped It is prudent to save the map files for logical volumes regularly and particularly after a change is made such as extending a logical volume or running reorgvg This may be done manually or via cron in which case care should be taken that the map copies are kept up to date A sample script to gather copies of all the maps is given in Appendix E 4 gather_maps on page 396 Again 1slv has a n flag which allows us to target a specific descriptor area 2 6 4 Checking fileset levels When strange or inexplicable behavior is encountered it may be worthwhile to check that the latest levels of AIX filesets are installed These can be downloaded from the IBM support Web site at http www rs6000 ibm com support The most important fileset is bos rte lvm Its level can be checked with the 1slpp command 2 6 5 Checking device availability lsdev Cc disk Or lsdev Cc pdisk provides a quick way to see if an hdisk or SSA pdisk is known and usable to the system If a disk cannot be seen with the 1sdev command the problem exists below the LVM layer lsattr El lt disk gt will check the pvid held in the ODM for a specific physical volume If more disks than expected appear within the ODM SCSI termination should be checked for the relevant bus A missing incorrect or loose terminator may allow signals to bounce up and down the bus
427. use the System Management Interface Tool SMIT smit rmfs fast path to run this command D 27 The savebase command The following summarizes the options for the savebase command savebase Saves information about base customized devices in the Device Configuration database onto the boot device Usage savebase o Path d File v d File Specifies the destination file or device to which the base information will be written 376 AIX Logical Volume Manager from A to Z Troubleshooting and Commands o Path Specifies a directory containing the Device Configuration database Causes verbose output to be written to standard input The savebase command stores customized information for base devices for use during phase 1 of a system boot By default the savebase command retrieves this information from the etc objrepos directory However you can override this action by using the o flag to specify an ODM directory By default the savebase command writes the information it retrieves to the boot disk Alternatively you can use the a flag to specify a destination file or a device such as the dev hdiskO device file The savebase command determines what device information to save using the PdDv base field corresponding to each entry in the CuDv object class Specifically the PdDv base field is a bit mask that represents the type of boot for which this device is a base device The savebas
428. used to build the logical volume name If the prefix is given then a two integer sequence number is generated Names are formed by concatenating the name prefix with a sequence number 290 AIX Logical Volume Manager from A to Z Troubleshooting and Commands If the n flag is used the CuDv class is checked to ensure the name does not already exist B 7 The getlvodm command The following summarizes the options for the getlvodm command getlvodm Is an object file command that volume group and logical volume gets information from the ODM This command is undocumented Usage getlvodm a LVDescriptor B LVDescriptor b LVID C c LVID d VGID e LVID F G LVID g PVID h j PVID L VGDescriptor LVDescriptor m LVID N LVID P p PVDescriptor Q VGDescriptor R r LVID S s VGDescriptor T VGDescriptor t VGID u VGDescriptor v VGDescriptor w VGDescriptor y LVID X VGDescriptor x VGDescriptor a LVDescriptor Return the logical volume name given the LVDescriptor which can be a logical volume name or LVID B LVDescriptor Return the label for this logical volume descriptor b LVID Return the volume group name for this LVID C Return all the configured physical volumes c LVID Return the logical volume allocation characteristics for this LVID The order is as follows logical volume type intra allocation policy flag inter allocation pol
429. ut the change to standard output and continues If unsuccessful in finding a valid value for Number the restore Command writes an error message to standard error and exits with a non zero return code Larger values for the Number parameter result in larger physical transfers from the tape device The value of the b flag is always ignored when the restore command reads from diskette In this case the command always reads in clusters that occupy a complete track Indicates that if the File parameter is a directory all files in that directory should be restored This flag can only be used when the archive is in file name format Appendix D Other related commands 367 f Device Specifies the input device To receive input from a named device specify the Device variable as a path name such as dev rmt0 To receive input from the standard output device specify a minus sign The minus feature enables you to pipe the input of the restore command from the dd command You can also specify a range of archive devices The range specification must be in the following format dev deviceXXX YYY where XXX and YYY are whole numbers and XXX must always be less than YYY for example dev rfd0 3 All devices in the specified range must be of the same type For example you can use a set of 8 mm 2 3 GB tapes ora set of 1 44 MB diskettes All tape devices must be set to the same physical tape block size If the Dev
430. utput Note Logical volume commands also cannot read or write to or from logical volumes in a volume group varied on with the s flag Logical volumes that attempt to write to a logical volume in a volume group varied on with the s flag Such as chvg Or mklvcopy may display error messages indicating that they were unable to write to and or read from the logical volume Appendix A High level LVM commands 281 Varies on a volume group but leaves the disks that make up the volume group in an unlocked state Use this flag as part of the initial varyon of a dormant volume group This flag only applies to AIX Version 4 2 or later m Note AIX Version 4 2 or later provides the flags b and u for developers who use n tailed DASD systems The base design of LVM assumes that only one initiator can access a volume group The HACMP product does work with LVM in order to synchronize multi node accesses of a shared volume group However multi initiator nodes can easily access a volume group with the band u flags without the use of HACMP Your must be aware that volume group status information may be compromised or inexplicably altered as a result of disk protect locking being bypassed with these two flags If you use the band uflags data and status output cannot be guaranteed to be consistent The varyonvg command activates the volume group specified by the VolumeGroup parameter and all associated logical volumes A v
431. vailable for allocation The logical volume modified with this command uses the Copies parameter as its new copy characteristic The data in the new copies are not synchronized until one of the following occurs The k option is used the volume group is activated by the varyonvg command or the volume group or logical volume is synchronized explicitly by the syncvg command Individual logical partitions are always updated as they are written to The default allocation policy is to use minimum numbering of physical volumes per logical volume copy to place the physical partitions belong to a copy as contiguously as possible and then to place the physical partitions in the desired region specified by the a flag Also by default each copy of a logical partition is placed on a separate physical volume Appendix A High level LVM commands 253 Note To create a copy of a striped logical volume all active nodes using the volume group must be at least AIX Version 4 3 3 or later Older versions of AIX will not be able to use the volume group after a mirror copy has been added to the striped logical volume TE Note There is an undocumented option to the mklvcopy command that suppresses the display of the warning when you try to add a copy toa striped logical volume Use this command if you want to automate the creation of copies for striped logical volumes without having to acknowledge the fact that you will not be able to imp
432. value is 512 The number represented by the Number parameter must be equal to or less than the number represented by the Maximum variable The maximum number of logical partitions per logical volume is 32 512 y NewLogicalVolume Specifies the logical volume name to use instead of using a system generated name Logical volume names must be unique system wide name and can range from one to 15 characters If the volume group is varied on in concurrent mode the new name should be unique across all the concurrent nodes the volume group is varied on The name cannot begin with a prefix already defined in the PdDv class in the Device Configuration Database for other devices Y Prefix Specifies the Prefix to use instead of the prefix ina system generated name for the new logical volume The prefix must be less than or equal to 13 characters The name cannot begin with a prefix already defined in the PdDv class in the Device Configuration Database for other devices nor be a name already used by another device VolumeGroup Volume group in which to create the logical volume NumLP Number of logical partitions to allocate to the logical volume Commands called allocp getlvname getlvodm lcreatelv lextendlv lquerypv lqueryvg putlvcb putlvodm and rmlv The mklv command creates a new logical volume within the VolumeGroup For example all file systems must be on separate logical volumes The mklv comm
433. value of EPERM if a failure occurs ro Specifies that the mounted file is read only The default value is rw rw Specifies that the mounted file is read write accessible The default value is rw The mount command instructs the operating system to make a file system available for use at a specified location the mount point In addition you can use the mount command to build other file trees made up of directory and file mounts The mount command mounts a file system expressed as a device using the Device or Node Directory parameter on the directory specified by the Directory parameter After the mount command has finished the directory specified becomes the root directory of the newly mounted file system Only users with root authority or members of the system group and have write access to the mount point can issue file or directory mounts The file or directory may be a symbolic link The mount command uses the real user ID not the effective user ID to determine if the user has appropriate access System group members can issue device mounts provided they have write access to the mount point and those mounts specified in the etc filesystems file Users with root user authority can issue any mount command Users can mount a device provided they belong to the system group and have appropriate access When mounting a device the mount command uses the Appendix D Other related commands 363 364 Device parameter as the n
434. ve up to two copies in a logical volume If the PhysicalVolume parameter is used then only copies from that physical volume will be removed A 30 The restvg command The following summarizes the options for the restvg command restvg Restores the user volume group and all its containers and files Usage restvg b Blocks f Device q s n p PPsize PhysicalVolume s b Blocks Specifies the number of 512 byte blocks to read in a single input operation If this parameter is not specified the default of 100 is used by the restore command Larger values result in larger physical transfers to tape devices 268 AIX Logical Volume Manager from A to Z Troubleshooting and Commands f Device Specifies the device name of the backup media The default is dev rmto n Specifies that the existing MAP files are ignored The n flag overrides the value of the EXACT_FIT field in the logical_volume_policy stanza of the vgname data file p PPsize Specifies the number of megabytes in each physical partition If not specified restvg uses the best value for the PPsize dependent upon the largest disk being restored to If this is not the same as the size specified in the vgname data file the number of partitions in each logical volume will be appropriately altered with respect to the new PPsize If a PPsize is specified that is smaller than appropriate for the disk sizes the larger PPsize w
435. vice is an ordinary file or a block special file the u and r flags backspace However raw magnetic tape devices do not support backspacing So when the storage device is a raw magnetic tape the uand r flags rewind the tape open it and then read it again Records are one block long on block magnetic tape but they are typically less than half as dense on raw magnetic tape As a result although a blocked raw tape must be read twice the total amount of tape motion is less than when reading one block records from a block magnetic tape once The structure of a streaming tape device does not support the addition of information at the end of a tape Consequently when the storage device is a streaming tape the uand r flags are not valid options An attempt to use these flags results in the following error message tar Update and Replace options not valid for a streaming tape drive No recovery exists from tape errors The performance of the tar command to the IBM 9348 Magnetic Tape Unit Model 12 can be improved by changing the default block size To change the block size enter the following at the command line chdev 1 lt device name gt a block size 32k D 32 The umount command The following summarizes the options for the umount command umount Unmounts a previously mounted file system directory or file Usage unmount t Type umount f a all allr Device Directory File Fil
436. volume group that is activated is available for use When a volume group is activated physical partitions are synchronized if they are not current A list of all physical volumes with their status is displayed to standard output whenever there is a discrepancy between the Device Configuration Database and the information stored in the Logical Volume Manager As a result the volume group may or may not be varied on The list must be examined and proper action taken to preserve the system integrity With quorum on and if the volume group cannot be varied on due to a loss of the majority of physical volumes a list of all physical volumes with their status is displayed This is also true if quorum is off and not all physical volumes are available For the command line options see varyonvg in Appendix A 39 The varyonvg command on page 280 or use the SMIT fastpath smit varyonvg SMIT provides the following fields Volume group name The name of volume group to be activated F4 for list Chapter 1 LVM commands 5 Resynchronize stale partitions The resynchronization of any stale mirror copies can be done during the vary on process Activate into system management mode Logical volumes can be operated on but not opened for input output Force activation This will force the vary on process even if some physical volumes are missing Warning Data integrity is not guaranteed Vary on VG in Concurrent Mode Used for concurren
437. volume incorporated in the volume group Volume group ODM classes Volume group ODM objects Cuat name odmvg attribute pv value 00017d375243d4020000000000000000 type R generic rep sl nls_index 0 100 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Volume group ODM objects CuAt name odmvg attribute vgserial_id value 00017d375251870e type R generic D rep n nls_index 637 Cuat name odmvg attribute timestamp value 37e3ebf429732e31 type R generic DU rep Not nls index 0 CuDv name odmvg status 1 chgstatus 1 ddins oun location parent connwhere PdDvin logical volume vgsubclass vgtype CuDvDr resource ddins valuel odmvg value2 41 value3 CuDvDr resource devno valuel 41 value2 0 value3 odmvg The ODM tracks the timestamp of the most recent VGDA for the volume group with an object having attribute timestamp This is set by the high level commands and is typically the last action performed by a high level script before completion and cleanup If a high level command fails it will not be changed and should thus hold a timestamp we can compare with the VGDAs on disk to see if they were altered by the failing command Chapter 2 Problem determination and recovery 101 The code used to change the ODM timestamp is of
438. w in 2 3 Corruption example 1 Simple ODM corruption on page 66 that file system log devices and mount points were lost when we exported imported and ran a mklv The mklv caused us to lose our LVCB information and when an LVCB is corrupt we are in the same situation If the LVCB is deleted importvg will still define the logical volume to the system that is accessing this volume group and the user can still access the raw logical volume However any JFS information is lost until the LVCB is re created We may also need to create a mount point manually if we are running importvg on a new system On some old versions of LVM if the LVCB was damaged and the volume group was taken to a new system and imported LVM could become confused about the number of copies existing and it was necessary to rmlvcopy and mklvcopy to re synchronize the ODM This is not the case with the current implementation as the code is smart enough to realize the number of copies from the partition maps inside the VGDA A corrupted LVCB will not cause problems for syncvg If the integrity of the data that has overwritten the LVCB is not an issue the following commands will rebuild the LVCB from the ODM echo AIX LVCB 0 dd of dev hd bs 1 count 9 updatelv lv_name vg_name The updatelv gets information from the ODM to re create the LVCB This is inadvisable unless you have identified what data is in the LVCB and taken the appropriate steps to ensure yo
439. when cfgmgr starts methods to probe the bus All necessary terminators for the configuration should be in place 2 6 6 Checking the ODM ODM data can be interrogated with odmget lt class file gt Specific objects or groups of objects may be obtained by using the q flag This is followed by 98 AIX Logical Volume Manager from A to Z Troubleshooting and Commands criteria which may include wild card searches using the little known LIKE comparison operator This is documented in the AIX base documentation A quick and dirty alternative to specifying criteria is to use grep p to pull stanzas out of the odmget output The investigator should check that the basic requirements for physical volumes volume groups and logical volumes are met within the ODM The ODM s relationship with the LVM is described in the first volume of this redbook AIX Logical Volume Manager from A to Z Introduction and Concepts SG24 5432 From a problem determination perspective we can distinguish certain basic objects ODM stanzas that are required for normal functioning of LVM The following section lists the basic objects that will be created as part of the configuration or creation of a e Physical volume brought in with cfgmgr e Volume group mkvg ft2 y odmvg hdisk6 hdisk7 hdisk8 hdisk9 e Simple logical volume mklv y simplelv odmvg 1 e Mirrored logical volume mklv y mirrorlv c 2 odmvg 1 Striped logical volume mklv y stripelv S64k
440. whether the logical volume is read write or read only The value is as follows 1 Read write 2 Read only r Relocation Sets whether there will be bad block reallocation The value is as follows 1 Bad block reallocation 2 No bad block reallocation v WriteVerify Sets the write verify flag The value is as follows 1 Write verify on 2 Write verify off w MWConsistency Sets the mirror write consistency flag The value is as follows 1 Mirror write consistency is on 2 Mirror write consistency is off 296 AIX Logical Volume Manager from A to Z Troubleshooting and Commands B 13 The Icreatevg command The following summarizes the options for the 1createvg command Icreatevg An object file command that creates a new volume group and installs the first physical volume in the volume group Usage Icreatevg a VGname V VGmajor N PVname n MaxLV D VGDASize s SizePP ft a VGname The volume group name for the new volume group VGname is the name of the dev entry V VGmajor The major number to be assigned to the new volume group N PVname The name of the first physical volume to install in the volume group n MaxLv The maximum number of logical volumes in the volume group an integer between 0 and LVM_MAXLVS 256 D VGDASize The size of the VGDA An integer between 32 and 8192 where n is the number of blocks 512 bytes to be reserved for one copy of th
441. wo mirror write consistency cache MWCC areas on each hdisk We can see the first one is running from bytes 02000 to 02767 that is decimal 512 to 1527 The alternate MWCC area immediately follows this 0002000 7 4 027 233 Y H o 0 o o 0 0 0 o 37e4 179b 2a0c 0548 0000 0000 0000 0000 0002020 0 o o o 0 O 0 o 0 0 0 o 0 o o g 0000 0000 0000 0000 0000 0000 0000 7 4027 233 f AH 37e4 179b 2a0c c548 o o o o 0 0 io o 0 0 0 0 0000 0000 0000 0000 0000 0002760 0 0 o o o o o O 0003000 0 0 0 o 0000 0000 0000 MWCC Area 1 Figure 12 Contents of the first physical sectors on an hdisk Blocks 2 and 3 The MWCC We now see the extremely important LVM information record This area should contain pointers to our VGDAs and VGSAs as well other key volume group information such as our VGID PP size LVM version level and so forth These items are documented in usr include sys lvmrec h 122 AIX Logical Volume Manager from A to Z Troubleshooting and Commands LVM_LVID Ivmarea_len vgda_len vg_id vgda_psn 0 0007000 R Q 207 016 0 0 sie 2 me 0000 UW vg Sa_psn 1 version Figure 13 Contents of the first physical sectors on an hdisk Block 7 The LVM information record There is a backup of this sector in sector 70 As you can see vgda_lenis 0x832 equal to 2098 so we have a small as opposed to big VG here vgd
442. xamples contain the names of individuals companies brands and products All of these names are fictitious and any similarity to the names and addresses used by an actual business enterprise is entirely coincidental Reference to PTF numbers that have not been released through the normal distribution process does not imply general availability The purpose of including these reference numbers is to alert IBM customers to specific information relative to the implementation of the PTF when it becomes available to each customer according to the normal IBM PTF distribution process The following terms are trademarks of the International Business Machines Corporation in the United States and or other countries AIX AS 400 CUA DB2 IBM Netfinity NetView RS 6000 System 390 400 The following terms are trademarks of other companies C bus is a trademark of Corollary Inc in the United States and or other countries Java and all Java based trademarks and logos are trademarks or registered trademarks of Sun Microsystems Inc in the United States and or other countries Microsoft Windows Windows NT and the Windows logo are trademarks of 402 AIX Logical Volume Manager from A to Z Troubleshooting and Commands Microsoft Corporation in the United States and or other countries PC Direct is a trademark of Ziff Communications Company in the United States and or other countries and is used by IBM Corporation under license ActionMedia
443. y His areas of expertise include system performance and customization scripting and high availability Thanks to the following people for their invaluable contributions to this project Gerald McBrearty LVM developer Ram Pandiri LVM developer Johnny Shieh LVM developer Mathew Accapadi AIX performance engineer Mike Wortman AIX file system developer Comments welcome Your comments are important to us We want our Redbooks to be as helpful as possible Please send us your comments about this or other Redbooks in one of the following ways e Fax the evaluation form found in IBM Redbooks review on page 415 to the fax number shown on the form e Use the online evaluation form found at http www redbooks ibm com e Send your comments in an Internet note to redbookeus ibm com x AIX Logical Volume Manager from A to Z Troubleshooting and Commands Chapter 1 LVM commands This chapter covers common tasks used in administering volume groups and the commands that are used This chapter will cover areas such as how to create list and modify volume groups logical volumes and file systems To use the commands effectively requires an understanding of the concepts covered in the redbook AIX Logical Volume Manager from A to Z Introduction and Concepts SG24 5432 The details of the command line options are contained in Appendix A High level LVM commands on page 199 Most of the commands can be accessed from
444. y settings will have to be re created manually Chapter 2 Problem determination and recovery 155 The lowfs2 file system is easier to fix since all we have to do is re create the mirror The safest way to do this is by running mklvcopy Although this costs disk I O we then have a known good mirror These final repair commands are shown as follows chfs a log dev lowlog lowfs1 gt mount lowfs1 1s l lowfs1 ls lowfs1 datal lost found mount lowfs2 ls lowfs2 data2 lost found mklvcopy k m lowlv2 map lowlv2 2 lsvg M lowvg lowvg hdisk6 1 215 hdisk6 216 lowlv1 1 hdisk6 217 lowlv1 2 hdisk 6 218 lowlv1 3 hdisk 6 219 lowlv2 1 2 hdisk 6 220 lowlv2 2 2 hdisk6 221 1075 hdisk7 1 215 hdisk7 216 lowlv2 1 1 hdisk7 217 lowlv2 2 1 hdisk7 218 lowlog 1 hdisk7 219 1075 a J The delicate procedure of VGDA repair is now complete VGDA corruption with two VGDAs on one disk For the case of a unique physical volume in a volume group one might think that a command such as echo c dd bs 1 seek echo ib 8 n 211000 bc of dev hdisk6 would bring corruption However LVM is very resilient in such a case and lquerylv can still be used to pull the maps out of the first VGDA All the following examples assume a Starting point of volume group corrvg on hdisk6 containing logical volume corrlv 156 AIX Logical Volume Manager from A to Z Troubleshooting and Commands echo c
445. y created This way the allocp logic will ensure that the same physical partitions are used AIX Logical Volume Manager from A to Z Troubleshooting and Commands 3 1 8 Step 8 Re create any unmirrored logical volumes and restore the data from backup 3 1 9 Step 9 Re synchronize the mirror copies syncvg p hdiskn If you are running AIX Version 4 3 3 or above you may want to take advantage of the syncvg P Nump option where Nump is the number of physical partitions to synchronize in parallel 3 2 Using the replacepv command Using the replacepv command is a much simpler option The original definition of the failed disk is left and the physical volume itself is replaced 3 2 1 Description of the test environment In our test environment we have used an F50 machine with five disks Here is the position and the name of these disks hdis hdis hdis hdis hdis hdig hdis hdis hdis lspv kO 0004163128f3de5a k1 000416314bd724be k2 0004163192b1d7 5 k3 000416314bd749 8 k4 000416314bdada38 lsdev Ccdisk kO Available 10 60 00 8 0 16 Bit SCSI Disk Drive kl Available 10 60 00 9 0 16 Bit IVD SCSI Disk Drive k2 Available 10 60 00 10 0 16 Bit IVD SCSI Disk Drive k3 Available 30 58 00 8 0 16 Bit SCSI Disk Drive hdis N rootvg None None None None k4 Available 30 58 00 9 0 16 Bit IVD SCSI Disk Drive We then create a volume group spanning on three 4 5 GB disks Chapter 3
446. y to the logical volume LVName s control block t Type Writes the logical volume type to the logical volume LVName s control block u UpperBound Writes the upperbound allocation policy to the logical volume LVName s control block N This option indicates a new logical volume control block is being written If this flag is not set then a control block must already exist on the logical volume to be updated v VGName Writes the volume group name to logical volume LVName s control block x VGAutoOn Writes the volume group auto_on value to the logical volumes LVName s control block LVName The name of the logical volume for which the control block will be updated created The putlvcb command writes the control block information into block 0 of the logical volume lvname Only the fields specified are written putlvcb can be used to write a new control block or update an existing one B 39 The putivodm command The following summarizes the options for the put1lvodm command putivodm An object file command that puts logical volume data values into the Configuration Database Usage putlvodm a IntraPolicy B Label c Copies e InterPolicy LVName n NewLVName r Relocatable s StrictState t Type u UpperBound y CopyFlag z Size LVID Appendix B Intermediate level commands 311 a IntraPolicy Sets the intra physical volume allocation policy the posit
447. ys be less than YYY for example dev rfd0 3 All devices in the specified range must be of the same type For example you can use a set of 8 mm 2 3 GB tapes ora set of 1 44 MB diskettes All tape devices must be set to the same physical tape block size If the Device variable specifies a range the backup command automatically goes from one device in the range to the next After exhausting all of the specified devices the backup command halts and requests that new volumes be mounted on the range of devices Specifies that files be read from standard input and archived by file name If relative path names are used files are restored with the restore command relative to the current directory at restore time If full path names are used files are restored to those same names 322 AIX Logical Volume Manager from A to Z Troubleshooting and Commands L Length Specifies the length of the tape in bytes This flag overrides the c d and s flags You can specify the size with a suffix of b k m or g to represent Blocks 512 bytes Kilo 1024 bytes Mega 1024 Kilobytes or Giga 1024 Megabytes respectively To represent a tape length of 2 Gigabytes enter L 2g This flag only applies to AIX Version 4 2 and above Note Use the t flag for i node backups only Limits the total number of blocks to use on the diskette device The value specified must be a non zero multiple of the number of sectors per diskett
448. ytes in the file being processed data can be lost Swaps halfwords This flag is usable only with the cpio i command Note If there is an odd number of halfwords in the file being processed data can be lost Creates a table of contents This operation does not copy any files Copies unconditionally An older file now replaces a newer file with the same name Lists file names If you use this with the t flag the output looks similar to that of the 1s 1 command Processes an old file for example one written in UNIX Sixth Edition format This flag is usable only with the cpio i command Appendix D Other related commands 333 334 m Note 1 If you redirect the output from the cpio command to a special file device you should redirect it to the raw device and not the block device Because writing to a block device is done asynchronously there is no way to know if the end of the device is reached 2 The cpio command is not enabled for files greater than 2 Gig in size due to limitations imposed by XPG 4 and POSIX 2 standards 3 cpio does not preserve the sparse nature of any file that is sparsely allocated Any file that was originally sparse before the restoration will have all space allocated within the file system for the size of the file 4 You can copy special files only if you have root user authority 5 All the flags must be listed together without blanks between them
449. yvg PROGAME exportvg COWDIR etc dojrepos ODMDIR etc dojrepos export ADR export ADR Parse comand line arguments set getopt get t badvg set badwg if S 1 0 Determine if there is a syntax error then dgomsg s 1 amdlvm cat 606 lvmmeg 606 gt n exportvg gt amp 2 dgomsg s 1 ardlvm cat 760 lvmsg 760 gt n SPROGNAVE gt amp 2 edt fi 0 0 shift LF shift J Here we see the script set up its environment the traps and other miscellaneous shell variables The command line arguments are also parsed with getopt In the case of exportvg this is trivial but as will be seen a far more elaborate process occurs with the other high level commands Following this a process of validation begins for the input parameters and gathering and testing the various LVM and ODM structures required for the script s operation 68 AIX Logical Volume Manager from A to Z Troubleshooting and Commands fie n 1 J 1 voname argument cn commend line then VQAVE S1 Determine the major nmr of the root device file ls 1 dev IPL rootvg sed s g gt tmp rootdevicess read skipfld skipfld skipfld skipfld ROOIMAIR skipfld lt tmp rootdevic ess VGID getlvodm v SVGNAVE test_retum check for error retum Determine the major nmr of the volume group entered V V
450. ze of the tape in feet by using the second form in which case the tar command assumes a default Density variable The third form allows you to specify both tape length and density Feet are assumed to be 11 inches long to be conservative This flag lets you deal more easily with multivolume tape archives where the tar command must be able to determine how many blocks fit on each volume Notes 1 Tape drives vary in density capabilities The Density variable calculates the amount of data a system can fit on a tape 2 When using 1 4 inch tape devices be sure to take into account the number of tracks on the tape device when specifying the value for the Feet variable For example a 4 track 1 4 inch tape drive with a 600 foot tape anda density of 8000 bpi can be specified using the s Feet Density flag as follows S 2400 8000 where 600 feet multiplied by 4 tracks equals 2400 feet Tries to create a symbolic link if the tar command is unsuccessful in its attempt to link regular link two files with the s flag Lists the files in the order in which they appear in the archive Files can be listed more than once Adds the files specified by one or more File parameters to the end of the archive only if the files are not in the archive already or if they have been modified since being written to the archive The u flag is not valid for any tape devices because such devices do not support the addition of informat
451. zed for example dev jfslog1 Note The only intended use for the logform command is to initialize a JFS log logical volume as a JFS log device The SMIT interface for creating a JFS and the crfs command allow only one JFS log device per volume group D 19 The logredo command The following summarizes the options for the logredo command logredo An object file command that uses the jfs log to re establish consistency in the specified file system Usage logredo b n filename b Replay the log file to a backup copy N Set debug mode The logredo algorithm reads the log backwards from logend to the value specified by the first log syncpt record encountered D 20 The Isfs command The following summarizes the options for the 1sfs command Appendix D Other related commands 357 Isfs Displays the characteristics of file systems Usage Isfs q c I a v VFSType u MountGroup FileSystem s a Lists all file systems default C Specifies that the output should be in colon format l Specifies that the output should be in list format q Queries the logical volume manager LVM for the logical volume size in 512 byte blocks and queries the JFS superblock for the file system size the fragment size the compression algorithm if any and the number of bytes per i node nbpi This information is not reported for other virtual file syst

AIX Logical Volume Manager from A to Z: Troubleshooting and

Contents

Download Pdf Manuals

Related Search

Related Contents