Home

EdgeWare OS User Manual - Cutting Edge Networked Storage

1. toot 0 Unlimited Unlimited 4 Unlimited Unlimited Backup grape 0 10000 42000 p 110000 12000 Ca Edit Quota For Browse gt EditGraceTimes gt Check Quotas gt Use the settings below to automatically set quotas on this filesystem when a new Local user is created Soft Block Limit Unlimited Soft File Limit Unlimited O Hard Block Limit Unlimited Hard File Limit Unlimited Apply gt Return to index Figure 39 Filesystem Quotas page for Users Page 57 MANAGING DISK QUOTAS 1 To manage disk quotas for users or groups click the relevant link under the file system column for the Volume you want to manage quotas for 2 This will lead you to the file system quota page for that logical volume Here you can set hard and soft limits for users on the volume Clicking on the button browse will give you a pop up browser window where you can select a user to specify quotas for 3 After you have selected or entered a user in the test box click on Edit Quotas 4 The next page Figure 40 lets you specify the hard and soft limits for the user You can specify limits based on number of blocks where 1 block here is approximately equal to 1 Kilobyte and or on number of files owned by the user 5 When you are done click on Update to apply the changes e Storage DO RAID We Manager Share Manager GG wou a Backup ax ee e whe
2. EdgeWare 3 0 0 1ce NAS Servers HELP iz D index Currently Scheduled Mirrors Identification _ Server to mirror Async Share to mirror Local Volume Destination 141661221678681 CE485192 shares Local Volumes L local OTTERS Sree Rowen Figure 108 Remote Replication configured HANDSHAKE High speed cluster operations such as synchronous remote mirroring and failover actually treat the pair of servers as a unit In order to aid the high bandwidth communication required between parts of this unit they often use special hardware such as point to point Ethernet connections on ports that are not being used for the local network Each system has to be capable of being master since recovery from neighbor failure without suffering downtime is the main purpose of the operation The setting up of the tightly connected pair that satisfies all this is called the handshake and must be complete before the actual deployment of the feature As an example of a handshake this subhead offers a two server cluster each using two Ethernet ports Each of the systems Leader22 and Follow22 connects to the same local network via eth0 Their eth1 ports connect to each other by a point to point cable The handshaking steps are as follows A Set up hardware configuration Open UI window for each server Impose static IP addresses netmasks and routes Impose hostnames for each port of each server Record all ports in both servers
3. DESIGNING A SNAPSHOT STRUCTURE The snapshot facility of the EdgeWare OS allows freedom to the user both to back up stored data in a known timed state and to be able to recover older versions of the data This freedom creates a burden on storage and CPU resources when it is used Therefore a design strategy is needed before the user plunges into the deployment of a long term snapshot structure This strategy is based on knowledge of the behavior of the snapshot machinery Key facts include e Snapshots randomize data writes This impacts you if you are counting on frequently writing big files like movies rapidly to EdgeWare storage a situation called long sequential writes Although EdgeWare mitigates this as much as possible by putting snapshot volumes on separate physical volumes when it can the snapshot mechanism causes much head movement when writing and even one snapshot can slow long sequential writes by a large factor EdgeWare tests September 2008 showed these falling to around 6 MB s On the other hand if like most customers you are running a database or mail server or other application that randomly accesses data one or two snapshots will not slow you much Also if you are mainly reading files even large files from EdgeWare storage snapshots will barely affect your speed e Snapshots multiply data writes If you are creating data or changing very old data then every snapshot requires an extra write to its Snaps
4. Note The configuration for Fibre Channel Arrays under EdgeWare 3 2 will only show for those Systems which have installed Xyratex hardware In systems with Xyratex hardware controllers you can use hardware RAID to create redundant storage just as with 3Ware Older systems shipping out with hardware controllers had the Operating System installed on a hardware RAIDS set configured from the factory In this case you did not have the option to delete and recreate hardware RAID since doing so would render the appliance unbootable Current systems have the Operating System residing on a local system disk or mirrored disk pair or on a solid state disk or DOM This isolates the hardware RAID from the operating system dependency and gives flexibility and total control of the hardware RAID storage to the customer A oca mirrored disk pair will appear as SOFTWARE RAID and should not be disturbed as it functions as the system disk Before proceeding with the Xyratex RAID documentation below please read the documentation above on 3Ware Hardware RAID The advice there is applicable to Xyratex and other RAID especially subhead Jn the Event of a Hard Drive Failure However the 3ware Manager described above is replaced by StorView the Xyratex UI which currently must be activated separately on your client not through EdgeWare EdgeWare Storage Configuration H Raid Configuration Provides configuration and maintenance Fibre Channel RAID Manager f
5. by that port The exception is that a firewall on your or the other system may prevent the message from getting through Notice that other ports of yours such as ethl may or may not be on the same subnet If they are not then they cannot hear each other without additional devices such as routers It does not matter if they are physically connected In the above example if ethl has IP address 192 168 2 10 and netmask 255 255 255 0 it is on a separate subnet from ethO even though they have the same netmask This is very important in multiple connectivity operations such as bonding and MPIO It is also used often in conjunction with separate switches to set up private subnets that allow high throughput communication without network contention NETWORK The network section provides configuration options for TCP IP protocols The main page gives a summary of the current network settings of the Storage Server and also allows the administrator to change some network settings Figure 70 When changing these always make the entries in this page last after finishing with changes in the properties of individual network interfaces Figure 71 and after finishing with any Advanced Configuration changes Figure 72 The last selection you should make is Save in Figure 70 Certain changes such as Hostname in Figure 70 and IP Address for ethO in Figure 71 may require a reboot after the Save in order to take effect See Maintenance gt
6. gen 9 Ss 2 Ki S amp A Volume to backup Molumes LVstorage ien Browse v Figure 43 Add a new backup of a volume 3 In the Backup to text field select the backup target you want to use Select Tape To backup to a tape device in the system Tar file To backup to a file on another logical volume Another Volume Live Copy To backup files to another logical volume Page 6 4 For Tar file or Live Copy above select a directory from the drop down list 5 For Tar file you must add a file name to the directory Click on the text window move the cursor to the end of the directory name and type in something like 20080826 LVstorage backup where in place of 20080826 LVstorage you place some identifier of the source and date instance of backup Do not terminate the name with a dot suffix like tar or tgz The restore program can recognize the format 6 For Make multiple volumes of the tape size select Yes if the size of your data is more than a single media and you have to span the backup over more than 1 media To span the backup on multiple media you will have to manually label the media to keep track of the order in which the backup will be written During the backup process once a media becomes full with data the tape drive will eject the media out At this point you have to remove the media and insert another empty media Once another media is inserted the backup process will proceed an
7. iscsiadm modify discovery i disable 4 Add a target server using commands a iscsiadm add discovery address lt IP address of Edgeware Server gt 5 List all targets using command Page 74 a iscsiadm list discovery address Statically connect to a single iSCSI target using command a iscsiadm add static config lt iqn iSCSI target name gt lt IP Address gt b iqn name will be unique to iSCSI target c example iscsiadm add static config iqn 2009 05 ce64c918 storage lvm2 vgstorage lviscsi0000 192 168 1 16 d Statically connect to additional targets as required Enable the static iSCSI service with command a iscsiadm modify discovery s enable Rescan for drives with command a devfsadm i iscsi this may take 15 30 seconds depending on the number of targets logged into Confirm disk availability with command a format PARTITION AND LABEL THE ISCSI DRIVE Format label and partition iSCSI disk a While in the format command select the correct disk by entering its number Page 75 fdisk and pressing enter c Confirm that you want to partition the drive d J i y partition pe by typing partition f Confirm the operation g Type print to display the partition information h Confirm the operation i Quit the format utility by typing quit twice Page 76 CREATE A UFS FILESYSTEM ON THE DRIVE 1 Identify the drive and slice where the new file system will
8. Page 175 E CUTTING rd EDGE ot oe a hee KR R A G E 1 Manufacturer s Warranty Information a Manufacturer Cutting Edge provides a return to factory warranty for all products offered Warranty period unless otherwise specified is for two years from date of shipment on any Cutting Edge hardware Replacement Products will be warranted for the remaining warranty period of the original Products b All parts replaced during the guarantee period shall become the property of Manufacturer c Inspection and repair of defective equipment under this guarantee will be performed only at Manufacturer s plant or any alternate site designated by Manufacturer 1825 Gillespie Way Suite 100 El Cajon CA The Manufacturer will employ commercially reasonable efforts to ship a replacement product within seven 7 working days after receipt of failed part and a Return Materials Authorization RMA request Delivery times will vary depending on customer location Manufacturer at its option may repair or replace the defective product as exclusive warranty remedies hereunder d Product returned to Manufacturer must first be assigned a Return Materials Authorization RMA number The RMA number must appear clearly on the outside of the shipping container Product shipped to Manufacturer without an RMA number will be returned to sender unopened When equipment is returned to the Manufacturer s authorized service facility for repairs the ordering activ
9. options for software and hardware RAID EdgeWare 3 0 0 1ce de a HELP Advanced configuration Controller Information Fibre Channel Arrays After making any changes in the Fibre Channel Arrays please reboot this Storage Server and any other Storage Server sharing the storage RAIDO RAIDO RAID system 0 RAID system 0 Array Last Create Fibre Channel Array of level Raid 0 bs Available drives Enclosure 1 Slot 5 SEAGATE 9QG4J7H2 Enclosure 1 Slot 6 SEAGATE 9QG47254 Enclosure 1 Slot 7 SEAGATE 9QG1JHBS Enclosure 1 Slot 8 SEAGATE 9QG42ZD6 No global spare drives present Return to index Figure 23 Fibre Channel device summary page When you select Fibre Channel Arrays in the RAID module main page you get an icon with a long identifier like 20000050CC204721 page not shown Select this and you get a page like Figure 23 This puts you into the driver s seat for generating arrays out of raw disks under the control of the Xyratex controller The techniques are very much like those for the 3Ware controller described above After adding deleting or changing any Fibre Channel arrays be sure to reboot to make them visible to other parts of EdgeWare Fibre Channel storage is frequently shared between paired EdgeWare CPU motherboard units heads in failover arrangements In such cases also reboot the other head at the same time Page 40 Creating RAID 1 EdgeWare Storage Configuration Raid Configuration S EdgeWare
10. s real device This vastly reduces overhead and makes efficient read and write strategies possible 2 FC Target is normally a point to point protocol one host at a time for each target This avoids the heavy burden of arbitrating between clients permitting efficiency at the cost of some flexibility It also makes it desirable and possible to use dedicated switch fabric connections between host and target for high volume FC Target devices 3 FC Target presents a oca virtual raw device which from the point of view of the application presents no network or file system complexities This is highly desirable for many applications such as databases 4 Unlike iSCSI FC Target does not require special initiator software in its host FIBRE CHANNEL TARGETS A Fibre Channel target is the server peripheral side of the Fibre Channel protocol In EdgeWare like iSCSI it presents oca virtua hard disk storage to another host system EdgeWare can present one or more than one FC target each of which may or may not be connected to an initiator host Each target must have its own dedicated Logical Volume which cannot be used for anything else A FC target volume Page 79 cannot support any NAS protocol or be a source or target of Asynchronous Replication although it is possible for it to be a Synchronous Mirroring source and or a Snapshot parent since those also are block level protocols Currently other network based capabilities such as Fai
11. 1825 Gillespie Way Suite 100 El Cajon CA 92020 800 257 1666 www cuttedge com
12. 257 1666 www cuttedge com r CUTTING EDGE ot oe eae ER A G E R 2 Maintenance Provisions a SERVICES AND SUPPORT PACKAGE Manufacturer will provide to Customer the technical support and maintenance services Services for the following support packages selected by the Customer Support Package Standard Gold Support Offering 1 year Next Business day parts delivery Next business day Field Engineer onsite Phone Email and Web Support are available 7x24x365 SERVICE MODIFICATIONS Manufacturer reserves the right to add change or delete available Services or Support Packages for certain hardware or software platforms and configurations in its reasonable discretion upon 90 days notice If Manufacturer deletes an entire Support Package at any time during the term of this Agreement Manufacturer will refund to Customer a pro rata portion of the Annual Fee paid by Customer for the portion of applicable Services not rendered for the remainder of the term due to the deletion of the Support Package ADVANCE REPLACEMENT Advance Replacement means that replacement products or parts will be shipped to the Customer before Manufacturer receives the product or part with a problem If Advance Replacement service is in effect under the Support Package selected by Customer Manufacturer will use commercially reasonable efforts to replace malfunctioning Covered Product hardware that fails to meet its published s
13. Host Addresses lists Capture each server s ports in the other s NAS Servers Index mmom Dm A SET UP HARDWARE CONFIGURATION Page 142 The two servers should have the following configuration prior to more advanced options to verify usage e The first network port ethO on the back of each Edgeware Storage Server should be connected to the local networking switch e The second network port eth1 should be connected directly to the ethl port on the other EW3 server and by pass any network switch NOTE a crossover cable is not necessary as the hardware auto detects the direct connection After making these hardware connections turn on the servers and wait for them to boot up B OPEN UI WINDOW FOR EACH SERVER On your client machine open a browser instance for each server and log in Figure 109 Use whatever IP addresses the servers currently respond to These may not be the IP addresses that are finally settled upon It is OK at this stage for them to be DHCP or Auto h a Cutting Edge NAS Leader22 192 168 1 130 Microsoft Internet Explorer MWe File Edit View Favorites Tools Help a G nn O DAG Dee km ORE BORS d ess ET https 192 168 1 130 10000 x Efe links voog Ef v Y Bookmarks Bhi blocked Ycheck v N Autolink E Autoril sb Sendtow A O Settings e fe g HBL A Server Storage Networking Authentication Maintenance Clustering Ga O ini tio Server Name Leader22 Server IP Address 192 16
14. Shutdown module later in this Manual Page 88 Hostname in ste HELP 192 168 1 101 K AA N RA Maintenance Clustering LockheedServer Figure 70 Network Settings EdgeWare Se HELP Return to index Figure 71 Interface properties for eth0 Networking Na Authentication ring 192 168 1 158 255 255 255 0 The line items of Figure 70 all important follow Domain Name Network domain where Storage Server is located Page 89 oa ox Manual Is KAEA The name of the Storage Server on the network It is usually desirable to change the default name which is CE followed by six hexadecimal numbers to something descriptive as shown in Figure 70 Network Interfaces eth0 eth1 A summary of the Network Interfaces configured in the system Name of the Media type of IP address of Subnet mask for status network interface the interface the interface the interface You can edit the interface properties by clicking on the name of the interface Figure 71 shows an example of this The values shown in Figure 71 are not default DHCP values but static values entered by hand You must edit the interface properties even in the DHCP case for any non eth0 interface that you want to come up automatically at boot Default router Select None or from DHCP if your interfaces are set DHCP assigned addresses If interfaces are set to stat
15. Specify the Group Name property below then click Save Group Name A more human readable and understandable name for the security group Page 84 om BOS g SS ZG A Networking Authentication Maintenance Log Out since e EdgeWare Storage Configuration FC Target Mode os Configuration a Ra configuration for Fibre Channel gt Add Security Group Target Mode Su Support Storage EdgeWare 3 0 0 1ce Add Security Group Figure 66 Add Security Group page Once the security group is added you will see it in the main FC target configuration page Future 58i There will be two associated buttons to the entry The Remove button removes the security group and the Edit button allows you to assign LUNs and authorized adapters to the security group Fibre Channel Adapters enabled for Target Mode isp3 Z Disabled isp2 6 S Disabled pi 5 201 Disabled Enable ispo 4 300 Enabled Disable Exported Fibre Channel Target Mode Devices FCExport 183296 512 Mev VGExport LVExport Remove LUNs and Security Groups ii Default Add Security Group Figure 67 Security group added to main page To assign LUNs and adapters to the security group hit the corresponding Edit button You will be taken to the Security Group Edit page Figure 68 Page 85 tf D A EdgeWare Server etworking Authentication Maintenance Clustering Log Out es Storage
16. WORKGROUP WINS mode Act as WINS server Use server Neither Master Browser Elections Decline Participate Participant lowest PDC and ADS option Windows 2003 server PDC ADS No O Yes PDC ADS server Administrator Administrator Password ADS option KDC Realm Logging option Debug No O Yes Log File var log samba smbd log Samba Status Samba is stopped Save r Save and Restart F Figure 75 Windows Networking Options This is how share level Samba finds a username to check the password Step 0 If the service is marked guest only yes then steps 1 to 5 are skipped Step 1 If the client has passed a username password pair and that username password pair is validated by the UNIX system s password programs or by checking the SMB encrypted passwords database smbpasswd then the connection is made as that username Note that this includes the server serviceYousername method of passing a username Step 2 If the client has previously registered a username with the system and now supplies a correct password for that username then the connection is allowed Step 3 The client s NetBIOS name and any previously used user names are checked against the supplied password If they match then the connection is allowed as the corresponding user Step 4 If the client has previously validated a username password pair with the server and the client has passed the validation token then that username is used This
17. You now have a raw iSCSI device on your initiator system To use it you need to format it so that it can be seen by your initiator system as an additional local hard drive Page 71 FORMATTING THE iSCSI VIRTUAL DRIVE IN MICROSOFT WINDOWS By now you have created an iSCSI initiator device connected to an iSCSI target but your system needs to see it as an additional disk The following instructions are for Windows XP 1 Click on Start then select Control Panel select Performance and Maintenance if necessary select Administrative Tools and then click on Computer Management 2 Select the Disk Management icon and the Initialize and Convert Disk Wizard should appear Initialize don t convert After it finishes you will see all the drives available in your system see Figure 56 Note other versions of Windows may use the Disk Defragmenter icon instead of the Disk Management icon Computer Management BEE g File Action View Window Help e gt Ame es E Computer Management Local volume Layout Type File System Status Capacity Free Space Free Fault Tolerance Overhead K System Tools Sc Partition Basic NTFS Healthy System 232 68 GB 222 28GB 95 No ga Event Viewer kl Shared Folders z Local Users and Groups Performance Logs and Alerts 4 Device Manager amp Storage 1 Removable Storage Prick NEE Disk Defragmenter Basic C Z Disk
18. including the EdgeWare Server Double click the EdgeWare Server It will bring up a browser and the IP address will be apparent on the Connected to comment on the bottom of the screen You can continue and actually log in to your Server if you wish Please note that on a LAN the Network Discovery tool will also find other EdgeWare Servers and non EdgeWare systems if they are present TROUBLESHOOTING STORAGE SERVER DISCOVERY If you have another EdgeWare Server in your network always try the NAS Servers Index Use another tool Magellan has been known to work in cases where uPnP was unable to find an address for instance A Magellan CD can be provided by Cutting Edge Technical Support if your original one cannot be found If no Storage Server discovery tool works try the serial number name search If the serial number name search does not work and you are on a non DHCP network or connection try the default Server IP address 169 254 5 58 If there is still no result at this point you are justified in returning to step B Troubleshooting a failed server IP address creation in the preceding section of this manual Page 18 OVERVIEW OF THE EDGEWARE STORAGE SYSTEM QUICK START WEB MANAGEMENT Once the IP address is known the GUI management interface is accessible to reconfigure the network settings for your desired layout or to initialize the storage and network file services The base for this GUI is the popular Open
19. monitor 5m wd Sun Sat Add serice Figure 88 Detail list of services being watched To modify an existing service monitor click on that service under Service name You will be presented with a page such as is shown in Figure 89 DEFAULT WATCH LIST The default watch list has the following services being monitored Free space Free space on the logical volumes is monitored by the monitor freespace monitor In the monitor parameters you have to specify the logical volume you want to check for free space status and at what percentage of the logical volumes free space the monitor will send an alert The format is Page 119 Watched service details Name of service RAID Check every 5 minutes Description Monitors RAID set degradation and rebuild status Using monitor RAID monitor v Monitor parameters Monitoring period 1 Delete this period Alerts for period Alert Run when Additional parameters alarm alert Service failed vll3 Alert Run when Additional parameters hdl Service failed KAN Add a new period Save Lei Advanced Configuration H Delete gt Figure 89 Edit a service monitor Volumes logical volume name percentage of free space For example Volumes LVstorage 20 specifies that the free space monitor will send out the defined alerts when the LVstorage logical volume free space becomes less then 20 of LVstorage total size RAID Hardware RAID service is monitore
20. needs to be performed on the logical volume and you don t want to unmount a logical volume on a production unit which has data being modified Another use arises when you want to have the option of accessing old data even though you are expecting the data to change When the need for the snapshot is done the system administrator can just remove it without in any way disturbing the parent logical volume This facility does require that the snapshot be made at a time when the data on the logical volume is in a consistent state Many databases offer this capability The snapshot device is a read only copy of another parent logical volume that contains all the data that was in the parent at the time the snapshot was created This means we can back up that volume without having to worry about data being changed while the backup is going on and we don t have to take the volume offline while the backup is taking place The Snapshot Volume found in the Volume Manager tree is a logical block device that is needed to keep the snapshot device current The snapshot works by making a note in the Snapshot Volume of all changes in the parent so that the snapshot device can back out of the changes and recreate the original state whenever it is read The bigger the Snapshot Volume is the longer the snapshot device can survive without overflowing or filling up with change data and becoming unusable Even if overflow happens the parent logical volume is unharmed
21. 13 33 events 1 2 root 13 33 events 2 13 root 13 33 events 3 14 root JEE khelper 18 root 13 33 kthread 53 root 13 33 kblockd 0 54 root 13 33 kblockd 1 55 root 13 33 kblockd 2 56 root 1833 kblockd 3 57 root 13 33 kacpid 253 root 13 33 ata D Figure 93 Top part of Processes module page Page 126 COMMAND LINE The Command Line module in the Maintenance section provides an encrypted command line access to the system through your web browser You may use this interface to call commands just as you would via a keyboard console interface or SSH session The commands issued from the Command Line module however must satisfy two requirements e They must not require a response as the Command Line module is NOT interactive e They must not require interruption to stop Be sure your command satisfies both these requirements before using this powerful EdgeWare capability Otherwise you will lose the Command Line screen and may fill up a file with redirected output A good example of a usable command is ping c 3 192 168 1 13 to find out whether the network connection to 192 168 1 13 is up and running Important Here the ce 3 is needed because interruption Ctrl C is not possible The bare ping command will never complete and no results will ever show up on the EdgeWare screen The EdgeWare Storage Server also includes a secure shell SSH program that allows SSH clients to log into a console over the network This
22. 3 0 0 1ce HELP Select drives from the list below Enclosure 1 AGATE 90G4J7H2 Enclosure 1 GATE 9QG47254 Enclosure 1 Slot 7 SEAGATE 9QG1JHBS Enclosure 1 Slot 8 SEAGATE 9QG42ZD6 Name R1 array0 Stripe size BAK v Backoff percent 1 y Enable write back cache ere F 4Return to index Figure 24 Fibre Channel device creation page Fibre Channel Array Creation and Deletion Select a type of RAID from the drop down list shown in Figure 23 and press Create Fibre Channel Array of level gt In the case of RAID1 10 a screen like Figure 24 appears The name has already been typed in by the user and the default selection of drives remains Correct this and or make other selections if necessary Then hit Create After a few seconds pause a screen like Figure 23 should appear but with the new array represented by a new icon under Fibre Channel Arrays Selecting this icon yields Figure 25 Notice that the device file name is incomplete the reboot mentioned above is needed before further progress can be made over this array Also notice that the Current status is Initializing a process that can take hours for any RAID level except RAIDO Page 41 Figure 25 Fibre Channel Array Ri array0 detail page To delete an array reach its detail page as in Figure 25 and then hit the Delete RAID array button The delete will not succeed if any superstructure suc
23. A Server Storage Networking Authentication D Clustering Log Out ren jaintenance Manual CS ses Era EdgeWare Storage Configuration Ouota Management if S ij Edit Quota O Provides user storage usage quota management on the logical volumes EdgeWare 3 0 0 1ce HELP Quota for grape on Volumes LVstorage Blocks used 0 Files used 0 Soft block limit O Unlimited 10000 Soft file limit O Unlimited 10000 Hard block limit O Unlimited 112000 Hard file limit O Unlimited 12000 Available blocks on disk 262016000 total 262015828 free Available files on disk 262144000 total 262143994 free Update List All Quotas s 4Return to index Figure 40 Adding quota for a user If you want to copy quota limits for the user you just specified to other users ME Click on Edit User Quota for the user from the Quota page Click on List All Quotas Click on Copy Quotas Select one of the options on the copy quotas page Figure 41 Adding and copying quotas for groups Adding and copying quotas for groups work the same way as they do for users Page 58 S User Maintenance Manual E y e s r Ed g eWare Server Storage Networking Authentication Clustering Log Out EdgeWare Storage Configuration Quota Management z Pa 1 Copy Quotas O management on the logical volumes EdgeWare 3 0 0 1ce x Storage S E e Copy quotas for grape on ALL filesystems to o
24. A logical volume under a volume group can only be resized to increase from its original size Decreasing the size of a logical volume is not possible Also resizing a NAS volume requires its shares but not its Page 48 data to be deleted after being backed up from Backup System Configuration in Integrity gt Save and Restore and then restored after volume size has been increased In the case of iSCSI volumes the data not just the configuration must be backed up and restored a time consuming process If you have already used up all the space in your volume group as shown by two equal numbers in Storage allocated from volume group GB contrast Figure 31 which shows lots of room to expand then you must expand the volume group before performing the steps below This is done by adding a physical volume see above The resizing procedure is as follows for a NAS volume 1 Note share name s back up NAS shares configuration on the logical volume if necessary 2 Delete all the shares on the logical volume use Storage gt Share Manager This does not delete their data It will be found intact in step 14 In Volume Manager expand the volume tree for the volume group containing the logical volume Click on the logical volume link You will be presented with an Edit Logical Volume screen bottom of Figure 29 Click Unmount the logical volume Click on Return to Index Volume size is now editable Enter a new increased si
25. AND BREAKING A FAILOVER CLUSTER Once failover is started configuration changes are supposed to take place without having to bring the server s down Whether this ideal can be realized depends on circumstances EdgeWare therefore arms you for several possibilities that you can integrate into a flexible response strategy In the descriptions below it will be assumed that the initial state is as built in the previous subhead with leader as Primary and follower as Secondary DESIGN OF RUNNING HEARTBEAT ON A CLUSTER With all the setup configuration and HA resources prepared the Heartbeat program is triggered as a CLUSTER WIDE program That is the system on which Heartbeat is run is actually the entire cluster even though each Linux instance seems to be running its own code the two or more boxes are so tightly coupled that it should really be considered the loading and running of one program The design is the heartbeat signal is broadcast by the Primary on the private communication line of the cluster The Secondary like a watchdog detects any failure of this signal Should it detect this it seizes control of the shared resource floating network address and storage backplane stops called fencing or STONITH for Shoot The Other Node In The Head the former Primary and starts its own copies of the HARs A restoration of a cluster will involve a new system being placed next to a cluster with a failed member all of the ab
26. Add Security Group Figure 59 Fibre Channel Target configuration page Before any FC adapter can be used for exporting of disk targets it must be factory configured for FC Target Currently this step must be performed by Cutting Edge support staff If you have ordered the Edgeware box with FC Target support configured you will see a list of adapters in the first section Fibre Channel Adapters enabled for Target Mode If you have previously ordered an Edgeware box with a FC adapter installed but it is not visible under the Fibre Channel Adapters enabled for Target Mode section please contact Cutting Edge Technical Support An Edgeware OS upgrade might be required Page 80 ENABLING AND DISABLING FC ADAPTERS FOR TARGET MODE SUPPORT Once a FC adapter is factory enabled for exporting of disk targets it will no longer be able to be used for mounting external FC arrays initiator mode described in the RAID module You can still enable and disable specific FC adapters for FC Target export for administrative purposes All the adapters will appear even ones with no FC cable attached To enable an adapter for FC target export click the Enable button See Figure 62 for the result of doing this for adapter isp0 To disable an adapter from FC target export click on the Disable button DEFINING LOGICAL VOLUMES A FC target is a Logical Volume and is created in the Volume Manager just like a standard NAS or iSCSI vo
27. N Manual e EdgeWare Storage Configuration FC Target Mode i 2 Configuration A Provides configuration for Fibre Channel i Edit Security Group J Target Mode Support Storage SZ EdgeWare 3 0 0 1ce HELPI RAID f LUN Assignments Group Webservers qom Manager Virtual Disk filles LUN Options Remove LUN Assignment KI Share Manager Add LUN Assignment ds Quotas Adapter WWN Assignments Group Webservers Ke E S ee a Backup Target Mode User FC WWN Remove from Access List j Add Authorized WWN gt ses Return to index CSS re get CS retary LW Figure 68 Custom Security Group Edit Page Here you can click on Add LUN Assignment to assign LUNs to the customer security group just like what you can do to the Default security group as described in the section above Keep in mind that the first volume assigned to each security group must be LUN 0 Note that an exported FC volume can be assigned to multiple security groups each with a different LUN However the first initiator that mounts the device will own it and the other initiators will only be able to access the same exported volume after the first initiator releases it By default no initiator adapter will be able to see any LUNs from the custom security group To enable an initiator adapter you need to add the WWN to the authorized list To do that hit Add Authorized WWN You will be taken to the
28. Use asynchronous writes to disk Yes No Clients must be on secure port Yes No SP w d Remote lef Acc ode Read write O Read only Trust remote users Everyone Everyone except root Nobody Disable Posix ACLs Yes O No Don t trust UIDs None O Don t trust GIDs None O Default O Treat untrusted qroups Default O Treat untrusted users as tee j Browse gt as Browse Save Delete gt 4Return to index Figure 78 Edit Export NFS share name page NFS EXPORT CONFIGURATION Export Details Directory to export This is the directory being exported to the clients listed in the Export to field on the main NFS page Active Controls whether this NFS export is active or not If not it is commented out in the exports file Export to The hosts that will have access to the exported directory Five ways of specifying which hosts are allowed to mount the directory are available Everyone Any host capable of contacting your system is allowed access This option should be used with care especially if your system is connected to the Internet Hosts s This can either be a single IP address a single hostname or a regular expression like foo com In the last case any host whose reverse DNS lookup by IP address matches the expression will be allowed access WebNFS clients Allows clients using Sun s WebNFS protocol to access this directory MNetgroup Allows all membe
29. access may be from a local area network LAN or a wide area network WAN Please contact Technical Support for more details if necessary INTEGRITY The integrity of the files making up the OS is a fundamental part of the system stability and reliability Most known security exploits involve replacing modifying or corrupting some of the key OS files This Integrity module does verification by comparing two databases with one being a reference database assumed to be from an OS in a known state The other database is generated from the current state of the OS Comparing the two databases then provides a list of changes based on two factors 1 a checksum and 2 access or modification time The Integrity module is also the home of Configuration Backup and Restore which permits big hardware and software changes without requiring you to re enter user lists and other administrative data by hand VERIFY The recommended usage of integrity is to initialize a database for a known system and then store the database in a safe location The database can be saved to a share location and copied to a floppy or other storage media The database is stored in the var db integrit directory as a hidden file i e with a period prefix before the file name Page 127 The initial database is named var db integrit cenas_new cdb and can also be copied to another system by using the scp secure copy command from an SSH session scp var db integrit cenas_
30. address and service failover 9 EdgeWare 3 0 0 1ce L E NAS Servers HELP No cluster resources have been defined on your system JQ _ Remote Replication Create Modify Resource To define a failover resource Re e Authentication Keys To configure the type of authentication for heartbeat events Heartbeat status heartbeat is stopped No process Start Heartbeat Monitor Click this button to start the Heartbeat monitor process This will enable automatic service failover on your system with the current configuration Figure 129 IP Failover module main page after breaking cluster Page 161 REMOTE MIRRORING SYNCHRONOUS MIRRORING SYNC The Remote Mirroring module also known as Synchronous Mirroring allows you to mirror a pair of volumes between two EdgeWare servers synchronously It can replicate data at the block level between a local host and a remote host Data modified on the source volume of the pair is immediately replicated to the other volume in the pair You can think of it as a network RAID 1 The source or primary volume may have a NAS share or iSCSI target enabled on it The host selected as primary either the local host or the remote host will define this resource for the mirror and the mirroring software will sync up this resource share target to the chosen destination secondary volume The resource can take up a most the whole primary volume Remote Mirroring works with two EdgeWare systems It
31. be created 2 Create a UFS file system with command a where c6t 4d0s2 is the iSCSI drive slice b Confirm the operation in CA d 3060 05 of 48 tracks 28 sectors 32 330024 689056 5 455 S55 d File system is now ready to be mounted MOUNTING THE FILE SYSTEM 1 Create mount point with command a AA n 2 Mount the drive to mount point using command a W W 3 Confirm mounting with command AA TT Page 77 QUICK SPEED TEST 1 Create a 4GB dummy file a time dd if dev zero of mnt edgeware_iscsi_0003 testfile 1GB count 4096 bs 1024k Roughly 40 seconds to write 4096MB 102 4 MB s 2 Read a 1GB dummy file a time dd if mnt edgeware_iscsi_0003 testfile 4GB of dev null bs 1024k b Roughly 32 seconds to read 4096MB 128 MB s LONG DATA AND STABILITY TEST 1 Create 10 000 64KB files 10 000 times with command for e in 1 10000 do for i in 1 100007 do time dd if dev zero of mnt edgeware_iscsi_0003 testfile i count 1 bs 64k done done 2 After files are created read 10 000 64KB files 10 000 times with command for e in 1 10000 do for i in 11 10090 do time dd if mnt edgeware_iscsi_0003 testfile i of dev null count 1 bs 64k done done Page 78 FIBRE CHANNEL TARGET MODULE FC Target Fibre Channel Target Module allows Edgeware to export logical disk volumes as single or multiple disks Clients also called Hosts or Initiators can either mount th
32. can be made to the configuration of a failover cluster without breaking it or swapping Primary roles These are essentially those found in the Create Modify Resource page except for the top two lines plus the contents of Advanced Configuration See Figure 120 1 Go to Clustering gt IP Failover gt Create Modify Resource on the Primary s UI The Resource Status should be Up green 2 Make changes to Advanced Configuration if desired and hit Save 3 Make changes in the Create Modify Resource page except for the top two lines See Step IV of the previous subhead 4 Select Save This returns you to the module main page 5 Select Apply Changes BREAKING A FAILOVER CLUSTER Note This is an emergency operation This procedure involves races and therefore Cutting Edge cannot guarantee the results of performing it Nevertheless we include it here because of its potential utility to customers In particular it is very useful during the setup phase of an installation when failover has just been enabled and significant user data has not accumulated if it is desired to back out and try again with a somewhat changed resource configuration The result of breaking a failover cluster is to place both Storage Servers back in the state they were in at the end of step II of the previous subhead IP Failover Heartbeat Monitoring Thus their HARs are set up on the Primary ready to be activated in a non failover setting but none of th
33. created 10 When you first create the RAID device of level 5 or 6 the software has to generate the parity and sync the drives Figures 16 and 17 are from a different creation run on a system that does NOT have mirrored system disks In Figure 17 you see a newly created RAID level 5 device dev md0 is being synced and is currently at 0 1 The RAID device is ready to be included in a volume group and used as a logical volume in the Volume Manager right away but performance will be slower than normal and there will be no redundancy or data protection until the syncing has completed This could take several hours even days depending on the size of the drives and the I O on the system RAID device options Device file dev mdo RAID level Redundant RAIDS Status Active Usable size 732594816 blocks Percent of resyne done 0 1 Persistent superblock Yes Parity algorithm Default Chank size 64kB Current status clean resyncing Partitions in RAID SCS device D partition 1 SCS device E partition 1 SCS device F partition 1 SCSI device G partition 1 Figure 17 Software RAID level 5 device is syncing The Percent of resync done value is for when the UI page was created and does not update automatically Use Return to index and cycle back to selecting the array s device name for a more current value Note that response may be slow see Screen Response subhead above Page 33 11 Figure 18 is from yet another cr
34. defines a primary normally local host that is the source of data to be synced The secondary normally remote host has the logical volume that will be synced up with the volume on the primary host The secondary volume must be at least as large as the primary volume and is best made exactly equal in size Once configured for remote mirroring the logical volume on the secondary host will be owned by the mirror you cannot use it for any other purpose After configuring a Sync pair the mirroring software drbd syncs up that is it makes the data blocks on the secondary consistent to the primary even if user data is not present on the logical volumes Figure 139 shows an initial sync up This may take several hours or even days and a considerable network bandwidth will be used to transmit the entire primary volume THE PECULIAR MEANINGS OF DELETE In remote mirroring there are two instances where the term delefe does NOT mean what it appears to mean Be aware of these because both of them are important and not destructive at all but key to the functionality of mirroring e Delete a mirror This means release the mirroring connection between two volumes on separate servers It does not mean delete the volumes After deleting a mirror both volumes are still there and capable of taking up separate existences e Delete a share This means release the connection between the share s data and its NAS authentication protocol s It do
35. described in Set Client IP Address and Netmask on Microsoft Vista above Otherwise continue through the options below There are five ways to find an EdgeWare Server once the EdgeWare port eth0 is connected through the network switch if a DHCP server is on the same switched network They mostly work without DHCP too e The EdgeWare Discovery Tool with a keyboard and monitor e Magellan client on any Java capable operating system OS e Use of the serial number for a name search e Bonjour OS X discovery via zero conf via an OS X client e UPnP XP and Vista discovery via a Windows client Page 14 THE EDGEWARE DISCOVERY TOOL The EdgeWare Discovery Tool is already integrated with the EdgeWare Storage Server All that is needed is a standard monitor and keyboard plugged into the newly installed system These can be removed after discovery is complete Step 1 e Once you have powered on the Storage Server and have a monitor plugged into the console serial port you will be prompted by Edgeware to setup the ethO port Step 2 e View the current active network configuration by entering 3 and pressing the Enter key Figure 1 EdgeWare Discovery Tool after Selection 3 is chosen Edgeware displays the current IP address Broadcast address and Subnet Mask of the CD Optical unit Figure 1 Page 15 Step 3 e Deactivate the DHCP and configure a static IP by selecting 2 e You will be prompted to input the
36. desired IP address Subnet Mask and the Default Gateway To re activate the DHCP simply select 1 from the list of options and the CD Optical unit will automatically drop all the static IP information and then a network restart will give it an address from the DHCP server USE OF THE SERIAL NUMBER FOR A NAME SEARCH The serial number can be used to get a name for the Storage Server 1 Note the serial number on the back of the unit The format is CExxxxxx x 2 Open an internet browser on a workstation on the same subnet as the EdgeWare system 3 Type in the URL https CExxxxxx 10000 where CExxxxx is the serial number with hyphen and last digit omitted or URL https IP address 10000 with the IP address assigned from DHCP You can find the IP address from your DHCP server leases by matching the MAC address of the unit The digits from the 3rd digit till the 9 digit in the unit serial number specify the last 6 digits of the MAC address of your unit 4 Login to the EdgeWare web based UI with default username admin and password setup BONJOUR OS X DISCOVERY VIA ZERO CONFIG VIA AN OS X CLIENT It enables the discovery of the servers on a Macintosh client Its requirements are stated below e Allocate IP Address without a DHCP Server e Locate services without the use of directory server e Translate between names and addresses without the DHCP Server UPNP VIA A WINDOWS XP OR SIMILAR CLIENT e UPnP allows disc
37. documentation tools test equipment and other material to provide the Services regarding the Covered Products Service Materials Manufacturer grants no title or license to the Service Materials and retains all right title and interest in the Service Materials Nothing in this Agreement implies any grant or license by Manufacturer to Customer in the Service Materials Customer will not to use the Service Materials or make them available to third parties without Manufacturer s prior written consent Manufacturer may remove Service Materials at any time and Customer will give Manufacturer access to permit this removal If Customer has requirement of on site service outside of normal business hours this service is available for a fee Manufacturer s Terms and Conditions Applicable to Software and Software Maintenance Cutting Edge 1825 Gillespie Way Suite 100 El Cajon CA 92020 800 257 1666 www cuttedge com r CUTTING EDGE MT A aa RK A a E 1 Wa a R rranty LIMITED WARRANTY Manufacturer Manufacturer warrants that for a period of 90 days from the date of delivery of the Product to Licensee i the media on which the Product is furnished will be free of defects in materials and workmanship under normal use and ii the Product substantially conforms to its published specifications Except for the foregoing the Product is provided AS IS In no event does Manufacturer warrant that the Product is error free
38. esse see see see see see ee ee 13 FINDING THE SERVER ON A DHCP BASED Nermwopk ese see see sesse ee ene see ge ese ene Gee se ee ee eke Ge ge ee ee GR Ge ee ee Re Ge ee ee Re ee 14 TROUBLESHOOTING STORAGE SERVER DISCOVERY esse sesse ese see se ese ene eke se ee ee eke Ge ge ee ee GR Ge ee ee Re Ge ee ee Re GR Ge ee Re ee 18 Overview of the EdgeWare Storage Systemsin smesso SE Ed Ee ER AE NE BE DoE AE Ke 19 QUICK START WEB MANAGEMENT tess tts ttes ttes trest test tes EtESEEESEEPSEEPSEASPEENPEENEEENEEESE EENE EE SEENE See 19 ACCESSING THE GUI E 19 NAVIGATING THROUGH THE USER GUDEN 19 KE Le 21 Stem NME lere SE contents EE EE EE EE EES sonteusnacaeaecBies 21 Management Mod le sees EERS Re DE We ERG Eege ge E Se 22 ass ee NE EE ER ER N ees 22 Wet GEE OR ET EER EE DEER EER EE TEE EES 22 SSL ENCRYPTION SEE OE EER EE EO OR OE EER EE EE OE 23 dye MR OE OE 23 leie Tee 23 GUT ELE LERE RE AR OE N 23 GUI OE Eier el RR EE EE EE EE ER 24 VIEW LOGIN SESSIONS EE EE ER EE SETENE 26 Storage sede ER EE 27 Ke OR el TEE 27 E RETTEN 28 SOFTWARE RAID EE EE EE EE EE EE EEN neds 28 Ed GR EE EE OR OE ER OE OR ced OE EE 34 FIBRE CHANNEL ARRAYS XYRATEX RAID sscescesesseseeseesesesseesesesecseesecesecsesseaesecsecaessesesaessesaeeeesessesaeeaereeeeas 39 Volume Manager Lee 43 USING THE VOLUME MANAGER FOR VIRTUALIZATION ees ese ese ese see ee ee ee ee ee ee ee ee ee ee ee 43 CREATING A NEW VOLUME GROUP 44 ADDING A PHYSICAL VOLUMES EE 45 ADDI
39. make the target more secure by enabling the CHAP authentication protocol on this target provide a username and password that the initiator has to supply to successfully connect to the target This properties page Figure 49 also shows the current connections from iSCSI initiators to this target Multiple connections are legitimate if MPIO MCS is in use with multiple network cables EdgeWare Server Storage Networking Authentica te Log Out Land tion Maintenance Clusterin g Manual EdgeWare Storage Configuration iSCSI Configuration 3 c Edit Local iSCSI Target k A8 W EdgeWare 3 0 0 1ce Provides configuration for iSCSI target and initiator D DO RAID HELP oo Manager La Target Name KI EE Alias 5 LUN 0 Quotas ed Local Device U een CHAP username ER CHAP Password Advanced Configuration Active Sessions to iSCSI Target Session ID SID Initiator Name ON External Command Sent Save Delete Return to index Figure 49 iSCSI target properties page HOW TO CONNECT WITH A MICROSOFT ISCSI INITIATOR EdgeWare is compatible with Microsoft Windows iSCSI Initiator Establishing the connection involves both storage and network configuration on both sides The steps are as follows Page 67 A First you need to activate the iSCSI target that you wish to use B Second you need to establish the network connection between target and initiator systems C Finally Microsoft Initiator
40. may vary With Windows NT you can send multiple pairs username password you have to fill in the Connect as field on any network connection dialog box and then you will be able to send a new pair of username password Security PDC member Domain In this security mode Samba server tells the connecting client that it is running in user level and passes all the username password pairs to an existing password server which must be a Domain Controller You should choose this Security level when you want to add the EdgeWare appliance as a member server to your Windows NT legacy domain Security ADS member Domain Using this security level you can join the EdgeWare appliance to an Active Directory domain This is possible if the domain is run in native mode or mixed mode Active Directory in native mode perfectly allows NT4 style Domain Members Workgroup Domain The domain or workgroup of your Windows network depending on the security level you choose Page 95 WINS mode Select the WINS mode for your network Act as a WINS server If you do not have an existing WINS server on your network and need to set up the EdgeWare appliance as WINS server select this Use server Enter your WINS server IP address here Neither If you do not need to configure WINS on the appliance PDC AND ADS OPTION Windows 2003 server PDC ADS If PDC or ADS is running the Windows 2003 operating system select yes PDC ADS server Enter the IP add
41. modules Additional documentation on Remote Mirroring More information on IP Failover Remote Replication Inclusion of more methods for NAS Discovery Migration of FTP from proftp to Vsftpd Included additional documentation and screenshots on Remote Replication and Remote Mirroring LDAP authentication explained Setting up IP Failover from scratch Page 3 8 a 7 id September 2008 November 2008 April 2009 July 2009 NAS Discovery constructive description added Constructive deployments for complex features Added server client HOWTO subheads Virtualization snapshot and backup elaborated LDAP server guidance added Handshake guidance added for clustering All clustering modules elaborated NAS iSCSI clarifications Added big partitions for software RAID Added Fibre Channel Xyratex RAID Resizing guidance Failover clarifications Feature revision Revision of Updates capability New Discovery Tool Fibre Channel Target FTP Solaris iSCSI Initiator interface Updates capability turned on Page 4 TABLE OF CONTENTS REVISION HISTORY EE 3 able OF EE 5 Connecting to EdgeWare Servers for the First TIMES RS NE ERG NEE RE RSG aai 10 A ESTABLISHING A GOOD HARDWARE NETWORK CONNECTION uses ese ese ese see se ee ese eke see ge ee eke Gee Ge ee ee Re Ge ee ee eke ee 10 B GIVING THE EDGEWARE SERVER AN IP ADDRESS AA 12 C FINDING THE EDGEWARE SERVER FROM YOUR CLIENT STORAGE SERVER DISCOVERY e esse
42. of Cutting Edge 1825 Gillespie Way Suite 100 El Cajon CA 92020 800 257 1666 www cuttedge com r CUTTING EDGE ot se b ae amp A G E R ENN Manufacturer will have the right to terminate or suspend in its discretion Customer s access to the Web Site in the event of any security breach All software that is available for download on the Web Site will be licensed to Customer in accordance with any license terms stated on the Web Site for the software and the terms of the parties license agreement for the Covered Products Customer acknowledges that Manufacturer will be entitled to track usage and other statistics on the Web Site which information may be used by Manufacturer to improve its services or otherwise for its internal business without restriction RESPONSE TIMES AND ERROR CORRECTIONS Customer may report to Manufacturer any failure of the Covered Products to substantially conform to their published specifications Error Manufacturer will respond to each call within the Response Time specified for the applicable Support Package to gather initial or additional information about the Error Manufacturer will use commercially reasonable efforts to correct or minimize the adverse effect of any reproducible and demonstrable Error reported to Manufacturer by Customer in a manner commensurate with the severity of the Error If Customer selects the Gold Support Package these corrections may include providing Customer
43. on the initiator system needs to be configured to see the target A ACTIVATE THE EDGEWARE ISCSI TARGET The first step is done in the EdgeWare UI by clicking on the Storage icon then selecting the iSCSI icon and if necessary clicking on the Start icon see Figures 50 and 51 EdgeWare Ki SN Va lt A Stor Figure 50 Start an iSCSI target Fa BOO 8 KRY User Networking Authentication Res Figure 51 An iSCSI target is running Page 68 B ESTABLISH THE NETWORK CONNECTION BETWEEN TARGET AND INITIATOR The second step is only sketched here as it is in the domain of the Networking gt Network module below in this manual The following design principles guide creation of a proper network connection for iSCSI e A network connection to be used by iSCSI should be static not DHCP It is not desirable for the IP address to change across reboots e If more than one physical network connection is to be used as with MPIO then they should be on different subnets e Unless iSCSI data traffic is expected to be quite light it is best for the initiator target connection to be dedicated not shared with many other LAN ports that are unrelated to the iSCSI setup This is typically accomplished either with point to point cabling or with a dedicated switch e Both initiator and target systems are likely to have other tasks than iSCSI The initiator is typically a server connected to a
44. other reason the EdgeWare Server does not receive an IP address from a DHCP server it will use Auto configuration to assign itself an IP address in the 169 254 0 0 network A typical address of this sort is 169 254 5 58 You cannot observe or control either DHCP or Auto The steps to be followed from your client perspective are 1 Make sure your client has an IP address and netmask in the correct range If you have DHCP this normally takes care of itself If not the Auto address range is 169 254 x y where in place of x and y you choose numbers between 1 and 254 not the same as your EdgeWare Server s address and the netmask is 255 255 0 0 2 Wait long enough after power up for the EdgeWare Server to boot up and receive its IP address Normally ten minutes is long enough If you have DHCP your client and your EdgeWare Server should automatically be given different addresses in the same range as determined by the same netmask If you have reason to suspect that this is not the case speak to your network administrator The client IP address techniques described below are mainly for the non DHCP case Please note that even if you have a LAN it is possible you do not have DHCP Ask your network administrator SET CLIENT IP ADDRESS AND NETMASK ON MICROSOFT WINDOWS XP Select Control Panel in the Start menu Select Network and Internet Connections Under Pick a Control Panel icon select Network Connections A graphic he
45. populate with a list of available servers on the network compare Figure 100 The Register a New Server approach is used where the broadcast would find so many servers that it would leave your screen too cluttered The Scan for Servers approach is able to look at one subnet out of several Notice in Figure 100 that the second to left icon for CE7E478A looks different from the others That is because that server has been identified as an EdgeWare Storage Server which required hand work as follows done before Figure 100 was captured Zf CE7E478A has not been registered in etc hosts be sure and make a detour to Networking gt Network gt Advanced Configuration gt Host Addresses and enter it with its IP address before proceeding The chicken and egg problem of how to get the IP address the first time you broadcast and get the icon with the spiderweb is solved by selecting the hostname 10000 link which if it is an EdgeWare Storage Server will offer you a login screen with the IP address Note this down and hit the browser s Back button then proceed to Host Addresses Having registered in Host Addresses return to Clustering gt NAS Servers Index find the desired neighbor server hit edit and a screen like that of Figure 101 will come up If this is a new discovery and still has its backwards N with a spiderweb as in Figure 100 you will have to adjust it so that it corresponds to Figure 101 This normally involv
46. schema file samba3 schema is added to the schema list when the OpenLDAP server is being configured Note Samba v 3 is looking for object class sambaSamAccount rather than the old SambaAccount which is also reflected in the new samba3 schema file Page 116 LDAP USERS AND GROUPS This page Figure 87 is used for displaying LDAP server database users and groups that the EdgeWare LDAP client can see If LDAP users and groups are visible the EdgeWare LDAP client will successfully connect to the server and retrieve the user and group list from the LDAP server Each of the users and groups listed should work to both authenticate and honor permissions from clients to the EdgeWare storage system Username User ID Real name Home directory Shell Samba IMAP Idickson Idap 1000 Larry Dickson home ldickson Idap bin bash No No No LDAP groups were found under dc site Return to index Figure 87 LDAP users and groups MANAGE CERTIFICATES This form provides options to manage certificates There are 3 different options given to manage certificates e CA Certificates e Certificate files e Certificate keys A suggested procedure is 1 Obtain a certificate using the server name for the Common Name CN This should correspond to the server name your LDAP clients will be using 2 Upon completion you will have a private key file host key and public key certificate file host cert Rename the files as needed 3 Upload the key file
47. steps A Establish a good hardware network connection B Give the EdgeWare Server an IP address C Find the EdgeWare Server from your client Storage Server discovery The critical step here is step B because it is completely invisible to you the customer You can only judge its success by its effects which is why step C is necessary A ESTABLISHING A GOOD HARDWARE NETWORK CONNECTION You need a client system in other words a standard workstation or laptop with a network capable operating system such as Microsoft Windows MAC OS or Linux Unix and a working network connection network connecting hardware either a point to point network cable that will connect your client directly to the EdgeWare Server or else two network cables one for the client and one for the EdgeWare Server each connecting to a hub switch or more complex Local Area Network and your EdgeWare Server In the point to point case it is not necessary to use a special crossover cable EdgeWare network hardware detects and corrects for a standard cable running point to point The issue of hot swapping changing connections on hardware while it is running arises during this setup task It is OK to hot unplug or hot plug a network cable as long as the connection is currently idle idle means data is not flowing across this connection In most modern desktop systems it is OK to hot unplug or hot plug the keyboard purple connector or the video monitor DO
48. that it will operate with any software or hardware other than that provided by Manufacturer or specified in the documentation or that the Product will satisfy Licensee s own specific requirements REMEDY Licensee s exclusive remedy and the entire liability of Manufacturer under this limited warranty and any other guarantee made by Manufacturer is at Manufacturer s option to repair or replace any Product that fails during the warranty period at no cost to Licensee Manufacturer will replace defective media or documentation or at its option undertake reasonable efforts to modify the Product to correct any substantial non conformance with the specifications RESTRICTIONS The foregoing limited warranties extend only to the original Licensee and not to any subsequent purchasers or third parties and do not apply if the Product a has been altered or serviced except by Manufacturer or an Manufacturer authorized service provider b has not been installed operated repaired or maintained in accordance with Manufacturer s instructions c has been subjected to abnormal physical or electrical stress misuse negligence or accident d is licensed for beta evaluation testing or demonstration purposes for which Manufacturer does not receive a payment of full purchase price or license fee e has been damaged or rendered defective by the use of parts not manufactured or sold by Manufacturer or f has been operated outside the usage parame
49. the GUI log entries for scheduled snapshots 2 Next enter the schedule for the snapshot creation using the pull down list at Job schedule 3 Select Create if you want to just make the schedule or select Create and Snapshot Now if you want to trigger a snapshot right now AND make the schedule Whether making the schedule actually causes regular snapshots to start happening depends on what you selected for Enable Snapshot Scheduling above EDITING AND MOUNTING SNAPSHOTS AND EDITING SNAPSHOT SCHEDULES R jSnapshot Volumes add a snapshot volume f GlLYstoragesnap 0820 1219282202 jLvstoragesnap 0820 ae Figure 35 Snapshot slice from Volume Manager page LYstoragesnap 0820 schedule The slice above from a Volume Manager page shows how two snapshots and a snapshot schedule appear in the Volume Group s tree This was generated by selecting Create and Snapshot Now in step 3 above Names ending in schedule are not Snapshot Volumes but snapshot schedules Among the two actual Snapshot Volume names the short one was the immediate snapshot made by Snapshot Now and the long one with the timestamp number 1219282202 terminating it is one created by the scheduler The timestamp number tells how many seconds after the Epoch 00 00 Jan 1 1970 the snapshot was made Selecting a snapshot schedule link brings up a snapshot schedule editing page looking much like Figure 34 except Delete i
50. the drives RAID 6 provides for an extremely high data fault tolerance and can sustain 2 simultaneous drive failures Minimum drives required for a RAID6 is 4 or N 2 where N is the number of desired usable drives Software RAID is widely used in Linux storage servers and is very reliable The software RAID is available for providing additional reliability when adding external SCSI drives or a JBOD disk set After selecting the software RAID icon in the RAID module the administrator can view all existing software RAID sets which are also listed inside the proc mdstat file where the Linux kernel tracks RAID sets Figure 15 below give an example of software RAID sets created from partitions on drives connected to standard IDE and SATA controllers SATA drives show up as SCSI drives in Linux systems Creating a Software RAID Device To create a software RAID device perform the following steps Page 29 1 Navigate to the RAID page under the Storage section 2 Click on Modify Partitions button 3 On the Partition Manager page click on Add primary partition link for the drive you want to include in a software RAID set This should be a drive that has not been used in any other RAID or anything else Verify by going to the other options on the main RAID page You can also check Volume Manager which drives are already configured in volume groups and hence not usable for software RAID until you delete those volume group structures For exa
51. the reference database Last Save Sep 24 13 50 Save Options x Create MDS sum D System Configuration Backup Backup gt Restore HM Figure 97 Part of Archive page Archive the Reference Database Here you can specify where to copy the reference OS integrity database to an easily accessible location such as a share defined on the system so that you can copy the database to another network location If desired specify this target share with the help of Browse it defaults to shares cenas_new cdb but Browse lets you point it to a share of your choice Decide whether or not you want MD5 it defaults to shares cenas_new_cdb md5sum Then hit Save in Save Options to store the result which can be around a megabyte or more in size It can then be recaptured later via Path to Archived Database above in Verify System Configuration Backup This option allows you to back up the configuration of the Storage Server such as user and group lists network settings share settings and much more and at a later time restore the Storage Server to those settings The backup feature should a ways be used from time to time as it can save the administrator long error prone hand labor Especially urgent is to back up configuration before doing major hardware or software upgrades NOTE To enable system configuration backup you have to create an OS integrity database file from the Verify page As long as there is an integrity
52. to arrive then check your mailbox If it fails to arrive go to Monitoring gt Logs gt System Logs gt var log maillog and use what you find there as a guide to revising your parameters and then go back to step 1 If it arrives your Email Notification is successfully set up for that target email address Logs are essential for troubleshooting and monitoring the EdgeWare storage system This module allows you access to them without need of console or SSH access or a customer support call Among the uses of this capability are If you desire a peek under the hood at details of an operation check logs before and after you do the operation If the system seems to be misbehaving the logs can help you solve the problem on your own or the logs can be cut and pasted into your communications with Technical Support Timestamps on the log entries help detect livelock loops or busy behavior that may be slowing down or overwhelming your system When unauthorized entry is being attempted log entries show the tracks of it Two components are available in the Logs module page EdgeWare Actions and System Logs Page 125 EDGEWARE ACTIONS The EdgeWare Actions page gives you the ability to search logged actions that have been executed by any EdgeWare web UI administrator user Enter your search options and click on Search to get a listing of the logged EdgeWare actions fitting your criteria The log file searched is var webmin webmi
53. to be a general database not just user and group authentication You must know your LDAP server s IP address as there is no broadcast option The LDAP module gives you the configuration and management options for integrating the LDAP server database into your EdgeWare Storage system as a user and group authentication tool Local users and groups can still be specified but the master authentication repository will be the LDAP server database when the LDAP client is configured as specified below Currently a standard set of schemas represents the authentication objects the LDAP client recognizes For EdgeWare the schemas are available if a conversion to LDAP authentication is planned to insure compatibility LDAP Server Each operating system variant has its own way of supporting LDAP and the guidance here can of course cover very little Nevertheless it is offered because the necessary information will be similar on all LDAP servers The settings described here are for SUSE Linux Enterprise Server 10 and were entered during install from SuSE CDs The server settings correspond to the EdgeWare client settings shown in Figure 86 Enable LDAP Server was set a change from default and defaults were left BaseDN dc site RootDN cn Administrator dc site Password unchanged from root password Encryption SSHA All the following checkboxes were set Append BaseDN to RootDN Activate Stat TLS Register at an SLP Daemon After open DAD was
54. volume It is important to understand however that there is no form of data backup when using a RAID 0 If a single drive fails you will lose the entire array This type of RAID should not be used for mission critical data RAID 0 is available for any system with two or more drives RAID 1 Mirroring This type of RAID is used to create a mirror copy of the information contained on one or more of the disks In this method a mirror is created for each drive containing data For example in a two drive RAID 1 setup both drives contain the exact same data Since there is 100 redundancy there is no risk of losing data if one drive fails RAID 1 is available for any system with two or more drives RAID 5 Disk Striping with Distributed Parity This type of RAID features both striping and redundancy This type of RAID uses a technique called distributed parity which allows data to be recovered if one drive in the RAID fails In addition data blocks are interleaved evenly across the drives in parity bits RAID 5 is only available for systems that have three or more drives RAID 6 Disk Striping with 2 Independent Distributed Parities This type of RAID is an variation of RAID 5 which allows for additional fault tolerance by using a second independent distributed parity scheme two dimensional parity Data is striped on a block level across a set of drives just like in RAID 5 and a second set of parity is also calculated and written across all
55. 0 800 257 1666 www cuttedge com r CUTTING EDGE te es hee G A G E R Pets system log files or any other reasonable information requested by Manufacturer to diagnose reproducible Errors In the event of software notification errors customer is responsible for contacting the Manufacturer describing the issue at hand and inputting UNIX shell commands in the Maintenance module of the Graphical User Interface GUI Customer is responsible for submitting via e mail UNIX shell command results to Manufacturer technical support for analysis COOPERATION Customer will provide Manufacturer with reasonable cooperation and assistance and with information as is reasonably requested by Manufacturer in connection with Manufacturer s performance of the Services Any information provided by Customer to Manufacturer may be freely used by Manufacturer for future versions of the Covered Products or otherwise without restriction Customer will provide Manufacturer with physical location of products upon installation DESIGNATED CONTACT Customer will designate a primary contact and an alternate contact in connection with the Services These contact persons will promptly notify Manufacturer of Errors provide the information set forth in the paragraph above entitled NOTIFICATION OF ERRORS and perform additional duties including but not limited to system restarts logging and running of operational readiness tasks including replacement of
56. 0 j A amp A on Controller 0 AA Target 0 dReturnto index Figure 130 Storage gt RAID gt Modify Partitions page If there is any discrepancy click on the partition link the number under No in Figure 130 SCSI device A and find its identifier e g dev sda1 Then go to Maintenance gt Command Line and execute fdisk 1 dev sdal replacing dev sda1 with the partition identifier you found Be sure and include the 1 which is dash ell You will get an exact byte size of your partition Do this on both primary and secondary and check they are the same or at worst the secondary is larger EdgeWare Maintenance H Configuration Command Line Encrypted command line access to the Command Line system ADVANCED USERS ONLY EdgeWare 3 0 0 1ce K z HELP gt fdisk I dev sda1 Disk dev sda1 doesnt contain a valid partition table Disk dev sda1 163 9 GB 163921572864 bytes 255 heads 63 sectors track 19928 cylinders Units cylinders of 16065 512 8225280 bytes Execute command fdisk dev sda1 vi Eatprewgus Clearcommands Figure 131 Maintenance gt Command Line get exact partition size Page 164 3 Create logical volumes that can be made raw 4 e If you are starting fresh create on both primary and secondary through Storage gt Volume Manager until you have a logical volume of identical size on both Even if it is intended for the mirror t
57. 8 1 130 Server Time amp Date Thu Sep 18 08 51 04 PDT 2008 Caruar Lntina NOEINA un 19 min 1 310 40ce on Leader22 Edgeware Linux 3 1 Ki S Internet Ta i TUE ss a Address https 192 168 1 132 10000 sl Ba Links gt Google G voog EG gt Y Bookmarks f 1 blocked YP check Autoli H AutoFill a gt Sendtow 4 O Settings e rage Networking Authentication Maintenance Clustering Pd Server Name Follow22 Server IP Address 192 168 1 132 Server Time amp Date Fri Sep 19 09 53 18 PDT 2008 Server Intime M 53 10 un 20 min A internet Figure 109 Browser instance for each server s UI C IMPOSE STATIC IP ADDRESSES NETMASKS AND ROUTES This is carried out in the Networking gt Network gt ethx links of each server See the Page 143 documentation of Networking gt Network above in this manual Because of the use of the Host Addresses list etc hosts it is not OK for IP addresses to change underneath your ports at any future time Therefore you must impose static IP addresses on each port of each server Check with your network administrator to find which static IP addresses are available on your LAN for the eth0 ports Traditionally the eth1 IP addresses are of the form 10 x x x since it is a private network Impose static IP addresses on both ethO and ethi on both servers Make sure the netmask is correct in each case the point to point subnet should not intersect
58. 94 everyone 95 avahi autoipd 101 messagebus 81 avahi 102 hacluster 501 Other group Browse La Create Ei 4 Return to index Figure 83 Local Users and Groups gt Create New User page Enter a username for the user Specify a real name for the user Give the user a password and confirm the password in the next field Password changed shows the last date on which this user s password was changed either through web UI or the passwd command Minimum days is the minimum number of days required between password changes Warning days is the number of days before the user is forced to change his password Expiry date is the date on which this account will expire After this date the user will no longer be able to login to the system Leaving blank will disable password expiration Maximum days is the maximum number of days allowed between password changes Once the time is elapsed the user will be prompted to enter a new password at login Inactive days is the number of inactivity days over the maximum days before the user s account is disabled Next you can provide the local group to which this user will belong Click on create to add the user in the system CREATE A NEW LOCAL GROUP da et a Click on the Create a new group link On the next page specify a group name Specify a group ID or accept the auto generated ID Add users to the group in the Members text box
59. AS Servers Index See the documentation for NAS Servers Index above in this manual Notice in this case that since the source is known it may be convenient to do the Host Addresses first and then if it is not already there capture the source in NAS Servers Index through use of Register a New Server E TARGET CONFIGURE AND SCHEDULE THE REPLICATION Here is the actual documentation for this module Clustering gt Remote Replication This description is for creating a new replication revising an existing one is similar Page 139 1 Click on Clustering then Remote Replication The module main page appears Figure 103 Now select Configure Remote Mirror Ed ge Ware Networking Authentication Maintenance Clustering Log Out end Server Storage Manual i EdgeWare Clustering Configuration Replication i oes d 2 Ly See ER of the data O Setup servers for mirroring A EdgeWare 3 0 0 1ce 4 HELP ig G Currently Scheduled Mirrors Figure 103 Replication module main page 2 Select the source server from the list of available EdgeWare servers only one appeared in Figure 104 Make sure you have configured that server from the NAS Servers Index page as described above Link type may be left at Normal link to server if desired AH g fd Ge amp A Server Storage Networking Authentication Maintenance Clustering Log Out Manual EdgeWare Clustering Configuration Replic
60. Add Authorized WWN page Figure 69 To obtain the WWN for your FC adapter refer to the manual of the FC adapter on your host system Once the authorized WWNs and the LUN assignments are configured in the custom security group the authorized adapters should be able to mount the exported FC target volume Page 86 Manual ow me RAME A Ed g ew are Server Storage Networking Authentication Maintenance Clustering Log Out Deer EdgeWare Storage Configuration FC Target Mode Configuration Provides configuration for Fibre Channel L e Add Authorized WWN Target Mode Support Storage EdgeWare 3 0 0 1ce HELP S RAID oo Manager e mmm Add Authorized WWN eet Add Authorized WWN Security Group Webservers CR Saclay WWN to Authorize Co wes X A Save CS rer va Return to index Figure 69 Add Authorized WWN property page HOW TO CONNECT WITH A FC SWITCH AND HOST Unlike iSCSI no special initiator software is needed Your Fibre Channel Target system merely sits on a FC fabric with any FC host at the other end FC fabric switch adjustments perform many of the same functions as iSCSI initiator software If there are multiple FC hosts be sure each target is masked at the switch so that it appears at most at one of the hosts or set up custom security groups as described above to accomplish this purpose A FC target is not designed to support read or write contention from multiple hosts FORMATTING THE F
61. All of these settings use information that derives from the LDAP server Base Domain Name This is the root node of the LDAP database Use SSL Check if appropriate Use TLS Check if appropriate LDAP Server Domain name or IP address of the LDAP server Never use the default of 127 0 0 1 as it would imply the EdgeWare Storage is the LDAP server a Capability that does not yet exist Port The port number that the LDAP server uses If none is specified the port connection defaults to port 389 Anonymous bind Check this box when an anonymous bind is being used i e no username or password is required for the LDAP server network access Page 115 User Settings LDAP Bind Name The user name used to bind to the LDAP database This name is given relative to the Base Domain name and derives from the LDAP server Bind Password Password for the bind user It derives from the LDAP server Root DN Check this box if the LDAP administrator account is used to bind to the LDAP database In this case the information to access the LDAP server and database is kept within a separate file with limited access to restrict anybody other than a local process to read the information Corresponds to appending baseDN to rootDN on LDAP server Use LDAP authentication Check this box when LDAP authentication becomes active with the settings described above for all network authenticated protocols If this checkbox i
62. B shows the amount of storage that is in use and not available for this logical volume The default Volume size number takes all the available storage Jt is usually NOT a good idea to take all the available storage as that leaves no room for unanticipated future logical volumes or snapshot volumes Select the file sharing protocol s in Network File Protocol or combination of protocols you will be enabling for the shares on this logical volume Use CIFS in a Windows only network NFS in a UNIX variant only network Mixed for multi platform network or ISCSI to export the volume as an ISCSI target If your iSCSI initiator is Solaris pick Solaris iSCSI Target If you have more than one Physical Volume a identical in the volume group under which you are creating the logical volume and if you wish for the speed advantage of striping a RAIDO like technique select yes for Stripe across physical volumes Figure 29 shows selected no for Stripe across physical volumes as the Physical Volumes in VGstorage are dissimilar The default in this selection which is the alternative to striping is concatenation or nose to tail stringing together of physical devices Hit Create top of Figure 29 Exit the confirmation screen middle of Figure 29 with Return to Index The result will be Edit Logical Volume page with Logical Volume details bottom of Figure 29 Exiting the Edit Logical Volume page will show the Vol
63. C VIRTUAL DRIVE Again unlike iSCSI no special procedure is required to format the FC Virtual Drive After the FC Target software is running the FC Host may be rebooted if necessary It will then see the FC Target as a local SCSI drive Follow standard disk formatting procedure on your host system Page 87 NETWORKING The EdgeWare Storage system provides networking by using the TCP IP stack within the Linux kernel In addition to TCP UDP is often used for NFSv2 and other network applications discovery and announcement protocols e g CIFS DHCP GUI etc Among all modules the Network module in the Networking section is the most vital to system administration Caution Improper configuration within the Network module may leave the system inaccessible except by VGA keyboard or serial console IP network hardware directs communications to each network port according to a combination of the port s IP address and its subnet where the subnet is derived from a combination of the IP address and the netmask of that port For instance if your ethO port has the IP address 192 168 1 10 and the netmask 255 255 255 0 then its subnet consists of all properly connected IP addresses from 192 168 1 0 through 192 168 1 255 Any message from any of these with destination address 192 168 1 10 will be heard by your eth0 port and any message output through your eth0 port with another properly connected destination address in this range will be heard
64. Click on create to add the group in the system Page 110 Windows Domain Users and Groups When the NAS is joined to a Windows domain this page shows the users and groups authenticated by that Windows domain For guidance in joining a Windows domain see Windows Domain Join in Networking gt CIFS module above in this manual NIS Domain Users This lists the NIS domain users when the appliance is set up for NIS authentication See the NIS module below in this Authentication section LDAP Users and Groups This lists the LDAP domain users and groups when the appliance is set up for LDAP authentication It will be covered in detail in the LDAP module below SHARE PERMISSIONS This page allows you to set user and group permissions on the shares you have created on the local volumes The first page gives you a summary list of the shares that have been defined on the system When you click on a share name you are directed to a page as shown in Figure 84 The left text box gives you the list of valid users for that share and the right text box gives the list of valid groups Adding users to the share 1 Type a valid user on the system into the text box labeled Valid users or select users by clicking on the respective browse button it will give you a popup window listing all the local or domain users for the system 2 Repeat step 1 for all users you want to add to the share 3 Check the default permissions that you want t
65. Comme wW Backup Enabled Protocd 5 e ernel ia iscsi Save View User Connections gt Delete dReturn to index Figure 37 Choose volume pop up window Options available in Step 7 CIFS Microsoft Windows Common Internet File System enhanced version of Microsoft s SMB NFS Network file system developed by Sun Microsystems and native to Unix variants like Linux APPLETALK Apple s network file system including resource forks FTP File Transfer Protocol HTTP Hyper Text Transfer Protocol for administrative users ASYNC Asynchronous replication source The first three items in the above options list are not exclusive to the operating systems mentioned For instance CIFS can be accessed by Linux clients and NFS can be accessed by Windows clients Each of these protocols is friendly to certain styles of authentication and permissions that are most prevalent on its native clients For more details see the Network section below Once the share exists it will be listed in the table in the Share Manager module page and can be edited Use of the share requires further setup in the Network or other sections Page 56 QUOTAS MODULE A disk quota defines the maximum amount of hard disk space allowed for a user s files There are two quota limits the hard limit and the soft limit The hard limit defines the absolute maximum When the hard limit is reached the system will not allow a user any more space This ef
66. DRESS On the leader go to Networking gt Network gt Advanced Configuration gt Host Addresses and record this new IP address with an unused hostname It should be related to the four hostnames of Handshake e g in the example given you could use Shared22 to relate to the 22 in Leader22 Follow22 etc Do exactly the same thing on the follower an exception to the rule that you do failover settings on one server only Figure 122 1 GER ENE EACH Een TEEN DION Provides configuration options for the IP Host Addresses O o Ee Oo protocol like IP addresses and DHS client EdgeWare 3 0 0 1ce HELP Add a New Host Address IP Address Hostnames 127 0 0 1 localhost localdomain localhost 192 168 1 132 Follow22 10 168 1 133 Follow22 eth1 192 168 1 130 Leader22 10 168 1 131 Leader22 eth1 192 168 1 158 Shared22 Add a New Host Address Return to index Figure 122 Follower s Host Addresses after Shared22 save Figure 122shows state after the Host Addresses change in the follower UI This information does NOT automatically transmit from the leader VI START THE HEARTBEAT MONITOR AND GET RESOURCES Page 154 IMPORTANT Before starting Heartbeat make sure any hardware or software RAID building is finished In Storage gt RAID select each array and check it compare Figure 17 The nature of the heartbeat is for the Primary to be nested outside the Secondary Figure 123 The reason fo
67. Follower s module main page should now correctly identify Primary Node as leader just like leader s module main page Do not change anything on the follower Page 152 EdgeWare Clustering Configuration Heartbeat if Prinses contiqisticnortione Tanke Create Modify Resource O address and service failover EdgeWare 3 0 0 1ce E Heartbeat configuration options o Hosts in cluster O Follow22 Leader22 v f eth0 192 168 1 132 Ethernet heartbeat interfaces O None EE Shared floating IP address Failover Resources Shared Storage Block Devices IP address for ping checks Save rr Advanced Configuration Delete r 4Return to index Figure 119 Follower s Create Modify Resource screen passive IV CONFIGURE THE HEARTBEAT SERVICE Now re enter Create Modify Resources on the leader Note Figure 120that the first radio button has moved EdgeWare Clustering Configuration Heartbeat if SE Create Modify Resource id address and service failover EdgeWare 3 0 0 1ce L Heartbeat configuration options Hosts in cluster Leader22 O Follow22 Ethernet heartbeat interfaces O EE EI Shared floating IP address 192 168 1 158 Failover Resources Y Shared Storage _ Block Devices IP address for ping checks Save rr Advanced Configuration r Delete n 4Return to index Figure 120Leader s configuration parameters At this point Advanced Configu
68. Local Area Connection 2 or similar should say Enabled or Unidentified Network or a name of a network See XP section above for limited connectivity If the network cable is hot unplugged from the client this should say Network cable unplugged Plugging it back in should cause it for a moment to flash up a graphic saying Enabled and then to go back to its previous state TROUBLESHOOTING A BAD CONNECTION All cables should be firmly seated However if continued pressure is required to get a connection and taking your finger off the cable s connector causes the graphic to spring back to Network cable unplugged then the cable should be marked defective and replaced A cable that pulls out easily without pressing the spring lever on its end connector should also be considered defective as bumping it may cause the connection to be lost If no cable can be found that works reliably then the server client hub or switch network connection must be considered defective If the fault appears to be with the EdgeWare Server s network connection pleae contact Cutting Edge Technical Support Otherwise replace the defective part and continue checking Page 11 B GIVING THE EDGEWARE SERVER AN IP ADDRESS EdgeWare Servers are preset to acquire an IP address from a DHCP server If you have a LAN connection and if there is a DHCP server on this LAN the EdgeWare Server automatically configures itself If the connection is point to point or for any
69. MART monitor fc hardware monitor Sync monitor shuttle b52 hardware monitor v Create shuttle hardware monitor ew 2i hardware monitor multipath monitor iSCSImonitor Return to index Figure 90 Add service selection using monitor drop down ADDING OR EDITING A SERVICE FOR MONITORING To add a service such as SMART select Add service on the Watch Lists page and pick a monitor on the Using monitor drop down list Figure 90 Then fill in the rest of the data and hit Create Watched Service Details Name of service This is the name of the service or component that is being watched as you wish it to appear in Monitoring Status Description A brief description of the service or component Using monitor Selects the monitoring scripts that will be used to monitor this service Each script is named with a service name Monitor parameters Advanced parameters to the monitoring program Normally left blank but compare Free space above Check every This sets the time interval at which the monitoring process checks the service Enter a realistic value usually several minutes or hours as too frequent checks can fill the logs or slow the system Add a new period This can be selected on the page that results from selecting an existing service on the Watch Lists page Monitoring periods define how often to check the conditions which should allow alerts to be sent You can define multiple monitoring periods for
70. Management 232 88 GB 232 88 GB NTFS fa Services and Applications Online Healthy System Disk 1 E Basic 232 00 GB 232 00 GB Online Unallocated MoS cee es N Unallocated RB Primary partition Figure 56 Computer Management Windows XP 3 Right click on the unallocated area of the new iSCSI virtual drive and select New Partition 4 Follow the New Partition Wizard and configure the new iSCSI disk to your preference We recommend you check the Perform a quick format box otherwise formatting may take hours 5 Wait until your new disk is formatted This should be under a minute for a quick format or perhaps a few minutes for a mirrored volume 6 When the iSCSI disk is formatted you will get a Healthy status display see Figure 57 Page 72 E Computer Management Ad Sy fie Action View Window Help fe x e gt om e 1S Computer Management Local volume Layaut_ Type File System status Capacity Free Space Free Fault Tolerance Overhead i System Tools E C Partition Basic NTFS Healthy System 232 88GB 222 28GB 95 No 0 fE Event Viewer EIN Partition Basic NTFS Healthy 232 00GB 231 92GB 99 No 0 Shared Folders z Local Users and Groups Performance Logs and Alerts 3 Device Manager Removable Storage E Disk Defragmenter Disk Management T Services and Applications 29 Wei 2 232 68 GB NTFS Online Healthy System Dis
71. Model ATA WDC WO360ADFD 00 7 swap P EE Controller 2 4 Extended EE 400 4500 IHC H en mmm D VM 400 4500 Add logical partition Return to index Figure 11 Showing available SCSI devices A and B Page 30 4 After selecting Add primary partition a page like Figure 12 appears A partition using all of the space of the drive is shown You can change this In Mass Storage partitions the type is always EFI GPT but the size can be many terabytes When satisfied with your Extent hit Create i d EdgeWare Storage Configuration Partition Management iji IF Create Partition O The Partition Manager modifies the EdgeWare 3 0 0 1ce partition file system ALL DATA WILL BE LOST Partition Details Location SCSI device A partition 1 Device file dev sdal Type EFI GPT Extent 1 1887966 of 1887966 Status Not created yet Size Not created yet Create H Return to index Figure 12 Create Partition page 5 Hit Return to Index and create partitions for more drives in the same way depending on what RAID level you wish to create using them Except for concatenated linear RAID Partitions used in software RAID should always be exactly the same size and should always be on different disks 6 When done you hit Return to Index for the last time and find yourself on the Partition Manager page compare Figure 11 and all your drives have red horizontal bars under Extent Start and End should have numbers but Use and Fre
72. NEE 114 W TE 118 MOT DOW EE 118 MONITORING E mret ES EE EE EE EE DE N ee ee N DE E N ee de Di 118 LOGS RE EER EER RE EET EE ER EER TREE RE TE EE EO 125 le e EE EE EE EE EE EG 126 se DIE SEE OE RE OO OE ER OE 127 SUS ER EE EE ORE EE OE EE OE EE 127 de ME N RE EE EE 133 RE INSTAULING e e 134 lr le AA eA AAEE EEEREN AEEA EE ENEA EAA 135 NAS SERVERS TEE 135 REMOTE REPLICATION ASYNCHRONOUS MIRRORING ASYNC ccccccsscsscsscesceseesecseesscesecsscstcceesecseesecesecseeseeseeeass 138 HANDSHAKE E 142 IP FAILOVER HEARTBEAT Mowtropme sees see see see see ee ee ee ee ee ee ee Re ee ee ee Re ee ee Re ee ee ee ee ee 148 FAIL EVENTS SWAPPING EDITING AND BREAKING A FAILOVER CLUSTER esse esse sees sees see ees see ee ee ee ee ee ee ee 158 REMOTE MIRRORING SYNCHRONOUS MIRRORING SYNC cessssesceseeseseeseesesecseeseeeeeecsesaeecesecsesaeeseeeaesaeeseeesereaes 162 SETTING UP A REMOTE MIRRORING LE OE EE OE ON 163 EXPORTING THE MIRROR AS A SHARE ccceccccesssscccssscsecessecsccessessccesusecsscsecesssssccssusececessscsececesseceesessecessseessueas 169 EXPORTING THE MIRROR AS AN ISCSI TARGET sccsccescesceseeecececeeseesecsececeeseeseeseesecsecsececeesecaeecaeseeseseeaeeseeeeeaes 173 SWAPPING AND DELETING REMOTE MIRRORING NC 174 Cutting Edge Terms amp eed se OE EE EE 176 Page 9 CONNECTING TO EDGEWARE SERVERS FOR THE FIRST TIME The process of connecting to an EdgeWare Server for the first time requires completion of three
73. NG A LOGICAL e E 45 DELETING A LOGICAL VOLUME EE 47 RESIZING Re ue Tel 48 SNAPSHOT e 49 DESIGNING A SNAPSHOT od dend EE EE EE EE Ee ENE 50 CREATING A SNAPSHOT OF A LOGICAL VOLUME esse sesse see see ees see ee ee ed Re ee ee ee Re ee ee Re Re ee ee Re Re ee ee Re ee ee Ki EDITING AND MOUNTING SNAPSHOTS AND EDITING SNAPSHOT SCHEDULES esse ese ese ese ese ee ee ee ee ee ese ee ee ee ee 53 DELETING SNAPSHOTS RE onnies a ESEESE EAEEREN EERE EEE A E EEE EE a 54 Share Manager Oe E 55 CREATING AND EDITING A SHARE ccsccesscssscssscssecssecsecsscssecsescssecseecssecssecssscssscssecsescssecseecseecasecaeeseeeceesenseeeeenes 55 Quotas ee EI 57 MANAGING DISK QUOTAS EE EE EE 58 QUOTAS FOR NEW USERS AND GROUPS ees eene ee ee ee Re Re Re ee Re Re Re ee Re Re ee Re ee 59 Backup leed AE EE EE R a 59 DESIGNING A BACKUP Ger 59 ADDING A NEW BACKUP OF A VOLUME ees sees see ee ee ee AA AA AA aS ni e n AA AA A n ee a ee 60 SCHEDULING BACKUP RE OO OE EE RE dE EE Ee 62 ase IE EE RE AR N N OE OR OE EE EE EE EE EE EE 63 ESE EE EE OE IE LIE EE OR N OOK 64 ISCSI TARGETS EE EE EE EE EE 66 CREATING AN ISCSI TARGET nosco EE ER EE ER EE OR 66 How TO CONNECT WITH A MICROSOFT ISCSI INtrtarop 67 FORMATTING THE ISCSI VIRTUAL DRIVE IN MICROSOFT WINDOWS esse esse sesse see ee ee ee Re ee ee Re Re ee AR Re 72 Connecting a Solaris iSCSI initiator in static mode to an Edgeware iSCSI target esse 74 PARTITION AND LABEL THE SCSI DRIVE EEN 75 CREATE A UFS FILE
74. NOT hot swap the mouse green connector as this will freeze up most client systems and require a reboot or power cyde 1 Make sure the client system is on fully booted up and ready for connection In the point to point case this means the Ethernet cable intended for the EdgeWare Server is connected to the client and reaches far enough to attach to the EdgeWare Server In the hub switch LAN case the client is connected to the LAN and the EdgeWare Server s cable is connected to the LAN and reaches far enough to attach to the Server 2 Connect supplied power cables to all the power connections on the power supply of the units at the back of the EdgeWare Server 3 On the back of the EdgeWare Server connect the Server s Ethernet cable to the network interface marked eth0 If there are no distinguishing marks and more than one network interface usually the left one or the top one is ethO Experimenting may be necessary 4 Connect the chassis front panel bezel shipped with the Server for EdgeWare 12 bay series units 5 Push the power button on the front panel to turn the Server on In some cases the power button may be located on the back or be very small or recessed In such a case a simple tool like a ball point pen will depress it 6 Check for a good connection Page 10 Step 6 refers only to a powered up hardware connection not to the possibility of communicating on the LAN which will not be established until step C It is pos
75. Other Network File and Print Services 0 0 MB dl UPrP User Interface wi E Outlook Express OOMB M FEA ndate Rant Certificates NNMR Description Contains a variety of specialized network related services and protocols Description Allows you to find and control Intemet connection sharing hardware and software that uses UPnP TM Total disk space required 54 5 MB Space available on disk 5623 0 MB Total disk space required 54 5 MB Details Space available on disk 5623 0 MB Figure 2 Installing UPnP in Windows XP r EIERE File Action View Help eo m SRR R gt m i m Rp Services Loca nme opon o Oos Startup Type Logon as _ By SQL Server VSS Writer Provides the interface to backupfrestore Microsoft SQL serv Manual Local System SSDP Discovery Service Enables discovery of UPnP devices on your home network Started Manual Local Service Ry System Event Notification Tracks system events such as Windows logon network and Started Automatic Local System Ba System Restore Service Performs system restore Functions To stop service turn off Started Automatic Local System Bs Task Scheduler Enables a user to configure and schedule automated tasks o Started Automatic Local System Sa TCP IP NetBIOS Helper Enables support for NetBIOS over TCP IP NetBT service a Started Automatic Local Service Extended A Figure 3 Locate SSDP Discovery Service under services UPNP NETWORK DISCO
76. P Failover page and check that the heartbeat monitor is up Get Resource shows a small blue button with an arrow and Apply Changes and Stop Heartbeat buttons are visible just like Figure 124 Do not change anything on the follower 3 Logged into the leader in Clustering gt IP Failover wait 30 seconds to a minute and refresh the screen by selecting IP Failover button on the left The result on the leader is Page 155 shown on Figure 125 If the green does not show up wait and try again As a last resort hit Get Resource A check of the follower will show Figure 126 Now the resources are captured by the leader and not possessed by the follower which is why the Resource Status is opposite for them Correct state at this point is as follows The heartbeat communication is happening over the private connection eth1 All HARs are active but connected to the cluster floating IP address not the leader s IP address The cluster IP address is visible on the leader s ifconfig as eth0 0 which is separate from the leader s old network IP address at ethO Pinging the cluster s IP address works after step 3 not before The leader shows green Up block and the follower red Down block under Resource Status Primary Node IP Addresses Node Services Resource Status Get Resource Leader22 192 168 1 158 hb_mounts services r Figure 125 Leader s response to Get Resource button selection EdgeWare Clustering Conf
77. S 9 HBS A Storage ZS rking Authentication Maintenan Log Out Manual Server Name CEZE47BA Server IP Address 192 168 1 31 Server Time amp Date Tue Sep 16 14 10 37 PDT 2008 Server Uptime 14 10 37 up 4 days 3 44 User Logged In admin User IP Address 192 168 1 33 PHYSICAL RAID information oo O RAID Status OK Number of Drives 8 All Working Spare Drives No Spares Temperature Status Temperature is normal USAGE volume information CS Total Usage 0 System Time Administrators iit Sambal Samba is stopped Total Connected Users 1 Figure 102 Remote server s top UI screen accessed from NAS Servers Index It is true that such a remote login can be used just like a standard client login to control the neighbor machine via its GUI However this is not recommended for two reasons One is the above mentioned confusion The other is that every command has to go through RPC from the home server which makes it very slow The recommended approach is after confirmation that the connection is made to log into the neighbor directly from the client in another instance of the browser Windows can be maintained side by side fast and easy to keep track of compare Figure 109 below And the same NAS Servers Index approach can be used to reach the original home server from the second server s GUI resulting in true peer to peer symmetry which is required for the han
78. SYSTEM ON THE DRIVE is ie ie es ee bes See seek ie ese aeaaea aa NS EN ea A ree Audi ses Ee dees 77 MOUNTNG THE FILE SYSTEM america eeni oes needs ee ee we ee ede ee ese dese ek oe ee ee ed de dee ee GE NRS 77 QUICK SPEED AE N RE N EER OR RE EE RR a 78 LONG DATA AND STABILITY TEST ies nee ER ees eed ee Dene van ree eene A oe eke ee ee Gee oe ie eg 78 Fibre Channel Target Module ses EE DE RE EE ER NE Ee EN GE Ge AO EE AR EE ee ARE Re ee ee AD Eer old 79 PIBRE CHANNEL TARGETSE Eet EE Eege 79 ENABLING AND DISABLING FC ADAPTERS FOR TARGET MODE SUPPORT ss ssssssssssssssssesseestsessrstsetsssesrseesenessrseserses 81 DEFINING LOGICAL Nie DE 81 ASSIGNING LUNS TO THE DEFAULT SECURITY GpOUp ese see sesse se ee eke se ee ee Re Ge ee ee GR Ge ee Re Ge ge ee Re GR ee ee ee 82 ADDING ADDITIONAL SECURITY GROUPS sesse see sesde see Skenk seek kk vee rosero touet vienesa sasis Ee si Vee Age SEE Se NEEN eek ese deed 84 How TO CONNECT WITH A FC SWITCH AND Hoer 87 FORMATTING THE FC VIRTUAL DRIVE sesse see see sesse sege se ee sd ese sd oe ee eg dd See Ge dd ee Ad ee ed Age Ed k 87 MEE oR de EE OE EE nawenenneanais 88 Manie ARE RE EE EE OE N OER EE HER 88 HTS EE 92 WINDOWS DOMAIN e EE 97 SERVER CLIENT ACCESS TO CIFS iis ese ee ee ge eE ee ee oe EEEE EEE OEE AAEE N E e 98 SE EE EE EE N E OE EE 99 EE 103 a INL 106 REMOTE REPLICATION ASYNC E 107 Au thenticatiOM EE 109 USERS AND GROUPS RR EE EI 109 SHARE di Se RE EE EE RE EE RE ER ER 111 CHE 113
79. Source Software OSS program called Webmin Webmin provides a solid base architecture and a means to leverage other work for the satisfaction of the customer GUI adds many custom storage management modules that allow quick and easy configuration with advanced features ACCESSING THE GUI The EdgeWare GUI like Webmin uses port 10000 for web browser access and discovery UDP broadcasts Once the hostname or IP address is known using discovery tools or Domain Name Services DNS the GUI is accessible from any web browser by the URL https IP address 10000 or https hostname 10000 Initial access is by user admin and password setup Figure 4 shows a web browser URL consistent with using HTTPS and port 10000 ne Fates Nise GEBOABIE 192 6B AL IE Mosslla Faretox File Edit View History Bookmarks Tools Help D GG DE Figure 4 GUI HTTPS and Port 10000 access line Webmin is organized as a tree start by selecting a section top row of icons then a module left side then continue by selecting component and even deeper features if necessary until reaching the page you need To back out closer to the trunk you normally either do an input or change or select Return to Index which is found near the bottom of the page Sometimes scrolling down is necessary if the page is too big to fit on the screen That backs you out one or more levels possibly up to module To change modules or sections select the corresponding ico
80. T MODULE This module provides management options for configuring access and settings for the web UI There are four sub sections under this section 4 ed Kl g ZS Ge amp A Networking Authenticatio jaintenance justering og Ou Manual EdgeWare Server Configuration Management Provides management options for EdgeWare Configuration O lt on ad access and settings for the EdgeWare 3 0 0 1ce weie e So Tv IP Access Control Language SSL Encryption Administrators Figure 6 Management module main page IP ACCESS CONTROL e Here you Can control access to the web UI based on IP information of client machines Select your choice of the three access modes and enter the IP addresses in the text area e If you want to restrict or allow clients from a particular subnet then the format to enter a subnet is lt IP address gt lt subnet mask gt e For instance if you would like to block or allow the 192 168 0 0 subnet enter 192 168 0 0 255 255 255 0 e If you wish a reverse name lookup to verify the IP address has a hostname and vice versa you should select the check box below the entry box Selecting the check box enforces reverse name lookups LANGUAGE Display the Language for the EdgeWare OS which is English as default Page 22 SSL ENCRYPTION Here you can create SSL keys for you server Enter the requested information in the text fields and click on Create Now CERTIFICATE AUTHORITY To set up yo
81. This means that the mirrored volume you are working on must be temporarily demoted to raw before its primary and secondary are swapped or before it is deleted This is done in the same way as described above Step II parts 3 4 for mirror creation SWAPPING STATUS OF PRIMARY AND SECONDARY Remote Mirroring is implemented on the Local and the Remote Volume Once Remote Mirroring is configured we can change the properties of the Local Volume by using the following feature Make Local Secondary Pressing the button will make the localhost Volume secondary This action will result in renaming to button to Make Local Primary Make Local Primary To take the local Volume back to primary click on this button Note When we make local Secondary on one server we need to login to the other server and make that primary so that they can handshake ALWAYS change primary to secondary and THEN change the other servers secondary to primary Having both Storage Servers primary on the same volume at the same time may cause data corruption or deadlock DELETING THE MIRROR To delete a mirror does NOT mean to delete the data on the volumes comprising the mirror Instead it is a non destructive process The mirror connection is removed leaving two separate volumes one on each Storage Server After delete is done f the mirror is fully synced these two volumes start up identical but can take on an independent existence after that If the mirror is delet
82. VE OR IMPUTED NEGLIGENCE STRICT LIABILITY OR PRODUCT LIABILITY OR OTHERWISE FOR ANY SPECIAL INCIDENTAL CONSEQUENTIAL OR INDIRECT DAMAGES INCLUDING BUT NOT LIMITED TO LOSS OF USE LOSS OF DATA BUSINESS INTERRUPTION LOSS OF REVENUE LOSS OF BUSINESS OR OTHER FINANCIAL LOSS ARISING OUT OF OR IN CONNECTION WITH THE SOFTWARE OR OTHER GOODS OR SERVICES FURNISHED TO LICENSEE BY MANUFACTURER EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGES LICENSEE SHOULD MAKE PERIODIC BACKUP COPIES OF THE DATA STORED ON THE CUTTING EDGE NETWORK STORAGE NODE OR OTHER CUTTING EDGE AUTHORIZED AND SUPPORTED HARDWARE AS A PRECAUTION AGAINST POSSIBLE FAILURES ALTERATION OR LOSS OF THE DATA CUTTING EDGE IS NOT RESPONSIBLE FOR DAMAGE TO OR LOSS OF ANY DATA STORED ON THE CUTTING EDGE NETWORK STORAGE NODE OR OTHER CUTTING EDGE AUTHORIZED AND SUPPORTED HARDWARE 2 Software Maintenance a SERVICES AND SUPPORT PACKAGE Manufacturer will provide to Customer the technical support and maintenance services Services for the following support packages selected by the Customer Support Package Exhibit A Gold Support 1 year Patches Fixes Minor Upgrades Next Business Day Response COVERED PRODUCTS Manufacturer will provide the Services solely for the eligible products set forth in price list that are purchased by or licensed to Customer from Manufacturer Covered Products SERVICE MODIFICATIONS Manufacturer reserves the right
83. VERY VIA A WINDOWS VISTA CLIENT e In Windows Vista uPnP has been renamed Network Discovery e Windows Vista Network Discovery uPnP discovery protocol allows the server to advertise its service to a control point of the network The control point searches for other servers on the Network and discovers them e Normally Network Discovery works automatically as described above in Set Client IP Address and Netmask on Microsoft Vista If not it can be activated as described below e Network Discovery in Windows Vista can be located or activated in the following manner i Select the Windows icon in the lower left corner of the screen ii Select Control panel Page 17 iii iv V vi Select Network and Internet Select Network and Sharing Center Find Network discovery and check whether the on dot is marked If not click the large dot with the v in it to expand Network discovery and click Turn on network discovery e Network Discovery once activated is used in the following manner i ii iii iv V vi vii Select the Windows icon in the lower left corner of the screen Select Control panel Select Network and Internet Select Network and Sharing Center Near the top icons representing your network connections will appear Double click the one that represents the network or connection containing the EdgeWare Server After a few seconds a screen will appear with all the connected servers
84. Ware Storage Configuration Backup Management P amp Provides data backup service for the Restore Backup H logical volumes EdgeWare 3 0 0 1ce Z HELP This page allows you to restore a previous volume backup Be careful when restoring as files modified since the backup may be overwritten Restore backup of filesystem Restore from O Tape Tar File Viocal Browse P Restore to directory LVstorage Browsen N Files to restore Everything in backup O Listed files Separate multiple file names with a comma Backup is split across multiple files tapes O Yes No Show list of files only for Test O Yes No E Hi z Restore Backup Now Return to index Figure 45 Restoring a backup Page 63 4 If the backup was spanned across multiple files or tapes then select yes for Backup is split across multiple files tapes 5 Select yes to Show list of files only for test if you just want to view a listing of files in the backup This will not restore data 6 Extra command line options should be left blank 7 Click Restore Backup Now this will start the restore process or print out a listing of files in the backup if selected in step 5 8 If your backup is on multiple media the tape drive will eject the media when it is ready for the next ordered media in the backup media set Just insert the next media and the restore process will continue automatically 9 When the restore proc
85. X Shutdown LA SNMP Version SNMP v 1 Monitoring e Public Community public ES eg Private Community private eer Contact Person Larry ER Physical Server Location El Cajon a Command Line IP Address 1 192 168 1 17 P Integrity eh IP Address 2 Le ge E Save gt Reset gt Download MIB file Stat Click on this button to start the SNMP server MAY TAKE A MINUTE 4Returntoindex Figure 91 SNMP after data entry but before Save and Start Page 123 If an SNMP MIB browser is to be used not just traps be sure and select Download MIB file while running EdgeWare UI within a Web browser on the client system that has the MIB browser Then go into your MIB browser and attach the EdgeWare MIB file Cuttingedge Storage Solutions MIB txt sometimes presented as all caps to your MIB tree Then give your MIB browser the EdgeWare IP address select the EdgeWare MIB and perform a Get Next and this should reveal the first EdgeWare MIB datum Now you have access and limited control over your EdgeWare Storage Server from the MIB browser Warning If MIB browser access is to be used it is a good idea to change Private Community to something other than private as this is a password not a user name and Private Community allows writes from your MIB browser that can do administrator level changes on your Storage Server Because of this security weakness EdgeWare severely limits the adjustme
86. a RAID set RAID increases reliability and often improves performance depending on the usage model discussed with the Volume Manager module EdgeWare OS supports up to four types of RAID configurations for the storage arrays These RAID levels can be used on a multi drive appliance whether or not a hardware RAID card is installed If there is no hardware RAID card installed you can use software RAID It is important to note that for a two drive appliance the only RAID levels available are 0 and 1 For an appliance with three or more drives RAID levels 0 1 and 5 are available At 4 or more drives RAID 6 also becomes available The RAID module contains two components hardware RAID and software RAID Generally EdgeWare systems have hardware RAID controller s The RAID sets in this case are controlled by the hardware controller s For one type of controller there is a component 3ware Hardware RAID which provides information about the hard drives connected to the RAID controller RAID sets and configuration options Another form of hardware RAID is the Xyratex controllers that export Fibre Channel devices For such systems there is a component Fibre Channel Arrays which provides information and configuration options for this kind of hardware RAID The Software RAID component allows you to view create and configure RAID sets via the operating system The lower end systems such as the 4i and Datacube do not have hardware RAID co
87. a monitoring job Each monitoring period can have more than one type of alert defined like mail alert snmp alert alarm alert etc Figure 89 shows a single monitoring period Monitoring period 1 for the 3ware hardware RAID service Page 121 Specified days and hours Define the days and hours the monitoring process will check this service at the time interval specified in the Check every setting Alerts for period Alert Type of alert to be sent The different types of alerts available are Purpose Additional Parameters Additional Parameters Example Mail alert Send alert via recipient email address user yourdomain com email file alert Save alert to a file name with full path Volumes LV1 raidalert file winpopup alert Send a popup NetBIOS host name of the MyServer1 alert to a workstation windows workstation Alarm alert Send sound Number of beeps to sound 3 alarm to chassis for one instance of the alert speaker snmp alert Send an SNMP community manager IP public 192 168 0 6 RAID trap based on address Watch Service Degraded_RAID additional parameters Extra message Trap alert Send an SNMP trap to the SNMP host with settings from the SNMP page Run when The event which should signal the alert Advanced Configuration The Advanced Configuration button will give you more options for the Monitoring Period Page 122 Days to check You can specify all days of the week to monitor or a range of se
88. aa Figure 72 Advanced Network Configuration Page 91 CIFS The CIFS module Figure 73 gives you the ability to set up native Microsoft Windows networking protocol Figure 73 shows the Samba Servers stopped before any setup has been done Networking H Server Storage Networking Authentication Clustering Log Out zech Maintenance Manual EdgeWare Network Configuration CIFS Windows Networking Options ae Z EdgeWare 3 0 0 1ce 4 HELP Share Name Path edgeware varfedgeware sharestorage shares sharestorage Global Configurations and Options o el Windows Networking Options Windows Domain Users and Grou Start Samba Servers Click this button to start the Samba CIFS support for the shares listed Figure 73 CIFS module main page Before proceeding to this CIFS module you should have a share set up as described in Storage gt Share Manager module above The share edgeware does not count as it is a system share with no external access permissions If Samba is stopped as in Figure 73 click the Start Samba Servers button Give your Windows client 10 minutes or so to pick up the new Samba CIFS server Then in your Windows client in the My Network Places page under Local Network an icon should be visible for the share in this case sharestorage Hovering over it will disclose the full name of form SERVERNAME sharename where SERVERNAME is all caps In this case it was LOCKHEEDSERVER sh
89. access without password is authorized Then the prefix in the URL line is ftp instead of the usual http or https HTTP The HTTP module is an EdgeWare administrator file manager for the HTTP protocol It is not a web server module for the m DAR See the web server section of the m DAR manual for explanation about that feature The HTTP module is not a wide area HTTP web server of the familiar apache type This module is a Java utility called File Manager which allows EdgeWare administrator users set up in the Server gt Administrators module described above in this manual to manage and transfer files and folders on the shares via the HTTP protocol It does not allow such access to standard users set up in Authentication gt Users and Groups or to anonymous browsers Use FTP above or install HTTP web server code on a client system for anonymous browser access You the administrator can use HTTP File Manager to upload download find delete copy cut and paste files in those shares which have HTTP protocol enabled by Storage gt Share Manager above You need Java 2 Runtime Environment JRE version 1 4 2_03 or later installed on your workstation for File Manager to run successfully You can get it at http www java com Select the HTTP button on the left of the EdgeWare menu After a pause of a few seconds and possibly a certificate approval Java gives you the screen shown in Figure 80 The left pane shows the di
90. aded with the words Local Area Connection or Local Area Connection 2 or similar should say Connected or Connected Firewalled or similar Right click the identifier for the network connection leading to the Server Select Properties Pick Internet Protocol TCP IP Select Properties For DHCP leave the radio button on Obtain an IP address automatically and exit without making changes For non DHCP static address specification move the radio button to Use the following IP address and fill in appropriate values for the IP address and netmask If a part 0 to 255 of the quad is less than 3 digits right arrow moves you to the next part In the Auto case the right IP address value is 169 254 x y where in place of x and y you enter numbers between 1 and 254 and the right netmask value is 255 255 0 0 Do not fill in gateway SET CLIENT IP ADDRESS AND NETMASK ON MICROSOFT VISTA If you have a point to point connection it should not normally be necessary to set a client IP address Vista should detect the absence of DHCP and do it for you Select the Windows icon in the lower left corner Select Network And your EdgeWare Server should be visible with an icon labeled Cutting Edge Edgeware NAS Double clicking will show its IP and open it If you are curious about your own IP click Page 12 the Windows icon type in emd and hit lt Enter gt and type in ipconfig and hit lt Enter gt in the resulting screen If you
91. al minutes for its initialization to get well underway 4 Click on Create The new Volume Group shows up expanded as in Figure 27 Note Screen response may be slow If an error screen appears give it a few minutes to initialize some more and try Steps 1 4 again Ke es E e g KERE A Storage Networking Authentication Maintenance Clustering Log Out Pd r H e N EdgeWare Storage Configuration Volume Management piece see ee ep oe functions like logical volume creation Molume Management O physical volume aggregation and Storage EdgeWare 3 0 0 1ce snapshot creation etc Se Add a New Volume Group SS Volume Tree Show hidden volumes Figure 27 Volume Group VGstorage is expanded Page 44 ADDING A PHYSICAL VOLUME Your original Volume Group has only one Physical Volume If this is not enough you can now extend the Volume Group by adding a Physical Volume 1 Select Physical Volumes add a physical volume in the expanded Volume Group tree see Figure 27 2 Select the physical device from the drop down list Leave Enabled for allocation on yes 4 Click on Add to volume group Provides logical storage management EdgeWare Storage Configuration Volume Management EG ee st d Add Physical Volume O functions like logical volume creation physical volume aggregation and EdgeWare 3 0 0 1ce snapshot creation etc HELP In volume group VGstorage Phy
92. al volumes which are block addressable storage units used to build up the volume group They have to exist before the volume group can be formed They can be added to or even removed from an existing volume group Usually they string together nose to tail but they can be joined in more obscure ways too like striping Even physical volumes are already abstracted they are typically the devices or arrays output by hardware or software RAID not the individual hard drives from which these are built The descendants are logical volumes and snapshot volumes which are cut from the volume group like partitions from a disk and presented to the outside world as block devices with all the power that block devices have They can underlie file systems can be iSCSI targets and can be raw targets as for mirroring The volume group has to exist before these can be made Snapshot volumes are treated as special because they are not independent each snapshot volume is related to a specific pre existing logical volume and aids in the snapshot feature to be described later EdgeWare uses this enhanced volume management to virtualize the real physical disks that exist on one or more of the EdgeWare Storage systems That provides the EdgeWare Storage system with the flexibility to merge different or even dissimilar physical devices e g iSCSI and local disks into a single pool of storage to create block level snapshots and to easily shrink or extend log
93. an be started restarted or stopped from either the SCST initiator page or the iSCSI target page both perform the same function iSCSI TARGETS An iSCSI target is the server side of the iSCSI protocol In EdgeWare it presents hard disk storage EdgeWare can present one or more than one iSCSI target each of which may or may not be connected to an initiator Each target must have its own dedicated Logical Volume which cannot be used for anything else An iSCSI volume cannot support any NAS protocol or be a source or target of Asynchronous Replication although it is possible for it to be a Synchronous Mirroring source and or a Snapshot parent since those also are block level protocols Among other network based capabilities that can be supported by iSCSI volumes are Failover Bonding and MPIO MCS CREATING AN iSCSI TARGET An iSCSI target is a Logical Volume and is created in the Volume Manager just like a standard NAS volume The difference is that you must specify iSCSI Target in the Network File Protocol drop down menu as shown in Figure 48 below Actually iSCSI Target is not a file protocol It denotes the absence of a file protocol and also notifies the EdgeWare iSCSI target software of the new volume s existence Specify a size in GB and a name for the iSCSI volume The volume name you specify will be added to the end of the string making the internet qualified name for the iSCSI target Clicking on the Create button will cr
94. and be replaced without bringing down the complex whole of which it is a part even temporarily A homely example is a chandelier in which you can replace a burned out light bulb without plunging the room into darkness EdgeWare offers several kinds of failover e Power supply failover the two plugs in the back squeaks but keeps working when a power supply dies e Network failover such as bonding or MPIO allows network components to go down without interrupting service e Storage failover or RAID redundancy allows swapping out of failed disks without losing data or stopping the storage device e IP failover allows a two headed device to share a single IP address on one side and a single storage pool on the other and to suffer a failure of one of its heads central processing units with operating system and software while preserving data flow between its network connection and its storage It is the last form of failover that is supported by this module THE DESIGN OF IP FAILOVER IP failover is so named because it permits a single IP address to be shared between two CPU motherboard units or heads commonly called NAS heads but this is confusing because it is possible they may be iSCSI targets and not NAS at all Everything on the LAN side of the handshaking Storage Servers including an iSCSI initiator if applicable sees only the one address The failover code called heartbeat sees to it that that address feeds the active head an
95. anges then Start Restart the replication support from the main module page NOTE To start the asynchronous replication of a share you enabled here you have to configure and set up a target volume in the module Clustering gt Remote Replication on another EdgeWare Storage Server Page 108 AUTHENTICATION The authentication module deals with user and group access security and permission Most of the NAS protocols depend on user and group authentication to determine how much access if any to grant to an external user attempting to connect through the network to EdgeWare storage External authentication for EdgeWare is primarily supported by Windows Domains and or by two externally served protocols Network Information Service NIS and Lightweight Directory Access Protocol LDAP In these cases the EdgeWare Storage Server usually acts as a client system and another system on the network must provide the services The exception is a Windows Domain run by EdgeWare s own Samba server USERS AND GROUPS This module Figure 82 deals with the creation and management of users and groups under local authentication or externally served schemes like Windows Domains Microsoft Active Directory NIS and LDAP As a rule it is local authentication that allows you the administrator to create new users and groups within EdgeWare or edit or delete them Externally served authentication may present EdgeWare with a aif accompli in the sense that the
96. arestorage see Figure 70 under Hostname If Samba was already running when this or any other changes were made select the Restart Samba Servers button to register and advertise the new configuration Edit Security Share Name Select Page This page Figure 74 is reached by selecting a share name in Figure 73 the CIFS module page In the case of Figure 74 the new share sharestorage was the one selected Page 92 EdgeWare Network Configuration CIFS i GE A S O rovides settings to ena Edit Security J communicate with Microsoft networks EdgeWare 3 0 0 1ce HE HELP Security and Access Control Writable Yes No Support Shadow Copies Yes No Guest Access None Yes Guest only Hosts to allow All Only allow Hosts to deny None Only deny Valid users Browse F Valid groups Invalid users d Browse F Invalid groups i Save Lei Return to index Figure 74 Edit Security CIFS share name page Normally before proceeding to this page you will have set up users and or groups in the Authentication gt Users and Groups module described later in this manual By filling in through browse and select valid or invalid users or groups you make it possible for certain users and not others to access the data on this share What you select here is ocal to the share Windows Networking Options On this page Figure 75 you can do basic g
97. art LI Figure 85 NIS client configuration after Save and Restart NIS Client See Figure 85 Enter the NIS client settings for your NIS domain NIS domain The NIS domain name This should be different from your DNS name NIS servers NIS servers should be entered here or you can choose to find NIS servers by broadcast Authentication Order If a same user is present as a local user and a NIS user choose the order with which the user will be authenticated Click on Save and Restart after making any changes to apply those changes NIS Users and Groups This lists the NIS users and groups found on the NIS domain by the NIS client Name Service Switch This page is for advanced users The form edits the etc nsswitch conf file which defines the search order of the network databases for different services and protocols running on the EdgeWare system Page 113 LDAP The LDAP module permits you to register the Edgeware Storage system as an LDAP cent LDAP Lightweight Directory Access Protocol is a protocol for network wide access to a database Directory in the name means database which is used by EdgeWare for authentication Before this module can be used there must be an LDAP server on the network and you must know its name and information to enter into the client settings LDAP is friendly to CIFS Samba Its password security is more robust than that of NIS but it is more complicated to set up because it is designed
98. as suffered a drive loss the status will change to degraded The Percent of resync done field shows the amount of RAID rebuilding that has been completed It is recommended that you perform maintenance monitoring and configuration of 3ware RAID sets from the 3ware manager Refer to 3ware Hardware Monitoring section below Ed ge Ware Maintenance Clustering Log Out ad 7 EdgeWare Storage Configuration Raid Configuration Steg i View 3Ware Disk Set O options for software and hardware RAID EdgeWare 3 0 0 1ce Storage RAID RAID device options oo Manager 3ware Controller 6 KI Gott Device file dev sde Disk Set Type Mirrored RAID1 EE Status Available d6 eege Usable size 279 GB et Percent of resync done 0 wW Backup Persistent superblock No n Stripe size CSS ses Storsave Policy E ER Disks in Set Port 2 LEOL4ZSH Port 3 BE2137HH Spare Disks Delete 4Return to index Figure 21 3ware RAID unit 0 in detail no hot spare disk present 3Ware Hardware Monitoring RAID needs monitoring to avoid loss of data at times when the physical disks do fail The 3Ware RAID monitoring and configuration should be performed by accessing the 3ware web based RAID manager 1 Log in to the Storage Server web management at https Storage Server IP address 10000 Start the 3ware manager under Storage gt RAID gt 3ware Hardware RAID Start 3Ware Manager if it is already running the option displayed wo
99. at for the primary side This works because of a major design feature of the Heartbeat it owns rather than shares its communication resource the point to point Ethernet connection Thus it can be treated as parallel rather than nested to the HARs and the order of startup is immaterial However EdgeWare currently requires that the Primary s HARs be set up before the heartbeat is turned on Nesting within the HARs is not immaterial They are shut down in the opposite order from the order in which they are started up The classes of HARs are shown in Figure 115 NESTING CATEGORY RESOURCES Outermost hb_mounts mounting of logical volumes Middle services smb nfs atalk iscsi_trgtd proftpd rsyncd Innermost drbddisk drbdswitch pl iscsi only Figure 115 High availability resources and nesting Here outermost is closest to the hardware and therefore longest lived first up last down The six services listed under services are treated as not nested or as independent of one another since they are shut down in the same order as they are started Note The leader starts out as Primary and the follower as Secondary But a fail event followed by a replacement of the leader may cause these roles to be reversed In all the discussion below the terms leader and follower stay with the hardware and Primary and Secondary follow the roles CHECK LIST FOR EDGEWARE IP SERVICE FAILOVER The protocol resources IP ser
100. ation ess x EE ER een EdgeWare 3 0 0 1ce 4 HELP Available EdgeWare Servers CE485192 4Returnto index Figure 104 Configure Remote Replication first screen 3 Click on Next A screen like Figure 105 appears T 4 Select the source share on the remote server from Asynchronous Shares on Server drop down list already done in Figure 105 Page 140 KAY A Maintenance dd Ss a CE485192 Local Figure 105 Configure Remote Replication second screen 5 Select Browse next to Local Volume on Server and select the destination volume from the popup Hit OK The result should be as in Figure 105 6 Click Next 7 Configure the mirroring schedule Figure 106 Authentication Maintenance Clustering Log Out a Figure 106 Configure Remote Replication final screen F TARGET TRIGGER THE REPLICATION 8 Click on Save Figure 106 to create the remote replication asynchronous mirror See Figures 107 with specific code number and 108 The replication will commence at the scheduled times Page 141 HELP Successfully created the mirroring schedule 141661221678681 Return to index Figure 107 Remote Replication Save response screen dd es D e J gA A E d ge wW are Server Storage Networking Authentication Log Out User Maintenance Clustering Manual EdgeWare Clustering Configuration Replication 2 Scheduled opi ofthe data O eneen e
101. c file etc hosts just as step D makes the IPs static You have imposed a total of eight entries counting both servers Be sure there are no typographical errors Both servers Host Addresses lists should contain the same data though it need not be in the same order Figure 113 F CAPTURE EACH SERVER S PORTS IN THE OTHER S NAS SERVERS INDEX This is carried out in Clustering gt NAS Servers Index on each server See the instructions for this module above in this manual Since you know the hostnames and the IP addresses of the Page 146 neighbor you can use Register a New Server if broadcast is not desired On each server capture both the eth1 hostname and the standard hostname of the other server so that the neighbor appears as two EdgeWare icons in the NAS Servers Index one with the standard name and one with the ethi name Clicking on either hostname 10000 link should lead to an automatic login compare Figure 102 Both logins will show a top page with the lead hostname and eth0 IP address If you use Register a New Server t is most convenient to enter the eth1 hostname first and then the standard hostname Page 147 IP FAILOVER HEARTBEAT MONITORING Many customers require to get as close as they can to assurance of being up 24 7 365 or at least not to have unscheduled downtimes However no component of hardware lasts forever Failover is the technical term for a design that allows a component to fail
102. ceed as follows Page 62 1 Enable backup schedule from Scheduled backup enabled at the top 2 Select the interval at which backups are to be made from the drop down list 3 Click on Create to save the backup job schedule RESTORE To restore from backup 1 Select the backup target you want to restore from For tape insert the correct media or the first media of the media group Figure 45 Select the location to restore the backup Click the browse button next to the Restore to directory text box a pop window will appear where you can select from available and valid locations Next select Everything in backup to restore all the data in the backup or select Listed files to restore specific files If you select the latter option make sure that the filenames are separated by a space and that the filenames must be relative to the file system that the backup was made from so if you backed up home and wanted to just restore home foo you would enter foo into this field or if want to restore a file foo in a directory home bar you would enter bar foo NOTE By default the restore process will not overwrite a file when a matching file is present on the volume being restored that is newer than the file in the backup S Storage RAID oo Manager KI Share Manager GG wow a Backup aintenance Clustering H Server Storage Networking Authentication M Log Out Manual Edge
103. d by a monitor RAID monitor which checks on the status of RAID sets defined on the RAID controller card s If the monitor finds any RAID set is unhealthy it sends out two alerts by default The alarm alert sounds three beeps every 5 minutes and the mail alert sends a mail out to the email address specified with a brief description of the problem It will have a line similar to Ware controller 0 unit 0 is I which means that RAID unit 0 on the 3ware card had detected inconsistency or an unclean shutdown and is re initializing or initializing the RAID set A 3Ware controller 0 unit 0 is D message will be sent out in a mail alert when RAID unit 0 on the 3ware card has degraded which most likely means that a drive in the RAID unit 0 is bad Hardware Hardware component temperature is monitored by shuttle hardware monitor This service monitor checks the temperature of the system motherboard system chipset and the CPU Critical temperature for the components is 68 C or 154 4 F Page 120 EdgeWare Maintenance Configuration Monitoring Add Service O Setup and manage warnings and errors FdgeWare 3 0 0 1ce HELP Watched service details Name of service Check every seconds Description Using monitor iSCS Target monitor v P iSCSl Target monitor Monitor parameters RAID monitor shuttle sk43g hardware monitor New monitoring period ESE ar i ardware monitor SEES HERE ew msi_ 207 harcware monitor Additional parameters S
104. d in DFARS 227 7202 Oct 1998 All other use duplication and disclosure of the Software and related documentation by the United States shall be subject to the terms and conditions set forth in this Agreement and the restrictions contained in subsection c of FAR 52 227 19 Commercial Computer Software Restricted Rights Jun 1987 or FAR 52 227 14 Rights in Data General Alternative III Jun 1987 Contractor licensor is Cutting Edge 1825 Gillespie Way Suite 100 El Cajon CA 92020 Customer may not copy publish modify prepare derivative works reverse engineer decompile disassemble lease loan sublicense timeshare or transfer any products or remove alter or obscure any proprietary notices on any products in any way without Manufacturer s prior consent Manufacturer licenses software to government customers only with restricted rights under this Agreement If applicable laws and regulations would grant any government customers greater rights than granted in this Agreement then the government customer will receive only the minimum rights required by those laws and regulations Use copying or disclosure by government customers is subject to applicable restrictions in any of the following paragraph c of the Commercial Computer Software Restricted Rights June 1987 clause at FAR 52 227 19 and the Restricted Rights Notice of subparagraph g 3 of the Rights in Data General June 1987 clause at FAR 52 227 14 Cutting Edge
105. d installs update fo Edgeware Updates f Keser EdgeWare 3 0 0 1ce L HELP Update list follows Scroll to bottom to apply or cancel Package Version Repo iscsi target x86_64 2 0 4 3ce ew extra opensip x86_64 1 2 1 4ce ew extra perl Math TrulyRandom x86_64 1 0 4ce ew extra Apply Cancel Figure 99 Updates list window The list of three updates found in Figure 99 will be replaced by the updates pending for your installation This list may be long perhaps more than a full screen s worth If after examination of the list you do not want to proceed to updating at this time click Cancel That will return you to the module main page Click Apply to download and apply the updates to the system This can take a number of minutes depending on the speed of your Internet connection The progress will be shown as each file is downloaded and installed on the system followed by Done upon successful completion of the update Never reboot the system while updating is still underway Unless a specific message appears saying to reboot you do not have to reboot the system at all It is possible the browser may time out while updating is still underway This does not mean the update has failed The proper response is to refresh the browser window usually a button at the top that shows a circulating arrow or arrows Hovering over it says Refresh in Internet Explorer Page 133 or Reload current page in Firefox Yo
106. d moves to the other one should the currently active one fail Since the two heads do not share a system disk this requires delicate handling in order to keep both Storage Servers consistent under changes One head the leader takes on the role of Primary or master of the high availability resources services or HARs that are to be failed over The other head the follower takes on the role of Secondary or passive recipient at failover start The general rule about configuration sharing is e Information about storage arrays and volumes Storage modules RAID and Volume Manager is stored on the storage arrays and therefore is automatically visible to both heads at least after a reboot e All other system information is stored on the system disks and must be copied across the network to make the two heads consistent This is accomplished by the heartbeat service running on both leader and follower Page 148 PRIMARY SECONDARY Hardware Hardware System System Network Network HARs Heartbeat Heartbeat HARS Figure 114 Order of resource and failover activation The dotted lines in Figure 114 refer to a network communication that must take place The rows from Hardware through Network are taken care of in the Handshake setup The last row on the Secondary side takes place only if there is a fail event It is worth noticing in Figure 114 that the nesting is shown as parallel between HARs and Heartbe
107. d so on till all the data has been backed up Give a name to the backup job for Label Choose to specify the Tape size or let Backup process calculate by querying the tape drive Block size value should be default unless you have a specific block size you want to enter Select yes for Software Compression only if compression is desired when backing up to Tar file Click on Continue WO ONAN In Figure 44 you can specify a schedule for this backup to run on If you do not want to specify a schedule right now you can click on Create and Backup Now to start the backup immediately By editing your backup you can later add change or remove the schedule ai a es E e g SS ZS A E d H eWa re Server Storage Networking Authentication Maintenance Clustering Log Out wer Manual EdgeWare Storage Configuration Backup Management TEER eae 7 Schedule Backup O See Storage EdgeWare 3 0 0 1ce RAID HELP Vou Manager Scheduled backup enabled Disabled Enabled at times chosen below KI mme Job schedule Every 10 minutes Every 10 minutes oS Quotas FEE Create Create and Backup Now a Backup ER Return to index ght YV Weekly on Sunday Figure 44 Specify a backup schedule SCHEDULING BACKUP In Figure 44 you can specify a schedule for this backup to run on By editing a backup that was made without a schedule you can later specify a schedule also Pro
108. database file created this section is automatically enabled Page 130 System Configuration Backup creates a tar file with all the different types of configurations you choose You should save that file on another system and when the need arises you can restore configuration of the Storage Server using that file EdgeWare Maintenance Configuration Integrity O Check OS integrity Backup Settings 4 EdgeWare 3 0 0 1ce HELP Last Backup Configuration Options Authentication Web_Management Shares Volume Network Scheduled_Jobs Volume_Backup System Time dd 9 9 9 9 9 9 IE Failover Process H Return to index Figure 98 System Configuration Backup Settings window Backing up Storage Server Configuration 1 Select the configuration types you want to back up The different types of configurations that you can back up and restore are the five complex configurations many items each Authentication This contains all the configuration information about local users and groups defined on the Storage Server Web Management This contains all the configuration information about the EdgeWare UI administrative users and groups Shares This contains all the configuration information about shares and network protocols Volume This contains all the configuration information about the volume groups and logical volumes Network Thi
109. dress 1921681195 oa Save Send Test Email Recipient s email address Hickson cuttedge com 1 Integrity Stop y Click this button to stop the Postfix server B Updates Restart Click this button to restart the Postfix server Figure 92 E mail Notification ready to send test email Page 124 The required parameters are Domain name to outbound mail This parameter specifies the domain that appears in the from address of mail which is posted on this machine It will be prefixed by Test in the test email Clients to relay mail from Specify the subnet or host from which mails can be expected Normally this is left blank as you are not really setting up a mail server only sending outgoing monitoring emails Destination to relay mail to This should be the domain name on which your target email address is found i e the part after the in the target email address the to address Proxy NAT external network address Specify the mail server on your local network that links it to the external email world If your email target is served in your local network this can be left blank The sequence of steps normally followed is OF P AM LOGS Fill out the above parameters Click on Save to save the Email Notification setup Click on Restart or Start to start the Postfix server Enter the target email address in the text box of Send Test Email Click on Send Test Email Wait a few minutes for your test email
110. dshake Page 137 HOW TO EXIT THE REMOTE SERVER S UI When you have accessed a second Storage Server from within NAS Servers Index of a home server even if only for a momentary look around the problem arises how to back out It is NOT advisable to log into the home server from the second server s NAS Servers Index The cleanest technique is as follows e Note the IP address and or hostname of the home server and the second server e If you are logged into the second server exit normally using Log Out icon in the upper right Skip this step if you did not log in that is you only went to the login screen e Use your browser s Back button to back over the screens of the second server These screens are only an inert record you are not really logged in there anymore e When you reach a screen from the home server you are live again and can pick up where you left off REMOTE REPLICATION ASYNCHRONOUS MIRRORING ASYNC Remote Replication enables you to replicate a share on an EdgeWare system to another EdgeWare system based on a schedule For instance when configuring from EdgeWare system A the source share would be on another EdgeWare system B and the destination volume for the mirror would be on EdgeWare System A Remote replication creates an exact replica of the source share data on the destination volume Three points need to be remembered about the structure of remote replication e It is normally s
111. du do RG du qa EDGEWARE USER GUIDE VERSION 4 0 1 JULY 2009 WWW CUTTEDGE COM Page 1 COPYRIGHT No part of this manual may be reproduced stored in a retrieval system or transmitted in any form or by any means electronic mechanical recording or otherwise in whole or part without prior written consent from Cutting Edge All rights reserved Copyright 2006 2009 Cutting Edge TRADEMARKS All trademarks and copyrights stated mentioned referenced or omitted are the property of their respective owners EdgeWare is a trademark of Cutting Edge DISCLAIMER Cutting Edge reserves the right to periodically revise this manual without prior notice Product features and specifications described are subject to change without notice Page 2 REVISION HISTORY za June 2006 e Initial Release October 2006 e September 2007 February 2008 GE Included Internet browser compatibility Added iSCSI documentation How to connect to an iSCSI target How to modify an iSCSI partition Additional information on remote replication Additional information and screenshots on remote mirroring Included documentation on IP Failover for network services and shared storage Moved Updates section from the Server module to the Maintenance module Reason being to reflect the changes made in the graphical user interface Included additional documentation and screenshots for remote replication under the Networking and Clustering
112. e or storage is customarily used to mean very large disk array space typically at least hundreds of gigabytes that is completely available to the customer to configure for use of systems outside the EdgeWare Server Other much smaller disk space is required for the EdgeWare Server itself its operating system logs authentication files and so forth This system disk space should be thought of as separate SCREEN RESPONSE Storage operations controlled by the EdgeWare UI often require most of the server s capability for an extended period of time especially when they are starting and finishing Because of this screen response may be slow Allow several seconds or even up to a minute for screen response to major creation and deletion operations such as RAIDs or volume groups Even updating a changeable screen such as a screen that shows percent completion may take ten seconds or more because the screen must wait for a pause point in the heavy CPU and disk operations of initializing the resource If a screen goes blank or gray or does not seem to respond to a change this may be a normal response to the delay discussed in the previous paragraph Do not interrupt the ongoing operation If it takes more than a minute to respond contact your organization s technical support Page 27 RAID MODULE The recommended mode of use of the physical disk storage whether internal or external to the EdgeWare Storage system is to be organized in
113. e Heartbeat configuration is there After reaching this point changes can be made to the HARs and failover rebuilt from step III on if desired 1 Disconnect all NAS clients and iSCSI initiators if any that are using the failover cluster as storage This may involve first closing down some user applications In iSCSI then delete the partition compare Formatting the iSCSI Virtual Drive in Microsoft Windows above in this manual then enter Initator and log off the Target and remove all traces of the volume compare How to Connect with a Microsoft iSCSI Initiator part C 2 If applicable perform a Configuration Backup see Maintenance gt Integrity above in this manual 3 Stop all high availability resources HARs that are running on the failover cluster whether NAS iSCSI or rsync For example to stop CIFS go to the Networking gt CIFS module main page on the Primary and select Stop Samba Servers Keep a list of all of these 4 Make sure a UI window is open for both the Primary and the Secondary Go to the Clustering gt IP Failover page on both UIs Page 160 10 11 On the Secondary red Resource Status Down hit Stop Heartbeat and await confirmation which may take twenty seconds or so that the heartbeat is off Important Do this on the Secondary first If you do it on the Primary first it will cause an immediate fail event On the Primary green Resource Status up as quickly as possible after ste
114. e default is 50 MB sec and is half the maximum bandwidth you can achieve on a Gigabit Ethernet see Figure 137 Page 167 Figure 137 Configure Mirror final page 12 Advanced Configuration link gives you a page with more settings for Mirroring Normally you do not need to configure these settings IV TRIGGER THE MIRROR 13 Click on Save You will be taken to a confirmation page see Figure 138 dd es Deg SS A A Manual LS Return to index Figure 138 Confirm your Remote Mirroring Page 168 14 Verify your configuration as shown in Figure 138 and click on Start mirroring to initiate the remote mirroring process The remote mirroring Sync process will make the data blocks between the two EdgeWare servers consistent Time to complete the synchronization depends on the network rate and storage capacity of the volumes This completes the setup and triggering of the remote mirror It leaves you with a new volume which is spread across two separate Storage Servers and has a name like dev drbd0 This new volume is still raw and it takes a while before it is in a completely consistent state see Figure 139 The next couple of subheads will show how to promote the new mirrored volume to a useful state as a NAS or iSCSI volume Data written to the resource primary volume will be immediately replicated to the secondary volume Once the initial Sync shown at 2 1 in Figure 139 is complete
115. e exported volumes connected directly to a Fibre Channel adapter or through a switch FC Target devices work very similar to iSCSI devices with the added advantage of speed Nominal transfer rate for FC host adapters is now 4Gbps while the best price point for iSCSI is still at 1Gbps That gap is not so significant for end to end connections where the server or the storage device is usually the slow link It does however begin to make a real difference as multiple switches pile up in a storage fabric The Fibre Channel FC Target protocol is very similar to iSCSI but it works over Fibre Channel hardware and HBA on the initiator host side Here the Storage Server behaves as the peripheral on the FC bus and some other system is the host It is important to distinguish this from the Fibre Channel Arrays Xyratex RAID whose interface is found in the RAID module above In that case the Storage Server is the host and the Fibre Channel Array is the peripheral It is even possible for both Fibre Channel Target and Fibre Channel Arrays to run at the same time in the same Storage Server using separate switch fabrics The FC Target protocol has four major characteristics that distinguish it from NAS 1 FC Target is a block level protocol which means that it cares nothing about file system or permission complexities that each NAS protocol must deal with In addition the blocks are always mapped linearly from the host s virtual device to the target
116. e should be empty Now hit Return to Index in that page and find yourself on the Software RAID page 7 Use the drop down list to select the type of software RAID to create and click on the create software RAID device of level button Shown in Figure 13 is the process of selecting RAID Linear Notice four old software RAID devices these are the mirrored system disks devices and they Should be left untouched EdgeWare Storage Configuration Raid Configuration y S f Provides configuration and maintenance RAID Sets f options for software and hardware RAID EdgeWare 3 0 0 1ce ki HELP Software RAID Pap Pap Pap RAIDI i Concatenated Linear Create Software RAID device of level H 4Return to index Figure 13 Selecting type of software RAID level to create Page 31 8 Now select the drive partitions to include on the RAID device from the text box labeled Partitions in RAID by pressing CTRL and mouse button as shown in Figure 14 Then click on create button Figure 14 Selecting disk partitions to include in RAID set 9 The software RAID device should now be created as shown in Figure 15 Click on the link for the software RAID device in this case dev md4 and it will take you to its properties page Figure 15 Software RAID Linear device created Page 32 dev md0 Concatenated Linear create software RAID device of level rt Figure 16 Software RAID5
117. eate and mount the volume as an iSCSI target and it will be listed on the iSCSI target page E d g e W are Server Storage Networking Authentication Clustering Log Out f m Maintenance M O functions like logical volume creation D d d h EdgeWare Storage Configuration Volume Management A Create Logical Volume d e A Gal Storage Ww EdgeWare 3 0 0 1ce Me ER etd iii x Provides logical storage management snapshot creatio DO RAID a In volume group VGstorage gome Manager KI Share Manager Logical volume details EES Volume name iSCSHLVstorage Volume size 250 6 kumm Storage allocated from volume group GB Dout of 279 Allocation unit size 4096 MB Network File Protocol CIFS AFP and or UNIX Mixed LU Sen giocoi CIFS AFP and or UNIX Mixed CS ses Dese r Sa Return to index Figure 48 Creating an iSCSI target Page 66 Reach the iSCSI targets page by selecting iSCSI Targets on the Storage gt iSCSI module page Figure 47 On this page you can view iSCSI targets defined on the system Each iSCSI target link takes you to the properties page for that target Figure 49 You can change some of the properties of the target after it has been created Alias A more human readable and understandable name for the target Max Connections in Advanced Configuration Maximum connection count allowed from initiators to this target or 0 if unlimited CHAP username and Password If you want to
118. eation run this time for a RAID1 It is what you get when the syncing is not done and you click on the link for the software RAID device Eeer Storage Co nfiguration Raid Provides configuration and maintenance EdgeWare Z GG H options for software and hardware RAID HELP RAID device options Device file aev man RAID level Mirrored RAID1 Status Active Usable size 10239872 blocks Percent of resync done 0 1 Persistent superblock Yes Stripe size Default Current status active resyncing Partitions in RAID SCSI device A partition 1 SCSI device B partition 1 EE Select a partition that is part of the RAID device and click this button 1 SCS device A pariton to remove it This may cause data to be lost Click this button to totally remove this RAID device Any data that it Delete RAID array VW contains will almost certainly be lost If this is shared storage you must delete this same array on the secondary also 4 Return to index Figure 18 Software RAID1 details 3WARE HARDWARE RAID Note The configuration for 3Ware Hardware RAID under EdgeWare 3 2 will only show for those Systems which have installed 3Ware hardware In systems with 3ware hardware controllers you can use hardware RAID to create redundant storage Older systems shipping out with hardware controllers had the Operating System installed on a hardware RAIDS set configured from the factory In this case you did not have the option
119. ed before it is fully synced compare Figure 139 then the primary is valid but the secondary is not If one half of a mirror has failed for any reason deleting the mirror allows the other half to be reconstituted as a standard non mirrored volume 1 Disconnect any NAS clients or iSCSI initiators running over the mirrored volume This may require pausing applications that use these clients or initiators 2 Shut down any services such as iSCSI or a NAS protocol running over the mirrored volume Page 174 E a ge wW are Server Storage Networking Authentication Maintenance User Clustering Log Out Mana EdgeWare Clustering Configuration Mirroring A Mirroring EdgeWare 3 0 0 1ce r Provides block device replication over the network to another server H HELP Lystor Follow22 LVstorage LYstorage mirror Mirror D is available Figure 147 Selecting a Mirrored Volume Go to the Clustering gt Remote Mirroring module main page see Figure 147 Click on the name of the mirror that you would like to delete This can be found under the Resource column under its local underlying volume name You will be taken to a Resource LVstorage page where your resource name takes the place of LVstorage Click on the Delete Mirror button compare Figure 139 Return to Storage gt Volume Manager and your underlying volume should be freed to be mounted if NAS or restored as an independent target if iSCSI
120. emed Covered Software upon release to Customer and Customer s access and use of Maintenance Releases will be subject to the same license terms as applicable for the Covered Software Software Updates are improvements bug fixes error corrections and patches that may include minor new features but not architectural changes or major new features which Manufacturer makes generally available to customers of a Support Package that includes Software Update Service Software Updates are typically labeled with a change in the version number after the first decimal e g 1 0 to 1 1 although not each change may be deemed a Software Update If Software Update Service is in effect under the Support Package selected by Customer as set forth in Exhibit A Manufacturer will make available Software Updates for the Covered software until the next Software Upgrade Software Updates will then be deemed Covered Software and Customer s access and Customer s use of the Software Updates will be subject to the same license terms as applicable for the Covered Software Software Upgrades are new software releases versions modules or feature sets that may include new features architectural changes and quality improvements that allow the Customer to upgrade from one version of Manufacturer s software to the next version e g 1 0 to 2 0 Software Upgrades will be made available to only those Customers who have separately purchased Software Upgrade
121. enters your folder A window with your new folder Address being a drive letter replaces the selection menu I Your share will also be visible in My Computer gt My Network Shares Your share is now permanent it and its data will survive reboots and power downs on both client and server When you reboot you will have to wait till the server is up and then log in again with user and password in My Computer gt My Network Shares It may be necessary to disconnect your network drive and map it again Page 98 NFS This networking section allows you to configure the NFS sharing options on the shares you have created from Share Manager NFS is the native Unix Linux file sharing protocol It is oriented toward trusted client systems hosts or IP addresses rather than user authentication In standard Unix Linux style it results in the server share becoming a branch of the client s universal file tree The server s side of this process is called exporting the share or the volume and the dente side is called mounting the share The NFS module main page Figure 77 shows a list of the shares that have been enabled for NFS sharing and a summary of the NFS export permissions Also you can start restart or stop the NFS server daemon from this page There are global selections NFS v4 better security and optimize NFS performance The global selections are usually left at their default but the start restart should be done after every change in t
122. er See Figure 117 Then hit Save so it saves these plus the defaults on everything else This returns you to the main page Figure 118 Now the primary node is identified and the Resource Status is Up green but Heartbeat is still down If you select Create Modify Resource you get a screen like Figure 117 except the Hosts in cluster is correctly identified as Leader22 left radio button The same screen shows up if you hit the link to the identifier under Primary Node EdgeWare Clustering Configuration Heartbeat if Heartbeat Monitor EdgeWare 3 0 0 1ce Provides configuration options for IP address and service failover HELP Primary Node IP Addresses Node Services Resource Status Get Resource Leader22 None services UP Heartbeat down Create ModifyResource To define a failover resource Authentication Keys To configure the type of authentication for heartbeat events Heartbeat status heartbeat is stopped No process Start Hearteat montor F Click this button to start the Heartbeat monitor process This will enable automatic service failover on your system with the current configuration Figure 118 Failover module main page after Save in Figure 117 Logged into the follower go to Create Modify Resource and you should find Hosts in cluster as the leader right radio button while Ethernet heartbeat interfaces is already selected as right radio button highlighting follower s own eth1 in list Figure 119
123. er This controls whether local logins are permitted or not If enabled normal user accounts in etc passwd may be used to log in Default Yes Write Enable This controls whether any FTP command which changes the filesystem is allowed or not Default Yes Recursive Enable When enabled this setting will allow the use of Is R This is a minor security risk because a Is R at the top level of a large site may consume a lot of resources Default Yes Maximum Concurrent Logins If vsftpd is in standalone mode this is the maximum number of clients which may be connected Any additional clients connecting will get an error message Default 0 unlimited Page 103 Ed g eWare Server Storage Networking Authentication Maintenance User Clustering Log Out Manual EdgeWare Network Configuration FTP server if Z e 7 N S F Upload amp Download files using the Fil kA Networking Options a Networking J EdgeWare 3 0 0 1ce E rey m HELP Do Networking Options Local User Yes No Write Enable Yes No as NFS Recursive Enable Yes No Maximum concurrent logins o m Anonymous User Yes No Anonymous Upload Yes No 1 Vi AppleTalk Anonymous Make Dir Yes No No Anonymous Password Yes No Idle Session Time Gog seconds Data Connection Timeout 420 beconds ts Time to wait for cn SEER Log Enable Yes No authentication D Http XferLog Enable Yes No Log File Narflog
124. erver Use server Neither Master Browser Elections Decline O Participate Participant lowest e PDC and ADS option Windows 2003 server PDC ADS ONo Yes PDC ADS server 192 168 1101 Administrator Administrator Administrator Password soosoo ADS option KDC 192 168 1 101 Realm LAB03 LOCAL Logging option Debug No O Yes Log File var log samba m log Figure 76 Windows Networking Options after entries d Click on Save and Restart This will save the settings and restart Samba not the server It should say Samba Status JOINED server TO REALM name name where server is replaced with all caps version of your hostname and name name is your realm name e g LABO3 LOCAL above 4 Check users a Return to CIFS module main page as by clicking on Return to Index b Click on Windows Domain Users and Groups in the center pane All users on the host server should be listed Page 97 SERVER CLIENT ACCESS TO CIFS This subhead gives a summary overview of the process of getting both sides of a CIFS server client relationship working This is for the default user level authentication but the others are similar A Connect server and client to the same local network carry out the instructions for Networking gt Network module above on the server and see that they ping each other The server side command is ping c 3 192 168 1 17 executed in Maintenance gt Command Line wait about 10
125. es and NAS Servers Index Target Configure and schedule the replication Target Trigger the replication mmo D It is easier to understand this design if you remember that the replication source is a server and the target is a dient in this operation Because of that all of the above steps are logically necessary A SOURCE SET UP THE SHARE WITH ASYNC ENABLED This is done in Storage gt Share Manager See that module s documentation above in this manual B SOURCE ACTIVATE THE REMOTE REPLICATION SERVICE ON THE SHARE This is done in Networking gt Remote Replication See that module s documentation above in this manual C TARGET SET UP A VOLUME FOR SOLE USE AS REPLICATION TARGET This is done in Storage gt Volume Manager See that module s documentation above in this manual Notice that this action on the target system is unrelated to the construction of the source share in step A above Nevertheless there are two points that ought to be noted here e The target volume must be big enough to accept a full copy of the file system found on the source share Unlike scheduled snapshots only one copy is made but it will continually be updated according to the replication schedule e The target volume s name should be suggestive of the source system and share D TARGET CAPTURE SOURCE IN HOST ADDRESSES AND NAS SERVERS INDEX This is done in Networking gt Network gt Advanced Configuration and Clustering gt N
126. es making changes as follows e Server type EdgeWareStorage e Link type Login via EdgeWare with username and password enter neighbor s values by default admin and setup do not use root as neighbor s username e Make fast RPC calls yes The rest of the broadcast settings default to correct values in the current version of EdgeWare Page 136 ES EdgeWare Clustering Configuration NAS Servers if i Edit Server O Contact managed servers EdgeWare 3 0 0 1ce HELP Hostname CE7E478A Port 10000 Server type EdgeWareStorage SSL server Yes O No Description From hostname and port Member of server groups New group Link type Normal link to server Login via EdgeWare with usemame admin password jeccce Login when icon is clicked on Make fast RPC calls Yes O No Server status Running EdgeWare 1 310 46ce Save Delete 4Return to index Figure 101 Edit Server window after editing When you have properly set up the neighbor server in NAS Servers Index it will appear with the EdgeWare icon like CE7E478A in Figure 100 and clicking on its hostname 10000 link will log you in automatically and place you in the top screen of the neighbor Figure 102 Close examination of Figure 102 shows that it is the neighbor CE7E478A at 192 168 1 31 and it refers to the home server 192 168 1 33 as its User IP Address This can be quite confusing om BO
127. es not mean delete the data on the share That data will be found intact when a new share is created on the same volume though specific information about its share configuration may have to be re entered CONSTRUCTIVE DEPLOYMENT OF A REMOTE SYNCHRONOUS MIRROR The constructive deployment of a remote synchronous mirror requires awareness of the state on both component Storage Servers Since the synchronous mirror is a block level device it joins raw volumes even if the data on the primary volume has already been created by a higher protocol such as NAS or iSCSI Once this is done the newly mirrored raw volume is an exact Page 162 virtual image of the primary volume that went into its making so all the superstructure can then be created or restored atop this new raw device This process can even be reversed deleting the mirror so that where once there was one mirrored copy there are now two identical copies The mirror itself for its own accounting requires some space at the end of the constituent volumes This means volume padding is necessary if a pre existing volume with user data is to be mirrored later and keep its data intact If volume padding is not done current EdgeWare default a backup and restore are required This is true for both NAS and iSCSI The steps below are easily seen to be logically required by the above description They are listed by Roman numerals because this is a super deployment whose steps e g Hands
128. ess is complete you will get a message Restore completed successfully on the page as shown in Figure 46 Note Figure 46 is from a tape backup not the Live Copy backup set up in Figures 42 through 45 Module Index Restore Backup Restoring backup with command star xp acl f dev st0 command star xp acl f dev st0 Volhdr beakup_0681004 star current newer star current iview391 exe newer Star current AdbeRdr60 enu full exe newer star current j2re 1 4 2 03 windouws i586 p exe newer star 3200 blocks O0 bytes total of 32768000 bytes 32000 00k Restore completed successfully da Retum to backups list Figure 46 Restore Backup result iSCSI is Internet SCSI Small Computer System Interface an IP based storage networking standard for linking data storage facilities By carrying SCSI commands over IP Internet Protocol networks iSCSI is used to facilitate data transfers over intranets and to manage storage over long distances The iSCSI protocol takes advantage of the fact that anything that uses a SCSI device knows only that the thing it is treating as a peripheral device is responding according to the SCSI protocol Instead of being hardware like a disk or tape with a SCSI driver it can be software called an iSCSI initiator that appears exactly like a SCSI peripheral But what the iSCSI initiator in fact does is to communicate over an IP connection with an iSCSI target on a differe
129. et a confirmation screen that says 3dmd process successfully started Hit Return to Index which returns you to the 3Ware Hardware RAID page then select a type of RAID from the drop down list and press Create Hardware RAID device of level gt In the case of RAID1 the following screen appears Create RAID Device EdgeWare Storage Configuration Raid Configuration S EdgeWare 3 0 0 1ce Controller 6 Storsave Policy protect v Ports to Use Figure 20 3ware RAID device creation page Normally the Stripe Size not present in Figure 20 because it is RAID1 need not be touched and the Storsave Policy is a tradeoff between speed and data stability in case of power failure Protect is the Page 35 most conservative Perform the fastest and Balance is intermediate The Optimize 3ware button on the summary page has a similar function You must also pick the ports using Ctrl or Shift to pick more than one shown selected After making your selections hit Create A confirmation screen appears Exiting the confirmation screen and selecting Logical Disk 0 leads to the details of Figure 21 If a 3ware RAID set encounters a failed drive it will resync to a healthy state by using a hot spare drive if one is specified If no hot spare was specified before a drive failed you will have to replace the faulty drive and then specify a rebuild of the RAID set When a RAID set is not healthy and h
130. etails to see information about the sessions connections and devices for that target Targets _ Sach Figure 55 iSCSI Initiator Properties Targets Click on Automatically restore this connection If you miss this step you have to do everything over again If you are not doing MPIO skip to step 8 If you are doing MPIO you must also click on Enable multi path then hit Advanced and type in Source IP to be the Windows side IP of this connection and Target Portal to be the EdgeWare side IP followed by 3260 the port number and also select CRC Checksum to be Data Digest and Header Digest After this click OK You have now connected your iSCSI initiator to the EdgeWare iSCSI target If you are not doing MPIO skip to step 11 If you are doing MPIO go back to step 3 and continue with your other connection s If you are doing MPIO when done with all connections go back to iSCSI Initator Properties Figure 55 select Targets tab select the target that has Connected status and click Details then click on the Devices tab select the first device click on the Advanced button click on the MPIO tab and change the Load Balance Policy to Round Robin Return to the top screen iSCSI Initiator Properties if necessary Select the Bound Volumes Devices tab If no volume shows hit Bind All Now your volume should show Hit OK
131. ever going to use MPIO be sure you check Microsoft MPIO Multipathing Support for iSCSI in the Installation Options section of the Installation Wizard 2 Locate the iSCSI initiator in your desktop and double click on the icon like the one below Figure 52 Initiator icon Page 69 3 Click on the Discovery tab and click on Add under the Target Portals section see Figure 53 below iSCSI Initiator Properties General Discovery Targets Persistent Targets Bound Volumes Devices Target Portals Address Port Adapter IP Addr ISNS Servers Name Figure 53 iSCSI initiator properties Discovery 4 Enter the iSCSI IP address of your EdgeWare system e g 192 168 2 218 and hit OK Figure 54 and you should be able to see the iSCSI target under the portals section z Add Target Portal Type the IP address or DNS name and socket number of the portal you want to add Click Advanced to select specific settings For the discovery session to the portal IP address or DNS name Port 192 168 2 218 3260 Figure 54 Add Target Portal 5 Click on the Targets tab and then click on the Log On icon and a pop up window will appear Page 70 10 11 iSCSI Initiator Properties General Discovery Targets Persistent Targets Bound Volumes Devices Select a target and click Log On to access the storage devices for that target Click d
132. fectively stops you from using that EdgeWare user account to create new files edit old files etc The soft limit is the desired maximum After exceeding the soft limit the system warns the user and starts a grace period When the grace period ends the system will not allow the user any more space EdgeWare Server Storage Networking Authentication Maintenance Clustering Log Out SEN EdgeWare Storage Configuration r Quota Management Disk Quotas Provides user storage usage quota management on the logical volumes Storage EdgeWare 3 0 0 1ce 4 c a VO Mume manager Volumes L Vstorage SGI LVM YG VGstorage LV User and Group Quotas Can only enable at users Filesystem LYstorage Active mount TE EET fr snare manager groups i C f E 1 Edit User Quotas Browse N Edit Group Quotas Browse nr Quotas g Backup m on Figure 38 Quotas page dd es E e AVE A Server Storage Networking Authentication Maintenance Clustering Log Out Manual EdgeWare Storage Configuration Quota Management Filesystem Quotas EdgeWare 3 0 0 1ce Provides user storage usage quota management on the logical volumes Storage g S e All User Quotas on volumes LYstorage i Seal Manager ef uit Quota For Browse W EwtGraceTimes Check Quotas gt Blocks 262016000 total 252015828 free 12000 granted Quotas User Used Sep Ump Hard Limit Used SoftLimit
133. figure Remote Mirroring is by setting up a logical volume on each server that is going to be used in the mirror If you are starting fresh make each volume the identical size If your primary volume already has data make sure the secondary volume is forced to be exactly the same size as the primary or if this is impossible larger than the primary Mote Remote Mirroring works best with two volumes of identical size Figure 130 shows the Modify Partitions information that is good to view before being sure that identical size has been achieved The disk model and the start and end cylinders are instructive If there is no discrepancy go to step 3 dad am J yA A Storage Networking Authentication Maintenance Clustering Log Out User Manual d EdgeWare Storage Configuration Partition Management if i Partition Manager Storage p yY EdgeWare 3 0 0 1ce d The Partition Manager modifies the O partition file system ALL DATA WILL BE LOST B RAID hit SSS SS Location IDE device A No Type Extent Start End Use Free Volume Manager Cylinders 2036 1 Linux d 40 boot 53 kid Size 1003 MB 2 Linux 4 2036 7 12 El snore manager Model M Systems DOC V1 12 Add primary partition Add extended partition Edit IDE parameters oS Quotas Location SCSI device A No Type Extent Start End Use Free Ct Cylinders 19929 A Linux O 19929 a Backup Size 153 GB Add primary partition Add extended partition Model ATA Maxtor 6Y160M
134. fy the first principle of backup design can include Page 59 e Use different media for instance magnetic tape backup of disk based storage and or offsite storage e Use a physically separate storage system see Asynchronous Replication in Network and Clustering sections below e Use different disks on the same system the backup to another logical volume Notice that in this case it should not only be another logical volume but also it should be on another volume group that is built from a different set of physical volumes Because software and user errors are at least as common as hardware failures it is important especially in the last option to make sure that the normal uses of your current data especially uses that can write or delete that data do not at the same time do the same thing to your backup data The backups should have different looking names and not be available for global operations EdgeWare s volume group based tree structure is your friend in this requirement but you must also set up your normal procedures to avoid inadvertent backup destruction In addition your procedures need to ensure the second principle of backup design See to it that the backups actually get made on a regular basis Once again EdgeWare does its part with its scheduled backup capability but there are other parts of the design that only you can contribute These include physically enabling backup flow with media shelf space tran
135. g drive other than the failed drive is removed data loss will result Typically drives are counted left to right then top to bottom but this is not always true Best is to record the hard drive number outside its slot in advance If there is any doubt temporarily unmount the degraded volume logical array or make it read only before experimenting with drive removal Replace the failed drive and rebuild as soon as possible A RAID 5 runs without protection while it is in degraded state Allowing this state to continue puts you in danger of losing another drive causing data loss If Auto Rebuild Hot Spare is underway wait for it to finish and then replace the failed drive with a new drive and set that new drive up as the new Hot Spare If there is no Hot Spare follow What to do first when a RAID set is degraded above except replace the failed drive with a new drive before step 7 NEVER Remove more than one hard drive at a time Page 38 In RAID 5 data protection modes your Cutting Edge Storage Server device can continue operating after a single drive failure per RAID 5 set Do NOT remove any other drives from a RAID set while there is a failed drive in the system or while the replacement drive is being restored Auto Rebuild of Redundant Array If a hot spare is specified and the array degrades an event notification is generated and the hot spare dynamically replaces the failed drive in a redundant array without user inter
136. ge even if the practical maximum barrier of item b above is not reached This means that rapid IO right after bootup or after a period of quiet can suddenly slow down after cache gets gobbled up Observe snapshot behavior over long periods of time days before concluding that it is satisfactory For a good snapshot design built around these facts you should take care to do the following e Predict behavior before committing to storage allocation and snapshot scheduling The numbers and formulas above will aid you together with your own installation s information about peak data creation and change rates and IOPS Cutting Edge personnel can aid with these predictions e Precisely define your snapshot needs and discard unneeded or aging snapshots immediately This is partially automated through EdgeWare s snapshot schedule and Maximum Concurrent Snapshots Lifetime Job Schedule Maximum Concurrent Snapshots However an administrator should be tasked with checking and cleaning up snapshots at least weekly e Progress the data offline so as to minimize needed snapshot lifetime See Designing a Backup System below in this manual and make use of the efficient EdgeWare capability of Asynchronous Replication CREATING A SNAPSHOT OF A LOGICAL VOLUME od Capea Storage Confiqurstes V iemg Manaqement eener tee mme Create Snapshot Volume prparenread eae Ww Edge ae 188 ien Leclerc wen In volume group VGstorage Veluma aa
137. get This subhead will cover how to export a mirrored volume as an iSCSI target V FORMAT OR PROMOTE MIRROR VOLUME TO ISCSI STATE To export a mirror volume as an iSCSI target you must follow these steps 1 Go to the Storage gt iSCSI module main page 2 Click on the iSCSI target icon and select the Create a new target option 3 Select Mirror 0 as the local device Figure 145 Now click on Save Select the Logical Volume Local Device Mirror0 v Mi 0 LUN 0 Save Return to index Figure 145 Select the Logical Volume 4 You will be taken to a confirmation page that the mirror volume has been successfully exported as an iSCSI target Figure 146 EdgeWare Storage Configuration iSCSI Configuration O Provides configuration for iSCSI target and A GS EdgeWare 3 0 0 1ce initiator HELP Successfully modified the iSCSI target ign 2008 8 leader22 edgeware drbd0 Returnto index Figure 146 Confirmation page for iSCSI target 5 Once the iSCSI target has been mounted you can start iSCSI services To map the iSCSI target as a local drive please refer to the How to Connect with a Microsoft iSCSI Initiator and Formatting the iSCSI Virtual Drive in Microsoft Windows subheads above in this Manual Page 173 SWAPPING AND DELETING REMOTE MIRRORING SYNC Changing status of a mirrored volume once you have it requires that you remember that mirroring is a block based operation
138. gure 135 we pick both ends of the point to point connection Select the source and destination mirror volumes These are the volumes that were created in step II above Select them even if there is only one in the drop down list Then hit Next Proceed through two pages of warnings following the instructions Figure 136 shows the second warning page Without volume padding NAS or iSCSI data must be restored from an external backup after mirroring EdgeWare Clustering Configuration r K Mirroring Provides block device replication over the A Mirror resources J network to another server ez EdgeWare 3 0 0 1ce d HELP Remote Mirroring will cause data loss on the logical volumes of both Primary and Secondary The current version of EdgeWare CANNOT SAVE customer data found on the Primary If you have any data on these logical volumes be sure to back up all of them before you continue v I read the warnings agree to configure Remote Mirroring am sure want to configure remote mirroring Please type in I am sure want to configure remote mirroring to continue Are you sure you want to configure remote mirroring YES NOT Returnto index Figure 136 Configure Mirror fourth page Select the synchronization speed see Figure 137 You can choose to give an upper limit to the synchronization rate A slower rate will take longer for synchronization to finish but will impose less burden on the network Th
139. h as partitions or dependent software RAID arrays still remains on the Fibre Channel Array in question If that is a problem remove the superstructure and try again to delete Page 42 VOLUME MANAGER MODULE Volume Manager is the heart of the storage management on the EdgeWare Storage Server It provides virtualization of physical storage Using the Volume Manager you can create and aggregate storage volume groups add physical storage devices to these volume groups create logical volumes and use them for such purposes as file systems iSCSI targets and replication or mirroring targets and create snapshot volumes for the logical volumes USING THE VOLUME MANAGER FOR VIRTUALIZATION The structure and terminology used here are standard for the Logical Volume Manager a worldwide cross platform virtualization standard that is stable and robust You can examine it by opening your EdgeWare GUI to Storage gt Volume Manager and moving around the tree found on that module page Outermost branch is the volume group which is a big sequence of block addressable storage possibly many terabytes It is simply understood as room for a line of data but it can grow and shrink because it is virtualized or strung together from a number of actual physical data storage devices The volume group is subdivided in two ways which are independent of orthogonal to one another They could be thought of as ancestors and descendants The ancestors are physic
140. hake can themselves have complex deployments described above in this manual I Set up the Handshake II Create or demote volumes to raw state III Configure the mirror IV Trigger the mirror V Format or promote mirror volume to NAS or iSCSI state SETTING UP A REMOTE MIRRORING PAIR This subhead will cover steps I through IV of the constructive deployment Step II has several branches which are easily understood when one realizes that the goal is to get the volumes into the proper raw state without in the primary case destroying any data that may already exist on the primary I SET UP THE HANDSHAKE Identify the two EdgeWare Storage Servers that will act as the primary and secondary in this mirroring arrangement Make sure they are reasonably close in hardware and software version and that they have storage that suffices to support the two mirrored volumes If high bandwidth point to point communication between them will be required place them in physical proximity with one another Then carry out the steps required to set up the Handshake described above in this manual II CREATE OR DEMOTE VOLUMES TO RAW STATE If you intend to mirror an existing volume always make sure its data is backed up before starting this procedure Although mirroring need not damage data on the primary there are volume padding and formatting steps required where small errors can make it necessary to do a restore Page 163 The first step to con
141. have a DHCP connection or a non DHCP LAN connection a similar process should take place except there may be many icons to choose from after the discovery If the EdgeWare Server does not appear you can set an appropriate client IP address in the non DHCP case as follows Select the Windows icon in the lower left corner Select Network Select Network and Sharing Center along the top Select Manage Network Connections on left list Right click the identifier for the network connection leading to the Server Select Properties Pick Internet Protocol 4 not 6 Select Properties For DHCP leave the radio button on Obtain an IP address automatically and exit without making changes For non DHCP static address specification move the radio button to Use the following IP address and fill in appropriate values for the IP address and netmask If a part 0 to 255 of the quad is less than 3 digits right arrow moves you to the next part In the Auto case the right IP address value is 169 254 x y where in place of x and y you enter numbers between 1 and 254 and the right netmask value is 255 255 0 0 Do not fill in gateway TROUBLESHOOTING A FAILED SERVER IP ADDRESS CREATION The need for troubleshooting on this level can become apparent only after step C below is completed and the troubleshooting of the discovery tool described below under step C has been tried without result This means the problem is not with the discovery too
142. he Users and Groups module documentation above in this manual compare Figure 83 above In Network section set up CIFS or other protocols to export the mirror share Now you need to map the storage to your local workstation The description that follows is for Windows XP other client operating systems are similar In your Windows workstation follow Start gt My Network Places gt View Workgroup Computers select your computer and if necessary log in with a user authenticated above You should see your share as in Figure 143 Right click your share and select Map Network Drive Assign the letter that you would like to use as a network drive decide whether to reconnect at logon and hit Finish see Figure 144 Your mirrored share is now connected Map Network Drive Windows can help you connect to a shared network Folder and assign a drive letter to the connection so that you can gage access the folder using My Computer eg ee Specify the drive letter For the connection and the Folder that you want to connect to Drive Y v Folder Leader22 testshare joo Example Vservertshare Reconnect at logon Connect using a different user name Sign up For online storage or connect to a network server Cancel Figure 144 Map Network Drive on Mirror0 Page 172 EXPORTING THE MIRROR AS AN ISCSI TARGET After remote mirroring starts syncing you can take that mirrored volume and export it as share or as an iSCSI Tar
143. he default permissions that you want to assign to the valid groups on the share R read W write and X execute Optionally you can specify these permissions to be propagated to all the files and subdirectories or all the subdirectories under the share by checking the relevant check box under Apply ACLs recursively to all files and subdirectories Click on save to apply the changes Deleting users and groups from a share To delete users and groups from the share use the backspace on your keyboard to delete the users and groups you want to remove and then click on save to apply the changes Page 112 NIS The NIS module gives you the configuration and management options for integrating your EdgeWare Storage system into your NIS domain as a NIS dient NIS also called yp is a simple protocol for sharing authentication around a network and recording users and passwords on a single server It is friendly to NFS It is recommended for use among trusted clients in a secure local network such as a science lab as its password security is not very robust A NIS server must be on the network and have a NIS domain name before this module can be used NIS client configuration NIS domain cuttingedge com NIS servers Find by broadcast Listed below Authentication Order Local authentication first CO NIS authentication first NIS Client Status NIS client is bound to NIS server 192 168 1 254 1 Save and Rest
144. he settings Fee BOSR SS ZS A User ver Storage Networking Authentication Maintenance Clusterin er Manual EdgeWare Network Configuration NFS ag NFS Exports r A Provides settings to enable NAS to EdgeWare 3 0 0 1ce communicate with devices running the Network File System Networking Network HELP Share Name Directory sharetest shares sharetest di aa AppleTalk Enable NFS v4 on this server _ Global Configurations and Options Optimize NFS Performance This will take effect after restart or start NFS exports Sle E DE RES Start r Click this button to start the kernel NFS support and export the directories listed Figure 77 NFS main page When you click the hyperlink for an NFS share you are presented with a page Figure 78 where you can edit the various security permissions and configuration options for NFS sharing on that particular share Page 99 4 ea VYE A E d g e W are Server Storage Networking Authentication Maintenance Clustering Log Out GE EdgeWare Network Configuration NFS TEE aren j Edit Export O communicate with devices running the Networking EdgeWare 3 0 0 1ce l Network File System l eer HELP EEE cifs Directory to export aS NFS Active Yes O No F Export to Everyone O Host s AppleTalk w ee WebNFS clients Netgroup J FTP Network Netmask
145. heartbeat events Heartbeat status heartbeat is stopped No process Start Heartbeat Monitor F Click this button to start the Heartbeat monitor process This will enable automatic service failover on your system with the current configuration Figure 116 Clustering gt Failover module main page before setup On the leader go to Clustering gt Failover module main page Figure 116 Page 151 EdgeWare Clustering Configuration Heartbeat if EE Create Modify Resource f A address and service failover EdgeWare 3 0 0 1ce H Heartbeat configuration options Hosts in cluster Leader22 Follow22 o ethO 192 168 1 130 Ethernet heartbeat interfaces O None ENAME Shared floating IP address Failover Resources Shared Storage Block Devices IP address for ping checks Save emm Advanced Configuration Delete F Return to index Figure 117 Create Modify Resource values on leader Go to Create Modify Resource which displays the heading Heartbeat configuration options There fill in the top two lines Hosts in cluster which currently is other host should be selected as the follower other host its LAN hostname not its eth1 hostname which will be on the drop down list if Handshake was done correctly i e it is Edgeware and can talk via fastRPC and a login And set Ethernet heartbeat interfaces as leader s own eth1 with right radio button and highlight correct list memb
146. his case Leader22 will really be Follow22 and the IP address will still be that of Shared22 In addition CIFS clients may be confused about share identity and may put up two icons of which only one will work In the case of NFS and other IP address based protocols the shared IP address works without any such confusion and continues to work across a fail event In the case of iSCSI activation proceeds the same as iSCSI activation without failover on both the target and the initiator side see How to Connect with a Microsoft iSCSI Initiator above in this manual The only change is the shared IP address on the initiator side step C4 of that subhead This will continue to work across a fail event Note configuration changes cannot currently be transmitted once failover has started This means if you for instance add more shares or users to a NAS volume after failover is underway and then suffer a fail event your recovered system will revert to the state remembered from the day failover was first triggered To get around this do two things e Make sure when you carry out step II that everything is included in your resource setup that will be vital to continue in operation in case of a fail event e If changes must be made after failover starts use Configuration Backup see module Maintenance gt Integrity above in this manual to capture the new configurations You can then do a Configuration Restore on the recovered
147. hot Volume One snapshot means two writes ten snapshots mean eleven writes This affects even the random IO applications like databases and mail servers In EdgeWare tests September 2008 database IOPS began to suffer significantly after about four snapshots Note that snapshots do not multiply reads which are little affected even by many snapshots Thus an approximate rule of thumb is IJOPS with snapshots IOPS without snapshots 1 proportion of writes number of snapshots For example if 20 of IOs are writes then five snapshots will approximately cut IOPS in half e Snapshots heavily impact resources including storage and memory This is true even if you hardly ever write This has several consequences Page 50 a You must plan ahead for snapshot storage usage The formula is Minimum dedicated snapshot storage 2 Rate of file growth and change Lifetime of snapshot Number of snapshots The 2 is a safety margin you may want more Each snapshot should get 2 Rate of file growth and change Lifetime of snapshot as its volume size when it is being created b Memory requirements enforce a practical maximum on the number of snapshots EdgeWare tests September 2008 indicated an absolute limit of 15 snapshots Software changes currently under development are expected to increase this absolute limit to 50 100 but that does not imply that a good design should go that far c Snapshot disk IO multiplies cache usa
148. hot swap components as may be reasonably requested by Manufacturer DATA Customer will be responsible for maintaining back up copies of all data residing in or relying in any way on the Covered Products Manufacturer will have no liability or responsibility for the loss of any data during the provision of the Services ON SITE SERVICE Manufacturer will determine in consultation with Customer whether it is necessary to perform Services at the Customer s facilities On site service will be performed during normal business hours If Manufacturer decides in its sole discretion that it is appropriate to perform the Services at a Customer facility ON SITE ACCESS Customer will provide Manufacturer services representatives with a full free and safe onsite access to the Covered Product and b a secure storage space designated work area and access to a telephone a backup copy of current software and data and reasonable use of necessary equipment and communications facilities as may be required to troubleshoot and maintain the Covered Products Customer will indemnify defend and hold harmless Manufacturer for any liabilities claims costs or expenses including but not limited to attorneys fees arising from or relating to injury to Manufacturer personnel agents or equipment during provision of the Services at a Customer facility SERVICE MATERIALS Manufacturer service representatives may use and siore at Customer s facility software
149. ibed in the Handshake subhead above II SET UP AND ACTIVATE BUT DO NOT CONNECT THE HARS On the leader make sure the RAID arrays and volumes are set up on the shared storage Viewing them through the follower GUI will then show the RAID arrays but not the volumes However a reboot of both heads after setting up the volumes will cause them to be visible on both leader and follower UIs You will probably have to reboot both heads after making any changes in the Storage gt RAID module After the volumes are set up configure NAS and or iSCSI as it is desired to appear in the failover system Note configuration changes cannot currently be transmitted once failover has started It is therefore important to get resources set up right otherwise the whole failover setup will have to be undone and redone Each resource be it a NAS protocol or iSCSI must be completely set up The best way of assuring this may be to start it and then check with a client or an initiator that it is running If you do this however you should make sure the client or initiator connection is shut down afterward Leave the resource running on y on the Primary If you use this approach in CIFS Page 150 Samba be prepared for some confusion on the part of Windows clients which will go on thinking their shares are connected to the old IP address after failover shares are available On the Secondary if necessary go into Storage gt iSCSI and remove the iSCSI
150. ic IP addresses you may wish to enter a default gateway Default route device Default route device in the system This should be none when interfaces are DCHP assigned If interfaces have static IP addresses you may enter the default interface for the network packets to be routed to If left blank this would be eth0 A ways enter a non ethi0 value here if that non ethO connection has a hardware route through eth0 such as a high speed switch accessible from the LAN Otherwise the configuration may insist on routing through eth0 and ignore the efficient connection DNS servers The DNS servers that the appliance will use to resolve domain names If your DHCP is set to assign DNS servers in the DHCP lease these fields will be automatically populated If not then you should enter DNS server IP addresses for your network Search domains Enter the domains you want the appliance to search for when resolving unqualified host names For example if you add sd cuttedge com and test cuttedge com to the list and then type a command ping ce 3 xyz the appliance tries to ping xyz sd cuttedge com and xyz test cuttedge com ADVANCED CONFIGURATION The hyperlink at the bottom of the Network page gives you advanced configuration options such as DNS Routing and defining Host addresses See Figure 72 DNS Configuration The DNS configuration page allows you to store settings for the Name Service Hostname The NetBIOS name of the appliance Page 90 Re
151. ical volumes When using the Logical Volume Management through the GUI this abstraction above the actual physical storage creates a file or block based storage system that is able to accommodate changing size and other requirements Page 43 dd a e 2 BD amp A Storage Networking Authentication Maintenance Clustering Log Out reel Manual em EdgeWare Storage Configuration Volume Management S Create Volume Group Storage EdgeWare 3 0 0 1ce a Provides logical storage management functions like logical volume creation physical volume aggregation and snapshot creation etc Se E Volume Manager KI Share Manager Volume group name VGstorage Allocation blocksize Default ke IG os Initial physical device TO Browse Or Warning All data on the selected device will be erased Oe on Return to index Figure 26 Creating a new Volume Group CREATING A NEW VOLUME GROUP To create a new Volume Group click on Add a New Volume Group at the top of the page You will be presented with Volume Group creation page as shown in Figure 26 1 Type in a name e g VGstorage for the Volume Group 2 Leave the Allocation block size default unless you need a specific block size for performance 3 Select an initial physical device from the list you can add more devices to the Volume Group later If you just created the physical device especially if it is a RAIDS or RAID6 wait sever
152. iguration Heartbeat Provides configuration options for IP Le Heartbeat Monitor address and service failover EdgeWare 3 0 0 1ce HELP Primary Node IP Addresses Node Services Resource Status Get Resource Leader22 192 168 1 158 hb_mounts services Create ModifyResource To define a failover resource Authentication Keys To configure the type of authentication for heartbeat events Heartbeat status heartbeat OK pid 8038 et al is running on follow22 follow22 Apply Changes Click this button to force the running Heartbeat monitor process to reload the current configuration Stop Heartbeat Click this button to stop Heartbeat monitor process Figure 126 Follower s response to leader s Get Resource selection passive VII ACTIVATE THE SHARED HARS Using the new shared IP address the HARs that were configured in step II can now be activated In each case this takes two steps e Trigger the HAR within the leader s module e Carry out actions in the client NAS or initiator iSCSI systems that complete the connection to the resource through the shared IP address or hostname In the case of CIFS Samba some confusion may be introduced because the hostname is hard coded to be the original resource name e g Leader22 in the case just described not Shared22 Page 156 Windows discovery will connect it to the correct IP connection however and failover will work correctly if the leader fails In t
153. ing the Server link Server with the closed folder icon next to it expands the Server section After expansion the Server section opens up the folder icon and lists the modules affiliated with this section The links are an indented list of module names as hyperlinks which can be selected using the mouse STORAGE SECTION Enclosed in the Storage section are the modules RAID Software and Hardware Volume Manager Share Manager Quotas Backup disk and tape and iSCSI NETWORKING SECTION The Networking section includes the modules for the IP address Network share management for Remote Replication and the file sharing protocols CIFS NFS AppleTalk FTP and Http These are protocols for which EdgeWare Storage Server acts as a server AUTHENTICATION SECTION The Authentication section includes the modules for Users and Groups Share Permissions NIS and LDAP NIS and LDAP are protocols for which EdgeWare Storage Server acts as a client MAINTENANCE SECTION The Maintenance section includes the modules Shutdown and reboot Monitoring Logs lists of running Processes Command Line access OS Integrity and software Updates Integrity also includes the important capability of configuration backup CLUSTERING SECTION The Clustering section includes the modules NAS Servers Index Remote Replication Asynchronous share mirroring IP Failover and Remote Mirroring Synchronous volume mirroring The last three are co
154. install or replace Field Replaceable Units and the units installation or replacement will be in accordance with handling procedures specified by Manufacturer CUSTOMER RESPONSIBILITIES Manufacturer s obligation to perform the Services is conditioned on the following MAINTENANCE Customer will a maintain the installation site in accordance with the applicable specifications for the Covered Products b maintain the Covered Products and the operating environment for the Covered Products in good working order and in accordance with the specifications in the documentation for the Covered Products c use the Covered Products in a proper manner by competent trained personnel and d implement all Software Updates Maintenance Releases and Error corrections and workarounds in a timely manner to be at the most current release of the product or at most one release back REMOTE ACCESS At Manufacturer s request Customer will allow Manufacturer remote access to the Covered Products to enable Manufacturer to perform remote diagnosis and service NOTIFICATION OF ERRORS Customer will notify Manufacturer promptly regarding Errors with reasonable detail so that Manufacturer can reproduce the Error Customer will provide a complete description of the Error including if applicable a description of any physical indicator of injury such as a LED alarm or event s on the Covered Cutting Edge 1825 Gillespie Way Suite 100 El Cajon CA 9202
155. installed from the CD User Authentication LDAP radio button was selected and Make me an LDAP client at localhost Thus from the SuSE system s point of view both server and client were itself Four more packages then installed from CD pam ldap and nss_ldap variants Finally the new LDAP user was installed ldickson idap After the client setup described below it shows up in Figure 87 from the EdgeWare Storage system s point of view proving that the LDAP server client relationship has been established successfully Page 114 LDAP Client Settings Base domain name dc site Port default 389 389 SSL TLS Settings Verify server certificate O Yes No Select CA certificate Server requires certificate Yes No Select certificate Select key User Settings LDAP Bind Name cn Administrator v Use SSL Use TLS Anonymous bind dc site v Root DN M Use LDAP authentication Windows Client Specific Settings User suffix Authenticate Samba with LDAP Group suffix Machine suffix Save Figure 86 LDAP client LDAP Client Samba v2 2 compatible window with settings In order for LDAP authentication to work the administrator must enter the LDAP client settings to configure access to the LDAP server over the network Figure 86 The following information needs to be provided in the setup page LDAP Client settings
156. ity shall be responsible for any loss or damage to equipment being returned by the ordering activity for repair to the Manufacturer s authorized service facility Manufacturer shall only be responsible for any loss or damage while the equipment is at the Manufacturer s authorized service facility and until it is returned to the ordering activity s location 2 Technical Services Phone Email and Web support is available during the warranty period during Standard Business hours M F 8AM 5PM Pacific Time Technical support is available via 800 257 1666 support cuttedge com http www cuttedge com support php All other Technical Support services from the Manufacturer are available per the charges set forth in the Pricelist Section Additional technical services may be available outside the scope of this contract from the Manufacturer or Authorized Government Resellers Manufacturer s Maintenance Information 1 Service Areas a The annual maintenance fees as listed in the price list are applicable to any ordering activity location within the geographic scope of this contract b If maintenance repair services cannot be performed at the ordering activity installation site the repair services will be performed at the Manufacturer s plant s listed below or any alternate site designated by Manufacturer 1825 Gillespie Way Suite 100 El Cajon CA 92020 Cutting Edge 1825 Gillespie Way Suite 100 El Cajon CA 92020 800
157. k 1 Basic New Volume D 232 00 GB 232 00 GB NTFS Online Healthy SCD ROM 0 DVD F No Media N Primary partition Figure 57 Computer Management New Volume 7 To verify that your iSCSI target has been installed properly go to your My Computer icon and open it You should be able to see the newly created iSCSI target Figure 58 Other versions of Windows may have different paths to the necessary disk allocation and formatting steps However the appearance of the pages when found is generally similar to the above screen shots For further guidance consult and show these screen shots to your network administrator or Windows system expert Page 73 My Computer File Edit View Favorites Tools Help Or EE Bus E System Tasks L view system information Se Local Disk C Kei Add or remove programs D Change a setting Devices with Removable Storage Other Places sl My Network Places O My Documents D Control Panel en EE ANStor64 D D 4 Se Se CEdatabase on Cec U SS Cec CEshare Cuting Edge V p Cutting Edge on Cec CEshare pr Figure 58 Shares showing up under My Computer CONNECTING A SOLARIS ISCSI INITIATOR IN STATIC MODE TO AN EDGEWARE ISCSI TARGET 1 Log into Solaris server console or ssh 2 su to root 3 Disable all targets using commands a iscsiadm modify discovery t disable b iscsiadm modify discovery s disable C
158. l but is because no IP address has been given to the EdgeWare Server that is accessible to the client system In this case after exhausting all troubleshooting options of step C below consult with your network administrator Ask whether DHCP is active or not for the EdgeWare Server s connection In either case your network administrator may have techniques for determining the IP address of the EdgeWare Server If the problem is not solved by your network administrator please contact Cutting Edge Technical Support They will guide you through a step by step procedure for physically accessing the EdgeWare Server and setting a valid IP address C FINDING THE EDGEWARE SERVER FROM YOUR CLIENT STORAGE SERVER DISCOVERY Successful completion of step B above leaves you with a known subnet whether DHCP or Auto and a known client IP address but the IP address of the EdgeWare Server though in existence is unknown Since the IP address is needed to communicate with the server this is a chicken and egg problem that results in many delicately designed pieces of software called discovery tools Guidance in the use of these is given below FIND THE SERVER BY USING ANOTHER EDGEWARE SERVER The easiest solution of your problem is if there is another EdgeWare Server previously installed on your network In this case log in to that server from your client and go to NAS Servers Index under Page 13 Clustering Follow the instructions fo
159. lect days of the week Hours to check Here you can have the service monitored every hour or specify a time period in a 24 hr clock representation Maximum alerts to send The number of maximum alerts to send for one failure event The default value is Unlimited Click on Save to modify or create monitoring Remember to Restart Monitoring Services Deleting a monitoring period To delete a monitoring period check the box marked Delete this period and then click on Save at the bottom Remember to Restart Monitoring Services SNMP SNMP Simple Network Management Protocol typically monitors network devices testing packet errors bandwidth device status etc This page configures the SNMP monitoring service on the EdgeWare system which from the SNMP point of view is a peripheral device We begin with assigning the right SNMP version then proceed to name settings physical location and trap sink client not EdgeWare Storage Server IP address es Save the page and start the SNMP server to begin to create valid data in a MIB Management Information Base See Figure 91 In addition a MIB browser elsewhere on the network can then access EdgeWare system details ER es De g Be A E d H eWa re Server Storage Networking Authentication M Log Out Lae jaintenance Clustering Poari EdgeWare Maintenance Configuration Monitoring Da sir Maintenance EdgeWare 3 0 0 1ce e O Setup and manage warnings and errors
160. lete guidance Page 107 Only shares that have Async protocol enabled from Share Manager are listed Figure 81 The Global Replication settings link is for advanced tuning options and should usually be left as is __ EdgeWare Networking Configuration Remote r el e Replication O Provides management options for the D Remote Replication H asynchronous mirroring service EdgeWare 3 0 0 1ce d HELP Locally Available Shares Exported Share Name Path Comment sharetest shares sharetest test Global Replication settings stat Click this button to start the replication support Figure 81 Network gt Remote Replication module When you click on a share you can specify optional settings Share Description A brief description for the async share Hosts Allow Allow these hosts to asynchronously replicate this share You can enter an IP address or hostname each separated by a space If entering hostnames make sure you have hostname resolutions working in your network Hosts Deny Prevent these hosts from asynchronously replicating this share You can enter an IP address or hostname each separated by a space If entering hostnames make sure you have hostname resolutions working in your network Delete This does not delete the share It deletes the Async authorization from the share and its effect is the same as unselecting Async in the Storage gt Share Manager gt sharename edit window Click on Save Ch
161. ll client users including root will be treated the same by the server This is convenient for small networks used for scientific or testing purposes Everyone except root The root user on the client is treated as the untrusted user by the server but all other users are trusted This defends against certain root hacks Nobody All users on the client are treated as the untrusted user This option is best used when exporting to a system that you do not control such as a single user PC or when exporting to everyone Disable POSIX ACLs This enables or disables the POSIX access control list support for the export directory Yes default means disable ACLs this default improves performance Don t trust UIDs Map requests from UID specified here to the anonymous UID Note that this does not apply to any other UID that might be equally sensitive such as user bin Don t trust GIDs Map requests from GID specified here to the anonymous GID Note that this does not apply to any other GID that might be equally sensitive Treat untrusted users as This option determines which local user untrusted client users are treated as You may enter either a UID or select a user or choose the default typically the user nobody You can click the button next text field to get a pop up window that displays local users Page 101 defined on the system that you can provide as the anonymous user to which untrusted users are mapped to Treat untrusted group
162. loba setup of your EdgeWare NAS within your Windows network Much of what you do here makes sense only if you have users authenticated and connected to a given share as described above in the subhead Edit Security Share Name Select Page Security This field lets you set the security level of Windows networking The options are _ Security share level Caution Share level security is mostly for very old legacy Windows clients and low security guest access It is not recommended for current technology networks In share level security the client must authenticate itself separately for each share it wants to connect Client will send at least a password with each demand of connection to a share Client won t send any username with this operation because it is expecting that the server has a password associated with each share EdgeWare CIFS component Samba uses the UNIX authentication where a pair username password is authenticated and not a pair share password So Samba has to find out what username is associated with the password the client sent Then Samba has to give the client the permissions this user has for this share Page 93 EdgeWare Network Configuration CIFS P A Provides settings to enable NAS to Windows Networking Options communicate with Microsoft networks EdgeWare 3 0 0 1ce d ke HELP Windows Networking Options Security User level Local v Workgroup Domain Defaut
163. lover Bonding and MPIO MCS are not supported by FC targets in EdgeWare If these capabilities are needed please consider using iSCSI Do not confuse a Fibre Channel target with a Fibre Channel Array covered in this manual under RAID Configuring FC Target Mode support in Edgeware involves the following steps 1 Enable an adapter for Target Mode operation 2 Define Logical Volume and prepare it for export 3 Optionally define access rights via Security Groups To access FC Target Mode configuration click on the Storage icon then select the FC Target icon see Figure 59 ad es Deg KAREL A EdgeWare Server Storage Networking Authentication Clustering Log Out Laat Maintenance Manual e EdgeWare Storage Configuration FC Target Mode r x N Configuration O Provides configuration for Fibre Channel DA Fibre Channel Target Mode Configuration Target Mode Support Storage d EdgeWare 3 0 0 1ce RAID HELP an Fibre Channel Adapters enabled for Target Mode KI wg Target Mode Adapters Adapter ID Device ID State Action isp3 7 b01 Disabled Enable oS Quotas Em isp2 6 boD Disabled Enable UI Bacup isp1 5 301 Disabled Enable lt 2 ispo 3 300 Disabled Enable Sr GF FC Target Exported Fibre Channel Target Mode Devices Alias Size MB Block Size Options File Name Remove from Export List Add Exported Device LUNs and Security Groups Security Group Edit Security Group Remove Security Group Default ra
164. lume The difference is that you must specify FC Target in the Network File Protocol drop down menu as shown in Figure 60 below Specify a size in GB and a name for the FC volume Clicking on the Create button will create the volume compatible for FC target export It will be listed as an available volume on the FC target page where you can export it Ed g Wa re Server Storage Networking Authentication Maintenance Clustering Log Out e danual Provides logical storage management EdgeWare Storage Configuration Volume Management N 1 functions like logical volume creation N a 3 _ Create Logical Volume k d physical volume aggregation and snapshot EdgeWare 3 0 0 1ce creation etc Storage HELP RAID In volume group VGExport y Volume M Er EO Logical volume details K Volume name LYExpart 1 Volume size EN Share Manage Storage allocated from volume group GB 0 out of 7 Allocation unit size 4096 MB Network File Protocol v CIFS AFP and or UNIX Mixed ds sanae F S Windows CIFS Stripe across physical volumes UNIX NFS Te Cep Native Appletalk AFP F om iSCSI Target iSCSI Create E a A r iscsi ty v Return to index CS rc tar get C j Figure 60 Creating a FC target Reach the Add FC Exported Devices page by clicking the Add Exported Device button on the Storage gt FC Target module page Figure 59 On this page you can view the volumes available f
165. ma L vionnpe OE Valea sire Sim age eer ated irom volume geep GE 2E on ef Sen Adecation uati gire a ME Saapihot ol lagie al yoluane iaga Ch aa F Seki F RR te pes Figure 33 Creating a snapshot volume of a logical volume Page 51 A snapshot volume can be as large or a small as you like but it must be large enough to hold all the changes that are likely to happen to the original volume during the lifetime of the snapshot Snapshots affect the performance of logical volumes so create snapshots sparingly When automating snapshot creation with Scheduling Snapshots always enter a Maximum Concurrent Snapshots value to limit the number of snapshots at one time 1 Expand the Volume Group which contains the logical volume you want to snapshot Click on Add a snapshot volume hyperlink 3 You should get the Create Snapshot Volume page as shown in Figure 33 Enter a snapshot volume name This is a new name not the name of the parent volume you want to snapshot 4 Enter the snapshot volume size This should be large enough to hold any increase in size or other change of the parent logical volume during the lifetime of the snapshot 5 Select the logical volume that will be the parent or source of the snapshot 6 Click on Create to instantly create the snapshot or click on Schedule to automate snapshot creation based on a schedule SCHEDULING SNAPSHOTS If you click on Schedule on the Create Snapshot Vol
166. many more are available at your option WATCH LISTS On this page you can change the monitoring settings for the different components services that are being monitored as well as start monitoring for components or services that are not being watched A default list of services components should already be defined for the system These are normally adequate for monitoring a standard EdgeWare system In what follows a service is a running component hardware or software of the EdgeWare system and a service monitor or monitor is a timed check on the state of a service A monitor alert or alert is a response to an out of bounds state found by the monitor in the service or sometimes to a recovery from an out of bounds state There are different kinds of alerts see table below Important After making any additions deletions or changes you must click on Restart Monitoring Services in the Monitoring module main page Otherwise the changes will not show up in Monitoring Status or alert behavior You can get a detailed listing of the services being watched by clicking the servers link under the Watching group column Figure 88 shows the detail list of the default Watch List Host group watch details Host group to watch servers x Add host Services being watched Service name Monitor to use Check every Monitoring periods freespace freespace monitor 5m wd Sun Sat hardware ew 2i hardware monitor 5m wd Sun Sat RAID RAID
167. mple in Figures 11 and 12 below SCSI devices A and B are separate mass storage arrays each presented by a hardware RAID card and device C is the system disk not mirrored In Figures 13 15 by contrast you are modifying a system in which SCSI device C and SCSI device D are the mass storage The drives with location SCSI device A and B have the mirrored operating system partitions and the IDE device A is an unused DOM solid state drive not shown Figures 16 17 come off a system with three mass storage arrays and Figure 18 comes from yet another system Sizes and extents are given in cylinders Mass storage cylinders are always one binary megabyte 1048576 bytes long System disk cylinders can vary but are usually about 8 MB long Ui EdgeWare Storage Configuration Partition Management IN Partition Manager i EdgeWare 3 0 0 1ce partition file system ALL DATA WILL BE O The Partition Manager modifies the LOST Disk Partitions Role MASS STORAGE No partitions found Location SCSI device A A e Cylinders 1887956 Add primary partition Size 1844 GB Model XYRATEX F5402E Controller 1 Target 0 Role MASS STORAGE No partitions found Location SCSI device B SS Cylinders 1887966 Add primary partition Size 1844 GB Model XYRATEX F5402E Controller 1 Target 0 Role SYSTEM DISK No Type Extent StartEnd Use Free Location SCSI device C ab Linux 1 13 boot BE Cylinders 4500 2 Linux 14 268 7 57 Size 35 GB Linux
168. mplex deployments involving several steps NOTE Every parameter option needed by GUI and the EdgeWare OS is case sensitive Page 20 SERVER SECTION This is the first section that comes up when you enter the EdgeWare UI It offers server information RAID information and volume information that are current at the time the page appeared Hitting your browser s reset button will cause this page to reappear with newer information Navigating by picking module selections on the left side of the screen even module selections from the Server section will cause the page to be replaced by other modules main pages To make the initial page reappear select the triangle that is in the upper left near EdgeWare on your screen SYSTEM TIME MODULE In this module you can set the time and the date for the system and the real time clock hardware time by making selections from the drop down menus A ea e BS A Ed ge Ware Kaar Storage Maintenance Clustering Log Out aer l Networking Authentication Manual EdgeWare Server Configuration Time Edit the time you want displayed on entire System Time system Use the System Time edit link to EdgeWare 3 0 0 1ce change system time HELP E Synchronize system clock with time server Oo Time server hostname or address clock redhat com Set hardware time too Synchronize on schedule No Run on selec
169. n NAVIGATING THROUGH THE USER GUIDE In this user guide or manual navigating to topics mirrors the GUI layout where sections and management modules or modules for short define an outline or management tree The modules are shown on the left hand side of your web browser when accessing the GUI and the sections are on the top The modules are the base unit of deploying or removing additional management features Inside each module are management components or components for short typically many of them The convention of grouping modules in sections is a convenience to create a tree layout of related modules The resulting management tree is an outline to this guide as well All the topics in this manual are organized in the six sections corresponding to the GUI and summarized below Page 19 Some feature deployments such as Failover for example are complex enough to require action in several sections Each of these complex deployments has a lead section in Failover s case it is Clustering Such a deployment guide is found in its lead section in this manual and the description is organized constructively by steps with checks to map the necessary progress through other sections You have already encountered such a constructive guide with steps A B and C in Connecting to EdgeWare Servers for the First Time above SERVER SECTION The Server section includes modules for System Time Webedge Management and GUI Administrators Select
170. n array of clients and the target may also have NAS volumes These other tasks should be served with separate network ports and switches usually DHCP Having set up proper network ports on the EdgeWare side you must now do the same on your Windows system Connecting to EdgeWare Servers for the First Time at the beginning of this manual deals with many of the techniques needed here for configuring client IP addresses The Windows IP addresses meant for iSCSI should also be static Use the following IP address not DHCP Obtain an IP address automatically Be sure the netmasks and the subnets match the correct physical ports From your Windows machine you will be able to ping all the iSCSI addresses on the EdgeWare Server when this is SO C CONFIGURE THE MICROSOFT INITIATOR You have now exported iSCSI target volumes and they should be visible across the network connections you have established with the Windows system The third step involves the Microsoft Initiator which is a complex enough piece of code so that we will sketch its use in this manual even though it is not a Cutting Edge product What follows is our best effort at guidance though of course we have no control over changes that Microsoft may have introduced since this manual was written To configure a Microsoft iSCSI initiator there are several steps that you must follow 1 Install Microsoft Windows Initiator in your workstation or Windows client If you are
171. n log SYSTEM LOGS The System Logs page gives you access to 6 log files that are being continuously logged A Description line for each log file helps you pick which one to examine Click on any log file name under Log destination to get 20 recent lines of logged messages You can increase the number of lines displayed from the View log file page You can also filter the log display by searching for a string or keyword PROCESSES The Processes page Figure 93 gives you a list of the currently running processes in the EdgeWare operating system You can sort the processes based on PID Process ID User Memory used or CPU on multiple processor system There is also a Search option where you can search for processes based upon different criteria such as owner file system using file etc Click any individual Process ID number to get detailed information Warning User sort may take several minutes with a blank screen in the meantime It is not really hung EdgeWare Maintenance Configuration Process ss Running Processes O List of currently running processes EdgeWare 3 0 0 1ce HELP Display PID User Memory CPU Search Process ID Owner Started Command il root 1333 init 3 2 root 13 33 migration D 3 root EH ksoftirqd 0 4 root 13 33 migration 1 5 root 13 33 ksoftirgd 1 6 root 13 33 migration 2 th roat 1333 ksoftirgd 2 8 root kes migration 3 fel root jewel ksoftirgd 3 10 root 13 33 events 0 Tii root
172. n page Select the Logical Volume that underlies the mirror in this case LVstorage This should give something like Figure 140 saying Current status is For mounting on as remote mirror 3 Now click Mount the logical volume om BOS ARMY A Server Storage Networking Authentication Maintenance Clustering Log Out rel EdgeWare Storage Configuration 1 S Provides logical storage management Volume Management functions like logical volume creation P Edit Logical Volume O physical volume aggregation and Storage EdgeWare 3 0 0 1ce snapshot creation etc ge S H RAID BET In volume group VGstorage Volume Manager KI EE Volume name LVstorage Volume size 15 Storage allocated from volume Allocation unit ER group GB 16 out of 152 aze 4096 MB FEED Device file dev VGstorage LVstorage Current status For HE od QUES COR age Re remote mirror J oa ES nent Mount the logical volume Press the mount button to mount the current filesystem type on the logical volume Select a filesystem type and click this Change use moie All size files and few ACLs lt 50 ext3 vi button to create a new filesystem on l this logical volume This will erase any data currently on the volume Return to index Figure 140 Creating Mirrored Volume 4 A confirmation screen like Figure 141 should appear Notice that a drbd device is mentioned not the LVstorage volume name that was
173. n with mount t nfs 192 168 1 158 shares sharetest mnt nfs where you use your real server IP address in place of 192 168 1 158 and your real share name in place of sharetest even without any entry in etc fstab If this approach is used the mount will not automatically be restored after a reboot J Run the command mount to see that your share is mounted Run the command df to see how much space is in it FTP The FTP module lets you configure and activate the EdgeWare FTP File Transfer Protocol server This is a long accepted protocol that even predates the World Wide Web and is particularly useful for downloading large files Uploading files is also possible but typically restricted Since there are no share specific settings your only control on shares that you do not want to be FTP enabled in a server with FTP active is to uncheck FTP in the Storage gt Share edit for that share You can start restart or stop the FTP server operation from the main FTP page by clicking the appropriate link Select Denied Users to see a system fixed list of users that are never allowed FTP access Select Networking Options to get a page Figure 79 of global FTP settings applicable to all your FTP capable shares After saving these restart the server in the module main page NETWORKING OPTIONS The Networking Options page checks and changes the global settings for the FTP Server Let s look at the various entries of Networking Options Local Us
174. napshot now This will remove all but one of the snapshots generated by the old schedule The process can take a few minutes Then you can delete the last snapshot and the new schedule too Page 54 Creating and managing NAS shares is done via the Share Manager You can share a specific folder in any mounted logical volume created in the EdgeWare Server and make it available to users on the network When you create a share you can assign permissions to the share for any of these users who will then be allowed or denied specific operations when they access the share over the network Here users usually refers to network users as dealt with in the Authentication section rather than administrative users as dealt with in the Server gt Administrators module The exception to this is Networking gt Http Shares are not used by iSCSI volumes CREATING AND EDITING A SHARE H KE af J amp A r e A IE g eWare Server Storage letworking uthe ti M usterin ul Veer Choate File Shar SE A Provides creation and management of Hart ay ae network file shares on logical volumes Storage EdgeWare 3 0 0 1ce d iN 5 RAID e oo Manager m Share name et Volume or Directory to share Browse r os PS Share Comment w Backup e Enabled Protocols v CIFS VINFS vlAppletalk VIFTP V HTTF Y Async GS iscsi EE Save vr View User Connections Delete Return to i
175. ndex Figure 36 Creating a network share Under Share Manager click on the link Create a new share Specify a Share name see Figure 36 Select a volume or directory to share by clicking the browse button A pop up window should appear after a few seconds listing the available volumes Figure 37 If you need to select a particular directory under the volume double click the volume link to go deeper into the volume tree Press OK to confirm your selection You can write a brief description of the share under Share Comment Select the network file protocols that you want to enable for the share Disabling the protocols you will not need improves efficiency and security and more protocols can be added later by editing the share See list below Click on Save to create the share Page 55 Se ae z LA A ext E SN 23 https 192 168 1 158 10000 Choose Directo EI a L Vo elt a eat EdgeWare irectory 4kB 16 Jun 2008 13 34 x bn and nt of Create File SH ES LvDOM 6 22 Aug z2008 10 14 ree on ogicel volumes EdgeWare 3 0 0 J Storage E e Ga LVlocal 4kB 16 apr 2008 08 45 RAID BE B LVstorage 12 kB 21 aAug 2008 19 09 gt 21 Aug 2008 17 37 ET Sharalinformation B LVstorage snap 24 aug ew m ngen ES Lvstorage snap_1219363200 4 kB 21 Aug 2008 17 13 Ki Sharema BW Lvstorage 121 1 BKB 21 Aug 2008 18 47 Share Manager Volume or Directory to sha G HET Share
176. new cdb user host new_file_name cdb 4 N es De AR A E a ge w are Server Storage Networking Authentication M Clustering Log Out Pc aintenance Clusterin anual EdgeWare Maintenance Configuration Integrity Zi x System Integrity and Restoration O Check OS integrity Maintenance 3 EdgeWare 3 0 0 1ce L N Shutdown merr an ee f Verify the current OS against a checksum database KE Monitoring EE Select the reference system database OE Local database ema Processes O Path to Archived Database Browse gt ay e Line Verify m1 Integrity ke dReturn to index a t Updates Figure 94 Verify page To verify the current database see Figure 94 1 If consistency with an external standard is to be checked select Path to Archived Database radio button and specify a path in Upload the database The default location of the database is var db integrit cenas cdb Then hit Update and go to step 3 2 If self consistency with a previous state of your existing Storage Server is to be checked make sure Local database radio button is set It will rename the last checksum cenas_new cdb to cenas cdb and use that 3 Click on Verify 4 System database verification will commence A page will appear saying Running the checksum command and you can exit this at any time to view the current state of verification DETAILS 5 After verification is complete you can view the result
177. nt system This iSCSI target controls a real peripheral device of the type desired and drives it according to the instructions of the SCSI commands coming in on the initiator side The combination of all these things is called an iSCSI device Page 64 It is common for the terms initiator and target to be applied to the systems containing the connected iSCSI initiator and iSCSI target software respectively In addition the name target is also applied to the real peripheral device that is controlled by the iSCSI target software on the target system In our case that is a Logical Volume An iSCSI node name is also the SCSI device name of an iSCSI device The iSCSI name of an iSCSI device is the principal object used in authentication of targets to initiators and initiators to targets This name is also used to identify and manage iSCSI storage resources iSCSI names must be unique within the operational domain of the end user However because the operational domain of an IP network is potentially worldwide the iSCSI name formats are architected to be worldwide unique The iSCSI protocol has three major characteristics that distinguish it from NAS and have contributed to its great popularity in storage networks 1 iSCSI is a block evel protocol which means that it cares nothing about file system or permission complexities that each NAS protocol must deal with In addition the blocks are usually mapped linearly from the initiat
178. ntrollers so the only RAID type available is the software RAID However even when hardware RAID exists you can place software RAID on top of partitions from arrays you created by hardware RAID To do this first select the Modify Partitions bar and create the partitions SOFTWARE RAID Important If your appliance even a Hardware RAID appliance has mirrored system disks these will appear as SOFTWARE RAID and should not be disturbed as they are mounted in critical system locations Manipulate only mass storage with the software RAID tools described below The Role of the off limits disks is given as SYSTEM DISK while for the ones open to reconfiguration it is MASS STORAGE see Figure 11 Cutting Edge EdgeWare OS supports four types of RAID configurations for the data partition These RAID levels can be used on a multi drive appliance whether or not a hardware RAID card is installed If there is no hardware RAID card installed you can use software RAID It is important to note that for a two drive appliance the only RAID levels available are 0 and 1 For an appliance with three or more drives RAID levels 0 1 and 5 are available At 4 or more drives RAID 6 also becomes available Page 28 RAID 0 Single Large Volume This type of RAID is also referred to as striping RAID 0 interleaves blocks of data between several drives Even though multiple drives exist in the array to the user a RAID 0 appears as a single large
179. nts that can be made to your Storage Server through SNMP If you use SNMP to monitor and control heterogeneous systems please contact Cutting Edge for MIB details EMAIL NOTIFICATION This page is used to set up Email Notification for alf monitors Here you can enter the parameters to set up Email Notification and send a test email to make sure they are working Normally a mail server on your network must already exist before you set up this page to relay your messages to the receiving email address The only exception is if that receiving email address is served within the same local network as your EdgeWare Storage Server Frequently email accounts nduding business email accounts with your company s domain name are served by a distant email provider like gmail Figure 92 shows a typical configuration of this sort Of course replace cuttedge com with your appropriate company domain name and the Recipient s email address with your own EE es D e gi S A A EdgeWare Server Storage Networking Authentication D Clustering Log Out of EdgeWare Maintenance Configuration Monitoring fee i im E mail Notification O Setup and manage warnings and errors Maintenance EdgeWare 3 0 0 1ce l p N Shutdown EI Ze ry E mail Notification de Ee cuttedge com SE Logs Clients to relay mail from Destination to relay mail to cuttedge com Processes EI Proxy NAT external network ad
180. o Manager Share Manager All users on your system j za EE Selected users guest Browse G eem Members of selected groups a Browse nr U ts GS iscsi LE ed Return to index Figure 41 Copying guota settings of a user to other users OUOTAS FOR NEW USERS AND GROUPS You can also specify quotas which will be enforced on any new users and groups that will be created later on the system Enter the relevant values in the New User Quotas area near the bottom of the Filesystem Quotas page shown in Figure 39 BACKUP MODULE Backup allows you to back up your data in the logical volumes to another volume in the system or to a tape device If your system has a tape drive or tape library connected you can back up your data to that device Otherwise you can only back up to another logical volume This backup module operates in two directions backup which saves a copy of a logical volume or restore which copies over a logical volume with data from a previous backup DESIGNING A BACKUP SYSTEM Backups serve two purposes e Recovery from catastrophic data loss e Storing old versions of your data so you can revert to them if necessary For either of these to work you must apply the first principle of backup design See to it that an event that takes out your current data is unlikely at the same time to take out your backup This not only includes hardware failures but software and user errors Alternative approaches that satis
181. o assign to the valid users on the share R read W write and X execute Checking Allow unlimited access to all users will check R W X for all the users 4 Optionally you can specify these permissions to be propagated to all the files and subdirectories or all the subdirectories under the share by checking the relevant check box under Apply ACLs recursively to all files and subdirectories 5 Click on save to apply the changes Page 111 uthenticatior gt gene ii ig Si Maintenance lustering Log Out taar Networking Authentication Clusterin Manual Server Storage EdgeWare Storage Configuration Share Management m Provides creation and management of je Edit Security O network file shares on logical volumes J EdgeWare 3 0 0 1ce HELP For share sharetest Valid users Valid groups user everyone _ Allow unlimitted access to all users Users VIR OW MX Groups MIR lw Mx DIACLs to all files and subdirectories Save Return to index Figure 84 Editing share permissions Adding groups to the share 1 Type a valid group on the system into the text box labeled Valid groups or select groups by clicking on the button with the three dots at the bottom right of the text box which will give you a popup window listing all the local or domain groups for the system Repeat step 1 for all groups you want to add to the share Check t
182. o be iSCSI fresh volumes will be mounted by default e If you have a pre existing primary volume to be mirrored create only on the secondary On the primary demote the volume In the case of NAS this involves a Do a configuration backup of share information b Delete every share on this volume remember that does not destroy data In the case of iSCSI this involves a Pause every user application using this iSCSI volume on the initiator b In the initiator disconnect this target this causes a flush across the network c In Storage gt iSCSI delete this iSCSI target this does not destroy data If it is newly created or NAS make sure that you check the disable mounting box in each volume and click Unmount the logical volume It should return unmounted Figure 132 If in the NAS case it does not unmount or if in the iSCSI case it complains that it cannot be renamed or resized as it is currently in use you must return to step 3 and complete it In both NAS and iSCSI cases Current status should be Not in use This step must be carried out on both servers d AN D SG Sg SS 3 4 User Networking Authentication jaintenance Clustering Log Out Manual EdgeWare Storage Configuration Provides logical storage management H Volume Management functions like logical volume creation gt 4 Edit Logical Volume AJ physical volume aggregation and ti hot ation etc us EdgeWare 3 0 0 1ce Snapshot crea
183. o the new eth0 IP addresses Page 145 EdgeWare Network Configuration Network Create Host Address EdgeWare 3 0 0 1ce IP Address 10 168 1 133 Hostnames Follow22 ethi Figure 112 Network gt Advanced Configuration gt Host Addresses gt Add a New Host Address page A am ea S Ye A E d g 6 wW are S Storage Networking Authentication M Ch g Log Out User erver Storage Networking Authentication Maintenance Clusterin Manual Engear Na re Configuration Network e Provides configuration options for the IP Host Addresses O protocol like IP addresses and DNS client Networking EdgeWare 3 0 0 1ce 1 Pe Network HELP aaa Add a New Host Address gt Jp EE cIFS n SES e 127 0 0 1 localhost localdomain localhost aS WS 192 168 1 130 Leader22 D 10 168 1 131 Leader22 eth1 AppleTalk i 192 168 1 132 Follow22 Le CG 10 168 1 133 Follow22 eth1 _ EE Add a New Host Address Q w hd Return to index R allel ee Figure 113 Host Addresses page after step E is done E RECORD ALL PORTS IN BOTH SERVERS HOST ADDRESSES LISTS On each server go to Networking gt Network gt Advanced Configuration gt Host Addresses There see that both ethO and eth1 of leader and follower are recorded with their correct IPs and hostnames In each server you will probably find one eth0 of current server and have to add three compare Figure 112 This places the hostname information in a stati
184. of this third party software who are intended third party beneficiaries of this License RESTRICTIONS REMOTE ACCESS INTEROPERABILITY The Product documentation and the associated copyrights and other intellectual property rights are owned by Manufacturer or its licensors and are protected by law and international treaties Licensee may not copy or reproduce the Product and may not copy or translate the written materials without Manufacturer s prior written consent Licensee may not copy modify decrypt disassemble reverse compile or reverse engineer the Product or sell sub license rent offer on a service bureau or ASP basis or transfer the Product or any associated documentation to any third party Licensee may not use the Product except as embedded in or installed on a a Manufacturer Network Storage Node or b Manufacturer authorized and supported hardware as offered by an authorized Manufacturer distribution partner To the extent required by law Manufacturer will provide Licensee at Licensee s reasonable request with interface information needed to achieve interoperability between the Product and another independently created software Cutting Edge 1825 Gillespie Way Suite 100 El Cajon CA 92020 800 257 1666 www cuttedge com r NETWORKED A G E s CUTTING c EDGE oO R program upon payment of Manufacturer s applicable fee Licensee will observe strict obligations of confidentiality with respec
185. ollower s module main page after leader Primary failure Important note Killing the Primary includes any shutdown of the Primary including Reboot or Shutdown selected from within Maintenance gt Shutdown on the EdgeWare UI A failover cluster is sensitive to such events in a way that a standard Storage Server is not Study this subhead especially Breaking a Failover Cluster below before performing any reboot or shutdown on the Primary of a failover cluster SWAPPING THE PRIMARY AND SECONDARY ROLES The Primary and Secondary roles within a failover cluster can be swapped without a fail event This is not usually recommended because of the possibility that newer configuration information may be lost However it is an option if the Primary must be shut down or for some other reason should not be leading the cluster at this time 1 Enter the module main page of Clustering gt IP Failover on the Secondary s UI The Resource Status should be Down red 2 Select the blue button with the arrow under Get Resource Figure 126 After a few seconds the Resource Status should turn to Up green The former Secondary has then pulled Primary status away from the former Primary The configuration of the HARs will revert to the state found on the Secondary s system disk which is normally the state at the time failover was first triggered Page 159 EDITING THE RESOURCE CONFIGURATION IN A RUNNING CLUSTER Limited changes
186. one Jmportant It lists all Storage Servers whether they are being used as NAS or iSCSI or both The User Interface NAS Servers Index refers both to NAS and to iSCSI targets Clustering Log Out EdgeWare Server Storage Networking Authentication Maintenance FL cat mm EdgeWare Clustering Configuration Kb Gage O Contact managed servers EdgeWare 3 0 0 1ce x S allover LU 192 168 1 254 10000 edit CEZE478A 10000 edit F6412E 10000 edit FiberTarget012U 10000 edit v EiberTarge 210 10000 edit Register a New Server Click this button to find any EdgeWare servers on your local network scanfor Servers Click this button to check every address on the network 192 168 1 0 for EdgeWare servers Figure 100 GUI Servers found on the local subnet Page 135 The first step in establishing an EdgeWare Storage system cluster is to establish a cluster group In order to do this you need to initiate contact with some or all of the Storage Servers This is done by clicking in the Clustering icon and then selecting the NAS Servers Index icon Next you can search for additional EdgeWare servers by clicking on the Broadcast for Servers or Scan for Servers icons Or if you have already registered the hostname and IP address in Host Addresses see next paragraph you can specify a server by clicking on Register a New Server and typing in its hostname Either way the screen will
187. or exporting as FC Target Figure 61 Specify an Alias then click Save Alias A more human readable and understandable name for the target Page 81 Ed ge Ware Server Storage Networking Authentication es aintenance Manual e EdgeWare Storage Configuration FC Target Mode Configuration Provides configuration for Fibre Channel Export ive Device as FC Target Target Mode Support Storage gt EdgeWare 3 0 S RAID Export LVM Device as FC Target MET ager l S ER a Backup om on G wii Return to index Figure 61 FC target export property page Once you go back to the FC Target main page you will see the LVM device added for export Figure 62 Fibre Channel Adapters enabled for Target Mode EE EE Action isp3 Disabled Enable isp2 d OD Disabled Enable isp1 5 301 Disabled Enable ispo 4 300 Enabled Disable Exported Fibre Channel Target Mode Devices Alias izelMB Block Size Options FileName Remove from Export List FCExport 183296 512 dewVGExport LVExport Remove Add Exported Device LUNs and Security Groups Default edit Add SecurityGroup gt Figure 62 FC target main page showing exported LVM ASSIGNING LUNS TO THE DEFAULT SECURITY GROUP Next you will need to define security groups and assign Logical Unit Numbers LUNs to the exported devices As you can see in Figure 62 above the Default security grou
188. or s virtual device to the target s real device This vastly reduces overhead and makes efficient read and write strategies possible 2 iSCSI is normally a point to point protocol one initiator at a time for each target This avoids the heavy burden of arbitrating between clients permitting efficiency at the cost of some flexibility It also makes it desirable and possible to use dedicated physical network connections between initiator and target for high volume iSCSI devices 3 iSCSI presents a oca virtual raw device which from the point of view of the application presents no network or file system complexities This is highly desirable for many applications such as databases See Figure 58 for example the iSCSI drive is D x Edgeware Storage Configuration iSCSI Configuration Provides configuration for ST iSCSI Connection Manager iSCSI target and initiator Storage HELPT B cm d iSCSI Initiators iSCSI Targets Figure 47 iSCSI module page including initiator Page 65 One or more iSCSI devices can be built on any EdgeWare system All EdgeWare systems have an iSCSI target drivable by Microsoft Initator and other iSCSI initiators on systems elsewhere on the network Some EdgeWare systems Figure 47 also have an iSCSI initiator that can be used with other iSCSI targets elsewhere on the network This manual will describe the Microsoft Initiator NOTE In EdgeWare Initiator systems iSCSI services c
189. originally chosen This shows that the mirror is active Successfully mounted logical volume dev drbd0 Figure 141 Mirror mount confirmation 5 Exit the confirmation and notice that Current status says something like Mounted on Volumes Mirror0 6 Click on the Share Manager icon to get to the Share Manager module main page 7 Click on Create a new share which should yield a screen like Figure 142 Page 170 Figure 142 Creating a mirror share Leader22 CE Nas 192 168 1 134 Leader22 File Edit view Favorites Tools Help Address W Woeader22 Network Tasks E testshare Re Printers and Faxes fe Add a network place view network connections Set up a home or small office network 2 Set up a wireless network for a home or small office wy View workgroup computers Show icons For networked UPnP devices Other Places jy Workgroup 3 My Computer 3 My Documents Shared Documents KN Printers and Faxes Figure 143 Logged into Storage Server Page 171 10 11 12 13 14 15 16 Enter the share name Click on browse and select a name like Mirror0 the same as you saw in step 5 as the volume to be mapped Note Mirrored volumes have names like Micro Enable the desired protocols and click on Save You now have a mirrored share Next open Authentication gt Users and Groups module main page Set up the authentication of users and groups as described in t
190. ou wish in a browser window on the client If you will be connecting as root skip to step D C Create a user on the server in Authentication gt Users and Groups gt Local Users and Groups gt Create New User Give this user the same name as the user who will be connecting from the client This step is not necessary if you are connecting as root D Create a RAID array a volume group a logical volume and a share as described above in the Storage section E Set up NFS as just described If your client user is root select Everyone otherwise leave the default Everyone except root in Trust remote users Be sure the Export to setting permits the client system to connect When done select Save F Start or restart the NFS exports on the module main page G Return to the client s command line as root Make sure you have a mount point an empty directory over which the share will be mounted If necessary run a command like mkdir mnt nfs to create one H If your user will not be root add a line like 192 168 1 158 shares sharetest mnt nfs nfs user 0 0 to the end of etc fstab where you use your real server IP address in place of 192 168 1 158 and your real share name in place of sharetest and then switch to your non root user and enter the command mount mnt nfs to mount the listed share Then go to step J Page 102 I If your user is root then instead of step H you may do the whole mount operatio
191. ove initialization being carried out on the new replacement leader system and the heartbeat being started and control of the HA resources seized by the system follower that already has control presumably with no effect This amounts to replacing Figure 114 HARs Heartbeat with a diagram in which Heartbeat follows HARs in time and in fact the bootup of the new leader now the Secondary fits into the gap in time Page 158 FAIL EVENTS Killing the Primary leader causes after a pause the follower s DOWN red to turn to UP green Contrast Figure 128 to Figure 126 The resources then come up on the follower which is now the Primary and the shared IP points to the follower now All this happens within a few seconds preventing significant downtime EdgeWare Clustering Configuration Heartbeat i KEE Heartbeat Monitor id A address and service failover EdgeWare 3 0 0 1ce HELP Primary Node IP Addresses Node Services Resource Status Get Resource Leader22 192 168 1 158 hb_mounts services s Create Modify Resource To define a failover resource Authentication Keys r To configure the type of authentication for heartbeat events Heartbeat status heartbeat OK pid 8038 et al is running on follow22 follow22 Apply Changes gt Click this button to force the running Heartbeat monitor process to reload the current configuration Stop Heartbeat Click this button to stop Heartbeat monitor process Figure 128 F
192. overy of servers on XP and other Windows based clients UPnP discovery protocol allows the server to advertise its service to a control point of the network The control point searches for other servers on the Network and discovers them e UPnP Service in Windows can be located installed in the following manner i Go to Add or Remove Program under control panel ii Hit Add Remove windows component iii Select Networking Services and hit Details iv Put a check on UPnP User Interface and hit ok to install Figure 2 v Open Services msc under Start gt Run and ensure that SSDP Discovery Service is started Figure 3 Page 16 mt H besse ne Intel R GMA Internet ge Sort by Name i Hers Driver For Options i sie o 2eme A Windows Components Ve You can add or remove components of Windows XP hatworiking services To add or remove a component click the check box amp shaded box means that only part of the component will be installed To see what s included in a component click Details To add or remove a component click the checkbox A shaded box means that only part of the component will be installed To see what s included in a component click Subcomponents of Networking Services Details Intemet Gateway Device Discovery and Control Client 0 0 MB Components El Peer to Peer W MSN Explorer Al RIP Listener f Networking Services 0 3 MB E Simple TCP IP Services 37
193. p 5 hit Stop Heartbeat and await confirmation which may take twenty seconds or so that the heartbeat is off Important If you wait too long before doing this you will never get confirmation because the Primary will be unable to shut down Heartbeat If that happens you must hand remove Maintenance gt Command Line all heartbeat or ha processes with a kill or kill 9 To find these on both Primary and Secondary go to Maintenance gt Command Line and run ps ax grep v grep grep heartbeat ha On both Primary and Secondary go to Maintenance gt Command Line and delete haresources and ha cf rm f etc ha d haresources etc ha d ha cf From a third system ping the shared IP address See that it is down On both Primary and Secondary go to Networking gt Network gt Advanced Configuration gt Host Addresses and remove the shared IP address and hostname line On the Primary go through the resources that you deactivated in step 3 If the service has started select its Stop button If the service is still stopped Start it and then Stop it This is a cleanup step that prevents it from restarting on reboot Check Primary and Secondary IP Failover page It should look like Figure 129 Ed g eWare Server Storage Networking Authentication Maintenance Clustering Log Out Veer Manual EdgeWare Clustering Configuration Heartbeat d s p q O Provides configuration options for IP GP Heartbeat Monitor 2
194. p is already defined and is not removable If you assign a LUN to the Default security group it will be visible to all initiators attached to the Edgeware box Page 82 4A ate J 4E A EdgeWare Dirar Storage Networking Authentication Maint ustering og Ou Mare CS EdgeWare Storage Configuration FC Target Mode d m Provides configuration for Fibre Channel Edit Security Group gt Target Mode Support Storage EdgeWare 3 0 0 1ce S CR e LUN Assignments Group Default LUN Options Rem KI Share Manager Add LUN Assignment GG wos Adapter WWN Assignments Group Default Group Default is accessible by all Initiators by definition DE e Add Authorized WWN q EE k G FC Target Figure 63 Property Page FC Security Groups To assign a LUN for the exported device to the Default security group click the Edit Security Group button then select Add LUN Assignment Figure 63 Timm BOS PHBE A Storage Networking Authentication Maintenance Clustering Log Out Sea ei EdgeWare Storage Configuration FC Target Mode CS Pare i O Provides configuration for Flore Channel Storage EdgeWare 3 0 0 1ce RAID _ Add LUN Assignment E snare anager k f Rad ent Ap ze SEG Group Din Cras Local Device FCExport 2 ELI LUN FC Target Save r Return to index Figure 64 LUN assignment page fo
195. pecifications or parts that are Field Replaceable Units based on the following service criteria TIME OF REQUEST For Gold Support when Customer calls are remotely diagnosed by Manufacturer as a hardware problem before 12 00 p m Pacific Time Manufacturer will initiate replacement product or part shipments by the next business day Manufacturer business hours are 8 00 a m to 5 00 p m Pacific Time Monday through Friday excluding holidays observed by Manufacturer EXCHANGE OF PARTS All replacement products or parts will be provided to Customer on an exchange basis Customer must return the replaced hardware to Manufacturer within ten business days of receipt of the replacement hardware product Customer agrees that if it does not return the replaced hardware within the ten business day period it will pay Manufacturer s then current price for the replaced hardware Return of hard drive top plates are acceptable from Customer s in secure locations Non return pricing is available REPLACEMENT PARTS Replacement product or parts may be new reconditioned or refurbished The replaced hardware will become the property of Manufacturer Customer will ship all replaced hardware for exchange or repair to Manufacturer at its own expense and in accordance with the RMA number and shipping instructions provided by Manufacturer Manufacturer will pay shipping costs for the hardware being shipped to Customer except that Customer will pay any applicable
196. r RAID R Replicati ISMART Drive Status Scheduled Cron Jobs Setup Wizard Share Manager Share Permission Shutdown Synchronous Mirroring System Time Updates Users and Groups Version Control Volume Manager EdgeWare Groups Figure 7 GUI root administrator details GUI ADMINISTRATIVE GROUPS This component provides the functionality to create administrative groups for the web UI You can create groups and then add the administrative users you create under GUI Users to specific groups by selecting the group in Member of group when enter the click Create a new GUI User as shown in Figure 7 You can use GUI Groups to create administrative users quickly with module access control specified in the GUI group so that you do not need to select or unselect the modules when you create each user See Figure 8 and Figure 9 The normal use of groups is to define a subset of the modules list that an admin of this group can deal with The example of Figure 8 and Figure 9 is meant to deal only with quotas and a few related modules such as shares Note that when the user created in this group is expanded its access rights do not even show quotas you must expand its group before you can modify these access rights and then they are modified for every user in the group Page 24 EdgeWare Ss a Ki Gel bf A mu uu a maa aa aa au aa au aa CARES Figure 8 Create GUI Group Storage Networking Authen
197. r Broadcast for Servers as found in the NAS Servers Index section of this manual Your new server should appear on the display page and you can get its IP address from there FIND THE SERVER ON A NON DHCP NETWORK OR POINT TO POINT CONNECTION In the case of Windows Vista the normal discovery takes place as described in Set Client IP Address and Netmask on Microsoft Vista above If DHCP is absent it is possible that the standard discovery tools such as uPnP or Vista Network Discovery will not work In this case the fallback position is to run the Cutting Edge tool EdgeWare Discovery Tool or Magellan The following discussion is for Magellan EdgeWare Discovery Tool is discussed under DHCP but works for non DHCP too Install and run Magellan This may involve first installing Java found on the Magellan CD Click the Refresh button This reveals the EdgeWare system s IP address Access GUI at https IP address 10000 Login to the EdgeWare web based UI with default username admin and password setup Set static IP address information for your production subnet PAD esa The best option is to exit Magellan and go straight to step 5 access the GUI as soon as the Server s IP address is obtained If neither monitor nor Magellan is available the standard Auto address of 169 254 5 58 may be tried FINDING THE SERVER ON A DHCP BASED NETWORK In the case of Windows Vista the normal discovery takes place as
198. r FC Security Groups On this page you can view the exported volumes available for LUN assignment Figure 64 Specify the property below then click save LUN Logical Unit Number The first assigned device must be 0 Subsequent devices can be 1 to 255 Page 83 A es Deg SS A A E d g eWa re Server Storage Networking Authentication Maintenance Clustering Log Out BA Manual EdgeWare Storage Configuration FC Target Mode Configuration Provid figuration for Fibre Channel sch SE HE OD E Storage EdgeWare 3 0 0 1ce J HELPI RAID LUN Assignments Group Default i ee Manager a S S Virtual piss LUN options Remove LUN Assignment KI Share Manager FCExport 0 Remove Add LUN Assignment CG vous Adapter WWN Assignments Group Default Backup Group Default is accessible by all Initiators by definition Ka S Add Authorized WWN gt CSS Fe target Return to index v Figure 65 Edit Security Group page showing LUN assignment When you are done the Edit Security Group page will show the LUN assignment within the security group Figure 65 ADDING ADDITIONAL SECURITY GROUPS Additional security groups can be defined so that certain FC initiators will only have access to specific exported volumes To add a security group click the Add Security Group button on the Storage gt FC Target module page You will be taken to the Add Security Group page Figure 66
199. r this nesting is that the Primary broadcasts and the Secondary listens if the Secondary does not hear the heartbeat it takes drastic action PRIMARY SECONDARY Start heartbeat Start heartbeat Get resources Reconfig Restart Reconfig Restart Stop heartbeat Stop heartbeat Though normally not needed EdgeWare offers this step explicitly for clarity Figure 123 Heartbeat nesting in time without failover Following Figure 123 here is the sequence in EdgeWare systems for starting the heartbeat 1 Logged into the leader go to Clustering gt IP Failover and select Start Heartbeat Monitor The result is shown in Figure 124 The resources are still DOWN EdgeWare Clustering Configuration Heartbeat GEET d Heartbeat Monitor a address and service failover EdgeWare 3 0 0 1ce HELP Primary Node IP Addresses Node Services Resource Status Get Resource Leader22 192 168 1 158 hb_mounts services vow a Create Modify Resource To define a failover resource Authentication Keys To configure the type of authentication for heartbeat events Heartbeat status heartbeat OK pid 9003 et al is running on leader22 leader22 Apply Changes Click this button to force the running Heartbeat monitor process to reload the current configuration Stop Heartbeat Click this button to stop Heartbeat monitor process Figure 124 Leader Heartbeat module page after Start Heartbeat Monitor 2 Logged into the follower refresh the I
200. ration can be selected if desired Usually the only non default that is of any interest is an email target If you wish this it needs Email Notification bottom of Advanced page the same settings as described above in this manual for Maintenance gt Monitoring gt E mail Notification and also an address higher up on Advanced page Then save it If you do not need an email target you can probably skip Advanced Configuration Select Shared Storage of Failover Resources if desired but see below The Middle in Figure 115is the default Selecting Shared Storage adds Outermost selecting Block Devices adds Innermost selecting both does both Whether you are NAS or iSCSI or both ALWAYS select Shared Storage NEVER select Block Devices Page 153 You must now determine the shared IP address and hostname that your failover cluster will use Find an UNUSED STATIC IP ADDRESS available on your LAN It will take the place of the etfi0 address after failover is running Enter it under Failover IP address Figure 120 Check your work and select Save It will put you back in the module main page Node Services are listed Resource Status is Down red and Heartbeat is still down Figure 121 Primary Node IP Addresses Node Services Resource Status Get Resource Leader22 192 168 1 158 hb_mounts services DOWN Heartbeat down Figure 121 Leader s response to configuration save Do not change anything on the follower V REGISTER THE SHARED IP AD
201. rectory structure of the HTTP enabled shares in the system and the right pane shows the files under the directories you select on the left pane The menu at the top gives you different actions for files After a selection pop ups may need to be enabled on your system Downloading files A double click of the left mouse button on a file lets you open it with programs that Java can find on the administrator s system Such programs in Windows XP permit saving the file with a Save As In the case of a graphics file the htm selection will generate a directory and the actual graphics file can be found there after the download Page 106 Uploading files Click on a share in the share tree in the left window of the file browser Once the share is selected click on the Upload button Click on Browse button to select a file to upload and verify that Upload to directory has the right directory set Click on Upload button the window will close once the operation is completed and the new file will show up in the right file browser window where the share contents are displayed Other actions The row of buttons along the top of the Java window permits the administrator to perform powerful operations including whole directory tree deletes Take care before performing such operations on real customer shares in your Storage Server as the lost user data may be unrecoverable For moving whole directories from place to place use the copy and pa
202. red Products with other software or hardware not supplied or not approved by manufacturer f use of the Covered Products on or with an unsupported hardware or software platform g abnormal usage or misuse of the Covered Products including but not limited to accident fire water damage earthquake lightning other acts of nature and other causes external to the Covered Products h installation or maintenance of wiring circuits electrical conduits or devices external to the Covered Products i Customer s failure to provide and continually maintain adequate electrical power air conditioning and humidity controls in accordance with Covered Product specifications j excessive wear or deterioration of the Covered Products k removal of the Covered Products from the location originally specified by Customer or reinstalled without the prior written approval of Manufacturer I the Customer s or a third party s negligence m any breach by Customer of this Agreement n board level repairs made to the Covered Products other than by Manufacturer unless the board repairs were approved in advance in writing by Manufacturer or 0 Customer s refusal or failure to implement any Error correction or replacement part made available to Customer by Manufacturer Additionally Manufacturer will have no obligations for any Covered Product in which the tamper label has been altered or the hardware cover has been opened Only trained personnel will
203. ress or hostname of the Primary Domain Controller of the domain or any Active Directory Service Domain Controller of the domain Administrator The domain user in the Domain Administrative group for the Windows NT4 or ADS domain Administrator Password The password for the user entered in the Administrator field ADS option KDC Enter the IP address or hostname of the Kerberos Key Distribution Center this is usually your ADS domain controller same as the PDC ADS server field above Realm A network that includes a Kerberos server is called a realm This should be your ADS FQDN fully qualified domain name such as cuttedge com or sd cuttedge com NOTE Date and Time are critical for the domain joining to succeed Ensure that the date is same as on domain controller s and time on the NAS is set to be within 30 seconds of the time on your domain controller s After making sure you have entered the correct settings for Windows networking click on Save and Restart Samba Status message above this button should update to Samba Status Joined servername to realm rea mnam e Once samba status says NAS has joined the domain reboot the NAS from the Maintenance gt Shutdown module to complete the domain joining Windows Domain Users and Groups Once the NAS is joined to the domain this section gives you a listing of the users and groups in that Microsoft Windows domain Page 96 WINDOWS DOMAIN JOIN For many custome
204. rs it is important to receive user and group authentication from an external Windows domain The following set of instructions will assist you in joining such domains including NT4 Windows 2000 or Windows 2003 SP2 1 Stop Samba a Start at the main page of the Networking gt CIFS module Figure 73 b If Stop Samba Servers button is visible at the bottom select it If Start Samba Servers button is visible at the bottom as in Figure 73 do nothing 2 Set DNS Server s a Click on Network on the left side of the screen This places you in the Networking gt Network module Figure 70 b Go down to the bottom of this screen to the DNS Options and populate the DNS Server textbox with an IP address of a host server Shown already set to 192 168 1 101 in Figure 70 Click on the Save button 3 Enter ADS info and restart Samba a Click on the CIFS to the left This returns you to the CIFS module Figure 73 b Now click on Windows Networking Options in the center pane See Figure 75 c Populate the fields for your local host e Change WORKGROUP to your host server name WINS Mode to Neither Master Browser Elections to Decline in PDC and ADS option Yes and the IP address admin name and password in ADS option IP address and Realm name name An after picture is Figure 76 Windows Networking Options Security ADS member Domain Workgroup Domain O Default LAB03 WINS mode O Act as WINS s
205. rs of the specified NIS netgroup to access this export Page 100 Network and Netmask Any host in the specified subnet is allowed access For example if the network was 10 1 2 0 and the netmask 255 255 255 0 all hosts with IP addresses from 10 1 2 0 to 10 1 2 255 would be allowed Use asynchronous writes to disk This option improves speed at the cost of robustness by treating the entire EdgeWare server as the disk When this option is set the server replies to clients before data has been written to permanent storage The server also sends a FILE_SYNC response to the client indicating that the client need not retain buffered data or send a subsequent COMMIT operation This exposes the client to the same undetectable corruption as exists for NFS Version 2 with async if the server crashes before it has actually written data to stable storage Note that even if a client sends a Version 3 COMMIT operation the server replies immediately if the file system has been exported with the async option Clients must be on secure port If yes is chosen NFS clients must use a UDP port less than 1024 This provides additional security for UNIX clients but may interfere with some Windows NFS implementations Export Security Access mode Set Read only or Read Write access on the exported file system Trust remote users Determines which UNIX users on the client are trusted by the server Three options are available Everyone A
206. s You can even tweak them with simple Linux commands e g ping ce 3 Finally Maintenance gt Integrity supports EdgeWare s Configuration Backup and Restore capability which saves you the administrator from having to re enter major amounts of data by hand in case of big version or hardware changes SHUTDOWN Shutdown System turns the system off You will have to press the power button on the unit to get access to the UI again Reboot System restarts the system Wait several minutes until the server restarts and then refresh the browser to get access to the web UI again MONITORING The Monitoring module is intended for notifications of warning or errors in hardware and software components of the EdgeWare Storage system In addition it contains SNMP a major administration capability in its own right which can be connected to SNMP browsers to control your EdgeWare Storage Server among many other kinds of systems MONITORING STATUS This colorful page shows the current status of the EdgeWare Storage Server components that the monitoring program is monitoring It shows the component s status when it was last checked Page 118 and when the next check will occur At the bottom there appears a legend which shows different states of the service Which monitors are active at any given time is subject to your control in Watch Lists below The EdgeWare Storage Server comes with a minimal set of them active by default but
207. s also an option Deleting a snapshot schedule does not delete the snapshots it created but it does prevent any more snapshots from being created by this schedule Similarly a schedule that was merely recorded by selecting No for Enable Snapshot Scheduling can be triggered into action by changing that to a Yes and hitting Create or Create and Snapshot Now Selecting a snapshot link one of the ones that do not end in _schedule brings up a snapshot editing page This is very much like the snapshot creation page of Figure 33 except it allows Delete and it also allows the snapshot to be mounted Unlike regular Logical Volumes snapshots are NOT mounted by default For them to be useful you have to select them and mount them in this page Page 53 DELETING SNAPSHOTS Snapshots require attention If a snapshot is no longer serving a useful purpose delete it If a snapshot schedule has outlived its usefulness delete it and then delete the snapshots created by it Aging snapshots take up precious storage space they slow down their parent Logical Volumes and they have an annoying tendency to overflow Hint In some applications large numbers of snapshots are created by a schedule After deleting the schedule it is time consuming to delete the snapshots one by one A convenient method for deleting them all at once is to create a new snapshot schedule for the same parent volume with maximum count 1 activated but with no s
208. s as This option determines which local group untrusted client groups are treated as You may enter either a GID or select a group or choose the default typically the group nobody You can click the button next text field to get a pop up window that displays local groups defined on the system that you can provide as the anonymous group to which untrusted groups are mapped to After making any changes click save to commit the changes and then restart the NFS server from the NFS Exports page SERVER CLIENT ACCESS TO NFS This subhead gives a summary overview of the process of getting both sides of an NFS server client relationship working Here we assume the client is a Unix Linux command line system The client can also be BSD or other Unix like flavor the behavior of NFS is very standard and has changed little for decades A Connect server and client to the same local network carry out the instructions for Networking gt Network module above on the server and see that they ping each other The server side command is ping c 3 192 168 1 17 executed in Maintenance gt Command Line wait about 10 seconds The client side command is ping 192 168 1 158 followed by a Ctrl C as soon as data flow is seen to be happening Here in place of 192 168 1 17 put the real client IP address and in place of 192 168 1 158 put the real server IP address B Proceed to the server UI Due to the success of step A it should be possible to do this if y
209. s been kicked out of the RAID set Click on Remove Drive at the bottom Select the same drive again and click on Add Drive button below Next select any two of the healthy drives of the RAID set and also select the kicked out drive from earlier steps Click on Rebuild Unit button below RAID rebuild should start immediately Go to Home page to check the status of the rebuild process Specifying a hot spare The RAID Controller s give you the option to specify a hot spare from one of your available ports Hot spares should be selected after RAID1 or RAID5 creation Page 37 To specify a hot spare 1 While the system is running install the spare drive to an empty tray from the chassis and insert back Make sure the blue LED on the tray lights up If not make sure the drive tray is well seated in the drive bay and the tray lever is fully closed 2 Log into the 3ware manager 3 Click on the check box next to the available offline port in 3ware manager on the Configure page and then click on the Add Spare button NOTE Hot spare drives must be equal to or larger than the drives used for the redundant array If they are even a few bytes smaller they will fail In the Event of a Hard Drive Failure ALWAYS Verify the failed hard drive number Before removing a hard drive verify the location of the failed drive by checking the 3ware manager alerts and logs on the Cutting Edge Storage Server If an additional functionin
210. s contains all the configuration information about TCP IP networking LDAP and NIS and the four simple configurations only a single file or directory tree each Page 131 Scheduled_Jobs This contains cron schedules for backup snapshots and other timed activities Volume_Backup This contains all the Storage gt Backup information System_Time This contains time zone information Failover This contains all information needed for current failover settings NOTE Selecting Shares or Volume backs up the ACLs on the file systems also 2 3 4 After selecting the backup configuration types click on Process to backup the configurations A report will appear If it says some configurations failed that only means nothing was found presumably because your system has not initialized this feature It is also OK for files to be missing in complex configurations for a similar reason The key news is at the bottom Backup successful After seeing this click on the Download Backup link to save the Backup_conf tar file to another network location It gives you a popup and you normally store this tarball on your client system in a dedicated location It is advised that you change the default file name before saving for instance by inserting a date like Backup20080915_conf tar Restoring Storage Server Configuration IMPORTANT This operation overwrites current configuration information on your system Unless yo
211. s in accordance with the then current terms and conditions for their sale DOCUMENTATION Manufacturer will make available to Customer all published revisions or corrections to the documentation for the Covered Products that Manufacturer makes generally available to customers who have ordered Services for the Covered Products This documentation will be made available online on Manufacturer s support web site currently located at http Awww cuttedge com support php the Web Site or by any other means specified by Manufacturer TECHNICAL SUPPORT AND ERROR CORRECTIONS TECHNICAL ASSISTANCE SUPPORT CENTER During the hours stated for the applicable Support Package Support Hours Manufacturer will use commercially reasonable efforts to provide Customer with assistance to diagnose and resolve technical problems with the Covered Products whether hardware or software based through the Web Site by e mail by telephone at the email address and telephone number indicated on the Web Site or by any other means specified by Manufacturer MANUFACTURER SUPPORT WEB SITE Manufacturer will issue Customer a unique username and password to access the Web Site Customer will use the Web Site only for supporting its authorized use of the Covered Products and Covered Software Customer will maintain reasonable password security with respect to the username and password issued for the Web Site and will immediately report to Manufacturer any breach
212. s left unchecked the information will be persistent but the LDAP authentication will remain inactive This allows administrators to retain the LDAP configuration information Authenticate Samba with LDAP Check this box when Samba Windows Client users authenticate using the LDAP server and allows additional Samba configuration Checking this feature of LDAP unlocks the following Samba specific settings Samba specific settings User Suffix This specifies where users are added to the tree If no suffix is specified the value of the LDAP suffix will be used Usually we put ou people here Group Suffix This specifies the suffix that is used for groups when these are added to the LDAP directory If no suffix is specified the value of the LDAP suffix will be used Usually we put ou group here Machine Suffix This specifies where machines should be added to the LDAP tree If no suffix is specified the value of the LDAP suffix will be used Usually we leave this blank In order for Samba authentication to work properly the Samba schemas should be used by the LDAP server The Samba schemas are verified to work with EdgeWare s LDAP client Currently two versions of the Samba schema are available and tested The two schemas are typically called version 2 and version 3 If using OpenLDAP as your LDAP server the correct schema needs to be included in the LDAP server configuration For example the path to the samba schema being used path to
213. s of verification by clicking on the link Details within the Integrity module main page This page shows a summary of the result Figure 95 Page 128 es Deg KAREL A Server Storage Networking Authentication Maintenance Mansel Figure 95 Details page after running Verify 6 Clicking on the button Details will lead to a more detailed results page as shown in Figure 96 which shows a listing of changed and missing files It also shows logs and database files themselves which it is normal to see changed every time and notes some directories with changed files in them Figure 96 Part of detail listing of OS files changed after integrity verification was run The list from the detailed results of the integrity verification can be very long depending on how many changes occurred since the known database was generated From Figure 96 it is possible Page 129 to see both the access time and the SHA1 i e checksum differ between the two databases for the var db integrit cenas_new cdb file itself This is expected since the previous was probably your known database currently being compared and the new cenas_new cdb was generated just now ARCHIVE SAVE AND RESTORE The Archive component allows you to perform the following 1 Save the reference OS integrity database 2 System configuration backup and restore Archive the Reference Database Select the location to save
214. seconds The Windows client side command is ping 192 168 1 158 from within a command window Here in place of 192 168 1 17 put the real client IP address and in place of 192 168 1 158 put the real server IP address B Proceed to the server UI Due to the success of step A it should be possible to do this if you wish in a browser window on the client C Create a user on the server in Authentication gt Users and Groups gt Local Users and Groups gt Create New User Give this user the same name and password you want to use when connecting up from the client D Create a RAID array a volume group a logical volume and a share as described above in the Storage section E Set up CIFS as just described and make sure your step C user is placed in the Valid users set see Figure 74 by clicking on Browse and picking your user out of the resulting pop up and then hitting OK F Start or restart Samba servers G Proceed to the Windows client UI You may have to wait about 10 minutes for the new share to advertise on the network H Map the share to a drive On XP the procedure is to right click My Computer select Map Network Drive from the drop down menu and select a drive letter from the Drive list best leave the default drive letter At Folder hit Browse and select the new identifier share on server line format You will be prompted to enter the user and password Use the same one as in step C Then hit Finish when it
215. server documented above Page 144 4 es E e KRY A Storage Networking Authentication Maintenance Clustering Log Out Manual r EdgeWare Network Configuration Network s A Provides configuration options for the IP Network Settings O protocol like IP addresses and DHS client IS EdgeWare 3 0 0 1ce I J Networking N HELP daa e cIFS j _ foe wes Hostname Leader22 See Domain Name oo ts mo emp Ethernet 192 168 1 130 255 255 255 0 Yes EES eth1 Ethernet 10 168 1 131 255 255 255 0 Yes CH e Add New Interface allel ese z None or from DHCP O ae None or from DHCP O DNS 1921681101 Search None Listed servers domains swe Advanced Configuration gt Figure 111 Network main page after ethx configuration and new hostname Devise hostnames for each port of each server The LAN connected port hostname normally eth0 should be the same as the hostname of the server itself others have a hint of the port location Here is a typical set Leader22 Leader22 ethi Follow22 Follow22 ethi Only its main hostname should be entered in the Network module main page Hostname line of each server Figure 111 Hit Save A reboot Maintenance gt Shutdown is now required to change hostname and eth0 IP address and to make sure that DHCP goes away You will have to retarget your browser instances t
216. sible for the hardware connection to fail even if all the components are present and connected In such a case a defective component must be replaced Most client operating systems offer graphical indications of a good connection Some examples are given below other operating systems or versions use similar techniques CHECK FOR A GOOD CONNECTION ON MICROSOFT WINDOWS XP Select Control Panel in the Start menu Select Network and Internet Connections Under Pick a Control Panel icon select Network Connections A graphic headed with the words Local Area Connection or Local Area Connection 2 or similar should say Connected or Connected Firewalled or similar If the message is Limited or no connectivity the Windows machine does not have a communicating IP address Folllow instructions for setting client IP address in step B below If the network cable is hot unplugged from the client this should say Network cable unplugged or Network cable unplugged Firewalled or similar Plugging it back in should cause it to return to Connected CHECK FOR A GOOD CONNECTION ON MICROSOFT VISTA Select the Windows icon in the lower left corner Select Control Panel Select Network and Internet In the search box in the upper right type in Network connections and hit lt Enter gt A list will appear Under the subhead Network and Sharing Center find View Network Connections and select this A graphic headed with the words Local Area Connection or
217. sical volume details Physical Device RAID device AE Browse Warning All data on the selected device will be erased Enabled for allocation Yes O No Add to volume group gt Return to index Figure 28 Add Physical Volume to Volume Group VGstorage The expanded Volume Group tree will reappear with both old and new Physical Volumes visible not necessarily in the order of attachment To add more Physical Volumes repeat this procedure Physical Volumes must be added one at a time It may seem odd that this particular Volume Group is now strung together from completely dissimilar parts an IDE DOM solid state disk only 1 GB long and a big SCSI RAID Linear array presented by a 3Ware controller In fact such combinations have value in certain high speed scientific applications This is an illustration of the power of virtualization ADDING A LOGICAL VOLUME Once enough Physical Volumes have been assembled into your Volume Group the next step is to create a Logical Volume Your Volume Group is a generic line of storage Your Logical Volumes will have specific and usually different roles to play even though they may come from the same Volume Group This step is very much like partitioning a disk Think ahead about how many Logical Volumes you are likely to need and what they will be doing because once you allocate a size to a Logical Volume and build the superstructure that makes it useful to Page 45 your storage
218. solution order This series of drop down lists give you the ability to define the search order for hosts on network from the network resolution protocol databases Hostname to IP Domain Network The next The Light Weight The local address mapping from Name Information generation Directory Access name service Host Addresses System System NIS Protocol database DNS servers The DNS servers to query for DNS lookups Enter the DNS server IP addresses in order of importance Search Domains Domain suffix search order Host Addresses If you do not have a reliable host name lookup service such as DNS or NIS running in your intranet or you are having difficulty in getting hostnames resolved from the appliance you can enter hostname to IP address mappings to get faster lookups Just click on Add a New Host Address Enter the IP address and then enter the hostname s for that IP address Then click Create to commit the new entry The Host Addresses capability is a ways used when setting up a Handshake between two systems such as is required for Clustering capabilities such as Failover and Synchronous Mirroring For more details see the Clustering section below in this manual EdgeWare Network Configuration Network Res EF SE pions forthe iP 7 rovides configuration options for the Advanced Network Configuration A protocol like IP addresses and DNS client a EdgeWare 3 0 0 1ce K J HELP Host Addresses ET
219. sportation off site and for some operations even regular replacement of whole systems But they also include virtual backup support such as a policy for backup naming and placement of the necessary backup volumes on storage Finally the third principle of backup design is as far as possible to make a current copy before restoring from backup EdgeWare supports this through the Live Copy and Backup Now capabilities Your job is to make sure there is space to place them and a well known policy to do this If your current data is utterly destroyed then no such copy is possible But it is much more common that a part of your data has been messed up a version overwritten or somebody s work lost The steps to a safe restore are then Determine exactly what the restore is going to cover Make a copy of that part of the damaged current data and place it in a safe place Do the restore over the data area determined in step 1 Compare the results of your restore with your knowledge of the original current data ES E It is entirely too common that a restore does far more damage than the original data loss Don t let it happen to you ADDING A NEW BACKUP OF A VOLUME To start a new backup 1 Select the logical volume you want to backup from the drop down menu as shown in Figure 42 Page 60 Figure 42 Backup 2 Click on Add a new backup of a volume button You should get the page shown in Figure 43
220. ste buttons E Pop ups temporarily allowed To always allow pop ups From this site click here x E d H e wW are Server Storage Maintenance Clustering Log Out bat Networkin g Authentication l ol Upload Copy cut Pas Info x Delete N Find Refresh Networking shares shares sharetest Ce E Binfstest TT Network Bl sharetest Name Size User Group Date cre aquota user 0 B root root 03 Sep lockheed pdf 240 KB 1dickson localevery 10 Jun m mount out 531 B ldickson 20001 04 Sep SS NFS sample manual good png 53 kB root root 17 27 w AppleTalk ts ees Ki aile JS Figure 80 File Manager ready for Download REMOTE REPLICATION ASYNC In this module you can start or stop the Asynchronous Replication source daemon and provide optional settings for each share that has async enabled This page allows you to set options to shares on your local EdgeWare NAS that are to be asynchronously replicated to another EdgeWare NAS Often such shares are in snapshots which have to be mounted before an enabled share can even be defined on them This module does the work analogous to the other protocols in the Network section but does not complete the job of setting up asynchronous replication That requires another EdgeWare system to present a target volume to accept the replicated data See the Clustering section below in this manual for comp
221. step is skipped if revalidate yes for this service in the etc samba smb conf file Step 5 If a user field is given in the etc samba smb conf file for the service and the client has supplied a password and that password matches according to the UNIX system s password checking with one of the usernames from the user field then the connection is made as the username in the user line Page 94 e Step 6 If the service is a guest service in etc samba smb conf guest ok Yes or guest only Yes then a connection is made as the username given in the guest account for the service irrespective of the supplied password One consequence of this security mode is that you are not required to make a UNIX local account for each Windows account you expect to connect to your Samba server Security user level This is the default level and the most common It is simpler than share level When the server tells the client that he runs in user level security the client will first send a command containing a username and a password At this stage of the connection negotiation the server has no idea about which service the client wants to access So it has to base its authentication procedure only on the pair username password or on the machine name Once the access has been granted to the client that client can connect to any share without re supplying a password or a username password pair The user s permissions on different shares
222. system Because of these facts failover is most friendly to storage protocols with highly stable setups such as iSCSI which is completely stable because it is a block level protocol If volume changes are made to an iSCSI failover system a reboot may be needed to make the Secondary aware of the new volume structure AUTHENTICATION KEYS The authentication keys option is to configure the type of authentication for heartbeat events Clicking on the Authentication Keys button takes you to a new screen Figure 127 where you can select the type of authentication Three kinds of authentication modes are provided e CRC There is no security in this mode SHA1 Security is provided by password authentication e MDS Security is provided by password authentication The two actions in this screen are e Save To save the authentication mode e Return to Index Return to the main page of IP Failover without saving Page 157 4 EN ZS g ZS Gel A Storage H uthentication Matten ance Clusterin g Log Out Maral EdgeWare Clustering Configuration Heartbeat Provlilom configuration options for IP Authentication Keys f A address and service failover 9 EdgeWare 3 0 0 1ce L l La Cluster node authentication Authentication mode CRC No security SHA1 with password MDS with password Save nr Return to index Figure 127 IP Failover gt Authentication Keys page FAIL EVENTS SWAPPING EDITING
223. t is also used if dual_log_enable is set to YES Default var log xferlog DENIED FTP USERS This page lists users or system accounts that are a ways denied authentication by the ftp server FTP CLIENT CONNECTION This protocol can use any client Windows Linux Unix MAC First carry out step A through step F of Server Client Access to CIFS above in this manual except in place of step E set up FTP as described above and in place of step F start or restart the FTP server Now proceed as follows G Proceed to the client UI H Open a command window Page 105 I Type ftp server where in place of server you put the hostname or IP address of the EdgeWare server J As prompted enter the username and password that you set up in the EdgeWare UI in step C The username alone should suffice here without being preceded by the IP address K It should announce your login was successful Enter a dir command to see your list of shares which should include the sharename you just set up Notice that the path root shares is omitted L Enter your share with cd sharename where in place of sharename you use the actual name of your share An example would be cd sharestorage compare Figure 73 Use put and get commands to move data back and forth between your current local directory and the share Other FTP commands are listed by help It may also be possible to download FTP files through a browser if anonymous
224. t to the interface information EXPORT CONTROL Manufacturer s standard Product incorporates cryptographic software Licensee agrees to comply with the Export Administration Act the Export Control Act all regulations promulgated under such Acts and all other US government regulations relating to the export of technical data and equipment and products produced there from which are applicable to Licensee In countries other than the US Licensee agrees to comply with the local regulations regarding importing exporting or using cryptographic software U S GOVERNMENT RESTRICTED RIGHTS If the Product is licensed for use by the United States or for use in the performance of a United States government prime contract or subcontract you agree that the Product is delivered as i commercial computer software as defined in DFARS 252 227 7013 Rights in Technical Data Noncommercial Items Oct 1988 DFARS 252 227 7014 Rights in Noncommercial Computer Software and Noncommercial Computer Software Documentation Jun 1995 and DFARS 252 211 7015 Technical Data commercial Items May 1991 ii as a commercial item as defined in FAR 2 101 or iii as restricted computer software as defined in FAR 52 227 19 Commercial Computer Software Restricted Rights Jun 1987 whichever is applicable The use duplication and disclosure of the Product by the Department of Defense shall be subject to the terms and conditions set forth in this Agreement as provide
225. target and then hit the Stop button But leave iSCSI running on the Primary with zero connections to all targets III SET UP THE PRIMARY AND SECONDARY Return to Clustering gt Failover Create Modify Resources defines or edits the failover configuration Enter data here only in the leader Primary UI Hosts in cluster Select the secondary LAN host i e the one that will acquire the virtual IP address Ethernet heartbeat interfaces Default is none but check the second radio button which shows the eth1 IP address Failover IP address Virtual IP address step IV Failover resources Under Failover resources we get two options step IV e Shared Storage Checked when the server shares an array e Block Devices Not used IP address for ping checks This is for a ping server a Tertiary system neither leader nor follower for special purpose network failover Normally leave it blank In step III we set up the first two of these iz es E e J SS ZS A Ed ge Ware Server Storage Networking Authentication D Log Out User jaintenance Clustering Geen EdgeWare Clustering Configuration Heartbeat Heartbeat Monitor H e Provides configuration options for IP address and service failover EdgeWare 3 0 0 1ce L HELP No cluster resources have been defined on your system Create ModifyResource To define a failover resource Remote Authentication Keys To configure the type of authentication for
226. taxes duties or other costs for international shipments CHARGES Replacement hardware products or parts will be provided to Customer at no cost if the malfunction is caused by normal wear and tear of the Covered Product and not Cutting Edge 1825 Gillespie Way Suite 100 El Cajon CA 92020 800 257 1666 www cuttedge com rd CUTTING EDGE ot ee a a ee A A G E R dee to causes external to the Covered Product Otherwise Customer will pay Manufacturer s then current list price for the replacement LIMITATIONS ON SERVICES SERVICES NOT COVERED Notwithstanding anything to the contrary in this Agreement the Services do not include the correction of and Manufacturer will have no obligation responsibility or liability with respect to any errors defects or other problems caused by or resulting from a Customer s failure to implement any Maintenance Release or Software Update made available to Customer by Manufacturer b changes by Customer or third parties Manufacturer to diagnose reproducible Errors to an operating system network configuration or environment that adversely affect the Covered Products c any alterations or modifications of or additions to the Covered Products made by parties other than Manufacturer d use of the Covered Products in a manner for which they were not designed or other than as specified in the applicable documentation or specifications e the combination use or interconnection of the Cove
227. ted schedule Hourly be Day Date Month Year Hour Monday 18 e August A 2008 O SE 34 Mi 32 Mm Current Time Zone is Not Set GMT 10 Hawaii GMT 09 Alaska AE EIS GMT 08 Pacific Time US amp Canada GMT 07 Mountain Time US amp Canad GMT 07 Mazatlan GMT 07 Arizona a K3 Figure 5 System Time setting page System Time refers to two different time settings both can be configured to synchronize with a Network Time Protocol NTP server They are shown as one except that the hardware clock is mentioned The two differ in that the system time lives in software only as long as the operating system is active i e as long as the server is powered on and the Linux OS is active The hardware time is the reflection of the hardware NVRAM BIOS time Figure 5 shows the System Time UI Page 21 Synchronization of time is important not only for keeping track of actual dates and times on files but also for proper mirroring and clustering setup In a cluster even a simple mirroring cluster of two time differences may keep the proper things from happening or at least happening as expected This is where NTP servers are most useful You can specify a network time server on this page for synchronizing the system and or hardware time with that time server You can find a list of public NTP servers at the website mentioned below http ntp isc org bin view Servers NTPPoolServers MANAGEMEN
228. ters stated in the user documentation shipped with the Product Manufacturer s limited software warranty does not apply to software corrections or upgrades INFRINGEMENT INDEMNITY Manufacturer will at its expense defend any suit brought against Licensee based upon a claim that the Product as delivered by Manufacturer directly infringes a valid patent or copyright Manufacturer will pay costs and damages finally awarded against Licensee directly attributable to any such claim but only on condition that a Manufacturer is notified promptly in writing of such claim by Licensee b Manufacturer has sole control of the defense and settlement negotiations c Licensee provides Manufacturer all information and communications received by Licensee concerning such claim and d Licensee provides reasonable assistance to Manufacturer when requested Manufacturer will have the right at its option and expense i to obtain for Licensee rights to use the Product ii to replace or modify the Product so it becomes non infringing or iii to accept return of the Product in exchange or for a credit not to exceed the purchase price paid by Licensee for such Product The foregoing subject to the following restrictions states the exclusive liability of Manufacturer to Licensee concerning infringement RESTRICTIONS Manufacturer will have no liability for any claim of infringement based on i use of a superseded or altered release of the Product ii use of
229. the LAN and should be the same on both ends of the cable If your setup is more complicated than the one described here for instance involving a high speed switch that is accessible from the LAN then you may need to specify your high speed port as a router on the module main page Default Route Device being its IP address in order to avoid having your fast traffic routed through eth0 The point to point eth1 connection does not require this router setting Hit Save and Apply for eth1 Figure 110 and Save for eth0 This insures that all IP addresses are permanent across reboot The new IP address for eth0 will not take over until reboot Ed ge Ware Server Storage Networking Authentication H Ch 9 Log Out User aintenance Clusterin Manual q EdgeWare Network Configuration Network if Frowkdee congas options Tor tie ip i G Per mees Setup A protocol like IP addresses and DNS client Networking igeWare 3 U U 1ce WR Network HEF Gad SS Boot Time Interface Parameters Type Standard ag ie Name ethi MAC Address 00 16 56 80 A4 OF ea EN IP Assignment DHCP CH BOOTP Static OP eg IP Address up Le FTP Netmask 255 255 255 0 At boot Yes O No Q sae F Advanced Configuration gt Save and Apply DIE RED a Figure 110 Network eth1 configuration D IMPOSE HOSTNAMES FOR EACH PORT OF EACH SERVER This is continued in the Networking gt Network module main page of each
230. the Product in combination with equipment or software not supplied or specified by Manufacturer in the Product documentation where the Product would not itself be infringing iii use of the Product in an application or environment not described in the Product documentation or iv Product that has been altered or modified in any way by anyone other than Manufacturer or according to Manufacturer s instructions DISCLAIMER LIMITATION OF REMEDY AND LIABILITY PRECAUTIONS EXCEPT FOR THE WARRANTIES SPECIFICALLY DESCRIBED HEREIN Manufacturer AND ITS THIRD PARTY LICENSORS DISCLAIM ANY AND ALL WARRANTIES AND GUARANTEES EXPRESS IMPLIED OR OTHERWISE WITH RESPECT TO THE PRODUCT SPECIFICATIONS SUPPORT OR SERVICES DELIVERED HEREUNDER Cutting Edge 1825 Gillespie Way Suite 100 El Cajon CA 92020 800 257 1666 www cuttedge com r CUTTING EDGE te NA BR A G E R INCLUDING BUT NOT LIMITED TO THE WARRANTY OF MERCHANTABILITY AND THE WARRANTY OF FITNESS FOR A PARTICULAR PURPOSE NEITHER MANUFACTURER NOR ITS THIRD PARTY LICENSORS HAVE AUTHORIZED ANYONE TO MAKE ANY REPRESENTATIONS OR WARRANTIES OTHER THAN AS PROVIDED ABOVE THE COLLECTIVE LIABILITY OF Manufacturer AND ITS THIRD PARTY LICENSORS UNDER THIS LICENSE WILL BE LIMITED TO THE AMOUNT PAID FOR THE PRODUCT NEITHER MANUFACTURER AND ITS THIRD PARTY LICENSORS NOR LICENSEE WILL HAVE ANY OBLIGATION OR LIABILITY WHETHER ARISING IN CONTRACT INCLUDING WARRANTY TORT INCLUDING CTIVE PASSI
231. the mirrored pair is totally consistent It is OK to proceed to step V before this is complete but for the sake of full redundancy you may wish to wait until it is complete E d g 6 wW are Server ion Maintenance Clustering Log Out Storage Networking Authenticati el Manual EdgeWare Clustering Configuration Remote Mirroring i d O Provides block device replication over the N Resource LVsto rage network to another server EdgeWare 3 0 0 1ce i F EP NAS Servers HELP E index a JU _ Remote plication Leader Mirror state SyncSource Localhost Localhost status Primary UpToDate Follow22 Remote host Virtual Device Remote host status Secondary Virtual device Inconsistent Idew drbd0 Mirroring SEL Synchronization in process e Een Underlying Se Underlying disk Disk Synchronization LVstorage mirror 62 424 46 012 MB LVstorage speed dey Time left 0 41 33 Mount point Maurt paint done 21 s eg Make Local Secondary Delete Mirror L vPetumn to index Figure 139 Remote Mirroring Status Page EXPORTING THE MIRROR AS A SHARE After remote mirroring starts syncing you can take that mirrored volume and export it as share or as an iSCSI Target This subhead will cover how to export a mirrored volume as a share where you can map it as a network drive from your local workstation Page 169 V FORMAT OR PROMOTE MIRROR VOLUME TO NAS STATE 1 Open the Storage gt Volume Manager module mai
232. the page not shown in Figure 9 to apply your selection Under each administrative user configuration page as show in Figure 6 you can also restrict or allow access to a web UI administrator for each configurable module in the UI by clearing or checking the check box next to each module name Alternatively the other recommended path is to add a new GUI user say administer This administrator does not have to be a local User on the EdgeWare system but can exist solely for administration through the web browser One does not need to delete the default admin administrator though it is a good idea to change its default password setup The root administrator is not accessible by browser but has access to certain extra modules and should also not be deleted Page 23 E d g e W a re Server Storage Networking Authentication Maintenance Clustering P Log Out User Manual _ EdgeWare 3 0 0 1ce EdgeWare Server Configuration Administrators Administrators EdgeWare Users r P Contains access control configuration and option pages for the user D iSCSI Global ACL Administrators bau AppleTalk Backup CES Cluster Servers Cluster Shell Cluster Users Command Line ETE Frox FTP Proxy Heartbeat Monitar Http Integrity Kerberos5 LDAP Users and Groups Log Log File Rotation Management Mirroring Monitoring NAS Servers Index NES TE beet Ee
233. tic Manual QuoteAdmin QuoteMaster v Figure 9 Create GUI User and make it a member of a group from Member of group Page 25 VIEW LOGIN SESSIONS Here you can view all the current logons made by administrative users of the UI and also the configuration changes made in each login session so far Figure 10 dd es D e g BY A Storage Networking Authentication Maintenance Clustering Log Out User Manual EdgeWare Server Configuration Administrators if Current Login Sessions _ EdgeWare 3 0 0 1ce Contains access control configuration and option pages for the user 4 Abffb01873 1 cdB07 acceac3e5160132 admin k Mon Aug 18 09 07 37 2008 View logs saam Figure 10 Viewing current login sessions You can view the logs recorded for a session by clicking View logs hyperlink on the right of each session To disconnect a session click on the Session ID hyperlink of each session Page 26 STORAGE SECTION The primary use of the EdgeWare Storage system is for block or file storage There are four primary configuration steps to allow access to disk storage space 1 Configure the RAID set typically done at install 2 Make a partition on the physical disk if required in Partitions 3 Virtualize the storage with the Volume Manager and create logical volumes 4 Export the logical volumes using the Share Manager NAS or activate a target iSCSI The term mass storag
234. tion ete DO RAID DEE In volume group VGstorage Manager snare manager Logical volume details Kegel Volume name LY storage Volume size 152 S Storage allocated from volume group GB 152 out of 152 Allocation unitsize 4096 MB mmm Device file dev VGstorage LVstorage Current status Not in use Backi ue Save gt Delete gt CS ses SS Mount the logical volume Press the mount button to mount the current filesystem type on the logical volume Select a filesystem type and click this Ch der 2 button to create a new filesystem on we SGskrerget SCS this logical volume This will erase any data currently on the volurne Return to index Figure 132 Configuring Logical Volume iSCSI case Page 165 III CONFIGURE THE MIRROR The steps that follow should be carried out only on one local server We will follow a push strategy and set the Primary server to be the local server The RPC programming of drbd takes care of the other server 5 Enter the Clustering gt Remote Mirroring module and then click on the Configure Mirror button see Figure 133 E d g e W are Server Storage Networking Authentication Maintenance g Log Out ME ie e E S OE SEO Iren Provides block device replication over the A g network to another server 2 EdgeWare 3 0 0 1ce HELP Resource Remote Host Local Volume Bonge Volume Usage Configure Mirror Figure 133 Remote Mirroring mod
235. to add change or delete available Services or Support Packages for certain hardware or software platforms and configurations in its reasonable discretion upon 90 days notice If Manufacturer deletes an entire Support Package at any time during the term of this Agreement Manufacturer will refund to Customer a pro rata portion of the Annual Fee paid by Customer for the portion of applicable Services not rendered for the remainder of the term due to the deletion of the Support Package MAINTENANCE RELEASES SOFTWARE UPDATES AND SOFTWARE UPGRADES Maintenance Release is a set of related or unrelated bug fixes that Manufacturer makes generally available to customers who have purchased a Support Package that includes Maintenance Release Services Maintenance Releases are typically labeled with a Cutting Edge 1825 Gillespie Way Suite 100 El Cajon CA 92020 800 257 1666 www cuttedge com r y CUTTING EDGE oe se ONRAAD A G E R Eet in the version number after the second decimal e g 1 0 0 to 1 0 1 although not each change may deemed be a Maintenance Release For as long as Maintenance Release Service is in effect under the Support Package selected by Customer as specified in Exhibit A Manufacturer will make available Maintenance Releases for the version of the Manufacturer software included in or with Covered Products Covered Software until the release of the next Software Upgrade Maintenance Releases will be de
236. to delete and recreate hardware RAID since doing so would render the appliance unbootable Current systems have the Operating System residing on a local system disk or mirrored disk pair or on a solid state disk or DOM This isolates the hardware RAID from the operating system dependency and gives flexibility and total control of the hardware RAID storage to the customer A oca mirrored disk pair will appear as SOFTWARE RAID and should not be disturbed as it functions as the system disk Page 34 Storage DO RAD Volume Manager es E eg HBS A Server Storage Networking Authentication aintenance Clustering Log Out User EdgeWare Storage Configuration Raid Configuration E EE 3Ware Disk Set Management O options for software and hardware RAID EdgeWare 3 0 0 1ce 3 HELP 3Ware Hardware RAID Optimize 3ware Performance Select this checkbox to apply enhance settings to optimize 3ware performance NOTE Uncheck if you experience system instability CI Enable Optimized 3ware Settings Create Maintain RAID Sets Create Hardware RAID device of tevel Striped PAIDO Start sware Manager Available Ports Disk L6OL4AZSH 279 48 GB at port 2 on card 6 Disk B62137HH 279 48 GB at port 3 on card 6 er Es Figure 19 3ware devices summary page Figure 19 shows the summary page for 3ware in the case where no arrays have been created If you press Start 3Ware Manager you should g
237. tructured as a pull the active primary system A is the target where the copy is maintained of the source share on secondary system B e It is not storage symmetrical The target is a mounted volume and the source is a share e It is in the source or secondary system where the share gets tagged with Async as one of the Enabled Protocols in Storage gt Share Manager There also the replication support or rsync service must be started in Networking gt Remote Replication NOTE The destination volume on EdgeWare system A should be exclusively created for remote replication as any data on the destination volume that is not present on the source share is deleted by the remote replication process Setting up Remote Replication Because of the three points mentioned in the above paragraphs it is particularly desirable to have both the local target and remote source server open independently in different browser instances A considerable amount of setup has to be done on both and the slowness of RPC plus the confusion of jumping back and forth in one browser instance can be frustrating if you try to do everything within one server s NAS Servers Index The deployment steps are as follows A Source Set up the share with Async enabled Page 138 Source Activate the remote replication service on the share Target Set up a volume for sole use as replication target Target Capture source in Host Address
238. u are upgrading or recovering from scratch it is best to back up your current configuration to a backup file with a different name before restoring 1 2 4 Click on Restore Select the configuration types you want to restore Click on Upload backup files link and select the Backup_conf tar file that you saved when you backed up the desired configuration If you omit this step EdgeWare will restore the configuration from the last configuration backup file present in the Storage Server Click on the Process button NOTE After clicking on the Process button the configuration will be restored and the Storage Server will be rebooted within 10 seconds Page 132 UPDATES EdgeWare has the ability to download software updates via the Internet This requires http access to the updates server which is updates cuttedge com The Storage Server can be updated by clicking Maintenance from the top icon bar and then Updates on the left side You will get a module main page labeled Check for Edgeware Updates with a single button Check If you select this button EdgeWare will contact the updates server and check for available software updates This may take a up to a minute depending on the speed of your Internet connection If any updates are available they will be listed in the Updates list window of which an example is shown in Figure 99 EdgeWare Update Edgeware IS Edgeware Update EE RE Spa Se eatches for an
239. u may lose some detail output and just get a Finished screen but that does not affect the success of the update Note The EdgeWare 4 0 Updates page is mot a one size fits all approach to our customers Each installed Storage Server registered with Cutting Edge connects to an entry in our update database so that it will receive all the updates appropriate for that installation and no others This system has been specifically designed to protect your installation from unsuitable updates that are intended for other dissimilar installations RE INSTALLING EDGEWARE It may prove appropriate at some future date to re install an entire new version of EdgeWare from a CDROM or by other techniques If this is done be sure that all build processes such as RAID arrays replications or mirrorings are complete and not still underway If a percent less than 100 is displayed it is a sign that you should wait If all builds are complete when the re install is done your partitions if any RAID arrays and volumes including your logical volumes will still be there after the re install This is because information about them is saved on superblocks in the mass storage devices themselves Information on the system disk such as shares and local users will not survive the re install To avoid having to re enter all such information by hand use the EdgeWare capability of Configuration Backup described above in Maintenance gt Integrity
240. uld be Stop 3Ware Manager 3 Point your web browser to https Storage Server IP address 10001 Select Administrator from the drop down list of users Default password is amcc Page 36 NOTE Only run the 3ware Manager when you have to perform configuration changes for example RAID rebuild to the RAID sets When done stop the 3ware manager aware 3DM 2 Loekheeaserver Linux 2 6 20 40cenassmp Administrator logged in Summary Information Management Monitor 3DM 2 Settings Help Refresh Summary Administrator now logged in Controller Summary Serial GE 6 9650SE BLPML L32602457 460446 FE9X 3 08 00 016 2 26 08 003 2 6 20 OK Last updated Tue Aug 19 2008 10 36 48AM This page will automatically refresh every 5 minute s 3DM 2 version 2 06 00 007 API version 2 03 00 006 Copyright 1997 2007 AMCC All rights reserved Figure 22 3ware Manager What to do first when a RAID set is degraded Sometimes due to heavy load or unclean shutdown a RAID set may kick out a member drive and cause array degraded state If a RAID5 array degrades and does not start initialization automatically after the unit is running for 20 minutes you may have to re introduce the kicked out drive back into the RAID5 array To do this Lo AMPER ME Start the 3ware Manager from GUI Login to the 3ware Manager at https IP address 10001 as Administrator Default password is amcc Go to the Configure page Select the drive that ha
241. ule main page 6 Select the EdgeWare server that will mirror the host see Figure 134 use the LAN port name not the eth1 port name and then hit Next EdgeWare Clustering Configuration Mirroring ae 2 2 Provides block device replication over the tA Setup Synchronous Mirror O network to another server EdgeWare 3 0 0 1ce L 3 HELP Available EdgeWare Servers Follow22 eth1 Follow22 Next 4Return to index Figure 134 Configure Mirror first page Server Selection 7 Select the Local Host as the primary host see Figure 135 Page 166 8 10 11 Configure mirror settings network to another server G EdgeWare Clustering Configuration Remote Mirroring f d A SS 7 O Provides block device replication over the AJ EdgeWare 3 0 0 1ce K HELP Required Settings Server Settings Remote Host Primary Local Host Primary Hosts Available NICs eth1 10 168 1 133 e eth1 10 168 1 131 Avalable Volumes L Vstoraqe mirror ia LVstorage Next Return to index Figure 135 Configure Mirror second page after selections Select a NIC on both hosts from Available NICs that you want to be used to transfer the Sync data packets keeping in mind that remote mirroring uses up a lot of network bandwidth especially during the initial setup If you have NICs on more than one subnet you should choose the subnet with more bandwidth and less network traffic In Fi
242. ult 120 Time to wait for Authentication Specifies the maximum amount of time a client using active mode has to respond to a data connection in seconds Default 60 Log Enable When enabled in conjunction with xferlog_enable and with xferlog_std_format set to NO all FTP commands and responses are logged This directive is useful for debugging Default Yes Xferlog Enable When enabled vsftpd logs connections vsftpd format only and file transfer information to the log file specified in the vsftpd_log file directive var log vsftpd log by default If xferlog_std_format is set to YES file transfer information is logged but connections are not and the log file specified in xferlog_file var log xferlog by default is used instead It is important to note that both log files and log formats are used if dual_log_enable is set to YES Default Yes Note in Red Hat Enterprise Linux the value is set to Yes Log File Specifies the vsftpd log file For this file to be used xferlog_enable must be enabled and xferlog_std_format must either be set to NO or if xferlog_std_format is set to YES dual_log_enable must be enabled It is important to note that if syslog_enable is set to YES the system log is used instead of the file specified in this directive Default var log vsftpd log Xferlog File Specifies the wu ftpd compatible log file For this file to be used xferlog_enable must be enabled and xferlog_std_format must be set to YES I
243. ume Manager page and the new Logical Volume will appear in the tree Figure 30 Page 46 A A ee J SS A A E d ge W are Server Storage Networking e aaar Authentication Maintenance Clustering Log Out Masidi Provides logical storage management functions like logical volume creation A physical volume aggregation and snapshot creation etc d EdgeWare Storage Configuration Volume Management b Create Logical Volume Storage EdgeWare 3 0 0 1ce G ss HELP Er In volume group VGstorage oo Manager E stare manager Logical volume details ST Volume name LVstorage Volume size 250 Ss ees Storage allocated from volume group GB 0 out of 559 Allocation unit size 4096 MB Network File Protocol CIFS AFP and or UNIX Mixed g a Backup Se Stripe across physical volumes No O Yes GS sen Create VW 4Return to index HELP Please wait a few moments to create and format LVstorage mkfs returned xfs by mkfs xfs i size 2048 I sunit 8 f dev VGstorage LVstorage Logical volume details Volume name LVstorage Volume size 250 100 available Storage allocated from volume Allocation unit group GB 250 out of 559 GE 4096 MB Device file dev VGstorage LVstorage Current status Me Ci ARSE SUNRISE Volume access Read write Type linear This logical volume cannot be renamed or resized as it is currently in use disable mounting Unmountthe logical volume Press the unmount button to
244. ume page you are presented with a Schedule Snapshot Creation page as shown in Figure 34 PF an ae ZS SZ A Ed H eWare Server Storage Maintenance Clustering Log Out ae Networl king Authentication Manual Provides logical storage management functions like logical volume creation physical volume aggregation and snapshot creation etc S EdgeWare Storage Configuration Volume Management b Schedule Snapshot Creation O Storage EdgeWare 3 0 0 1ce S We Manager Schedule Selection IE Enable Snapshot Scheduling No O Yes 9 Maximum Concurrent Snapshots 2 GG woes Notification Email Reminder Notes ee ee Job schedule Every 10 minutes v CS iscsi Z Create Every half hour reate Return to index Figure 34 Snapshot creation scheduling 1 Enter your choices for the options Enable Snapshot Scheduling This checkbox allows you to select whether this scheduling should be active or not No means it will get recorded and you can activate it later Yes means it is recorded and active now Page 52 Maximum Concurrent Snapshots The scheduler will not exceed this number of active snapshots at one time It kills the oldest when needed to ensure this A ways enter a reasonable small number here Notification Email This email address is where email notification is sent when snapshots are created Reminder Notes Text here is sent out in the email notification or entered in
245. under Archive A re install does not affect data on the logical volumes of your mass storage whether they are NAS or iSCSI If you have difficulty accessing this data after the re install contact Cutting Edge Technical Support Do not attempt to recreate your arrays or volumes from scratch as this will result in lost data Page 134 CLUSTERING The Clustering section allows your EdgeWare Storage Server to discover other EdgeWare Storage Servers on the local network and to deploy features which require two or more intercommunicating EdgeWare systems These include Failover synchronous Mirroring and asynchronous Replication The generic series of steps to follow is 1 Establish contact in the GUI among local Storage Servers 2 Ensure that at least two of these have the same DNS and etc hosts file information i e use Host Addresses icon under the Network module 3 Form an authenticated peer to peer relationship handshake between these so that each can access the other through GUI or perform EdgeWare programming on the other 4 Configure remote replication async remote mirroring sync or failover using these In the case of remote replication steps 2 and 3 need not be a complete handshake as the pull is a one sided relationship NAS SERVERS INDEX The NAS Servers Index module allows administrators to contact the GUI managed servers on a local subnet either by broadcast or by specifying them one by
246. unmount the logical volume This action may fail if shares exist on this volume or other processes are using the mounted volume Figure 29 Logical volume LVstorage is created Not all available storage is taken DELETING A LOGICAL VOLUME In order to delete a logical volume you have to make sure that there are no shares defined on that logical volume from Storage gt Share Manager In addition if it is iSCSI it must be removed from the iSCSI Targets list in Storage gt iSCSI Any other uses such as asynchronous replication or synchronous mirroring must also be disconnected After this is done proceed as follows Expand the volume tree for the volume group containing the logical volume Figure 30 Click on the logical volume link An Edit Logical Volume page appears with Logical volume details box bottom of Figure 29 Click Unmount the logical volume A result announcement page appears Click on Return to Index The Edit Logical Volume page returns with Delete button Figure 31 Click Delete on that page and Delete now on the resulting confirmation page Hit Return to Index on the result announcement page and your LV should be gone pl tsi eer Page 47 da Ki 1 Cl amp A aa VGsystem ad Figure 30 Volume Manager page with Logical Volume link dd SS 2 kA S amp A Figure 31 Edit Logical Volume page unmounted ready to delete RESIZING A LOGICAL VOLUME
247. ur Certificate Authority CA enter the CA information and paste the certificate given to you by the CA ADMINISTRATORS MODULE This module provides the functionality to create administrative users and groups for the web UI These administrative users are different from and more powerful than the standard users dealt with in the Authentication section further on in this manual This module also offers the administrator the capability of viewing the current logons and administrative changes GUI ADMINISTRATIVE USERS This component contains access control configuration for the GUI user interface administrative users Inside the Administrators module the administrative users and groups have pull down list boxes directing the page to list the selected member s module access list You can then select to edit the GUI administrators properties using the hyperlink Edit User You then can select a higher level of security using the select list next to Password as shown in Figure 9 If you wish to keep the GUI root administrator in place and want the password synchronized with the SSH secure shell access from other systems then select Unix Authentication for the Password parameter The new password is now the default Unix Authentication password 4linux If you wish to keep the GUI root administrator s password separate from the SSH and UNIX password then select Set to and enter a new value You must then select Save at the bottom of
248. users and then permit it to be filled with real data it is not easy to trim back or make room for later needs Among the possible uses for Logical Volume space in the Volume Group are NAS protocols which offer File System storage to your users Asynchronous replication target which offers File System backup storage to an asynchronous replication source on another system iSCSI target which offers block level storage to your iSCSI initiator users Raw source or target for synchronous mirroring the source is later built up in one of the previously mentioned ways but the target on another system is left raw and should be the same size as the source Snapshot Volume which has a special relationship to a Logical Volume of one of the previous types It gets a separate tree branch and will be treated later in this manual Remember to leave room for Snapshot Volumes in the Volume Group if you are intending to use them because each Snapshot Volume must be in the same Volume Group as its parent Logical Volume After planning the use of the storage in your Volume Group create each of your Logical Volumes in this way Pd ME 10 If necessary click on the Volume Group to expand it Click on Add a Logical Volume You will be presented with a Logical Volume creation page Specify a name for the logical volume Give the amount of storage you want to allocate to the logical volume in Volume Size Storage allocated from volume group G
249. users and groups exist and are unalterable in this UI Then you must deal with them on their native server It is possible to have the total set of users and groups be the combination of those presented by several of these authentication services In this case be careful to distinguish ones with similar names and avoid having different users whose names are distinguishable only by case such as susie Susie and SUSIE Some protocols distinguish by case others do not this leads to partial setups which are difficult to clean up Local Users and Groups Windows Domain Users and NIS Domain Users LDAP Users and Groups Groups Figure 82 Users and Groups module selections Local Users and Groups Local users and groups are the users and groups that are defined on the NAS itself and are authenticated by the NAS operating system kernel The main page shows the currently present local users and groups CREATE A NEW LOCAL USER 1 Click on the Create New User link You will be presented with a page for creating a user as shown in Figure 83 Page 109 Or eS AR 10 11 12 User Details Username user Real name User Number Six Password eoccce Password Confirm coocoo User ID 20002 Password Options Password changed Never Expiry date Jan vy B Minimum days Maximum days Warning days Inactive days Group Membership Primary group Default group localeveryone Secondary groups root 0 localeveryone
250. using the Certificate keys option and the cert file using the Certificate files option 4 You may also want to use a CA certificate for verifying your clients You may get CA certificates from Credentia VeriSign Thawte Geotrust etc Typically the certificate will have a name like cacert pem and it can be uploaded using the CA Certificates option Consult source web sites such as www credentia cc for further information Page 117 MAINTENANCE The Maintenance section includes the management modules Shutdown and reboot Monitoring Logs Processes lists Command Line access OS Integrity and Updates Many of these are capabilities that reach under the hood of the standard appliance functions of the EdgeWare Storage Server and allow you to get information about the functioning parts of these features and even to micro manage some of them Maintenance does what its name says and serves the purpose of keeping track of the underlying physical processes that are necessary for your Storage Server to work well Temperature voltage storage fill state SMART disk warnings and software integrity are among the things that can be tracked Maintenance supports special connectivity to the external world such as email notification popups SNMP alerts and even beepers These are used to provide warning of out of bounds conditions Maintenance also permits software updates and integrity checks and watching current logs and processe
251. vention Rebuild will automatically be launched as background process and an event notification will notify the user when the rebuild process is complete if 3ware manager is running and e mail notification is setup Replacing a hard drive in the Cutting Edge Storage Server In RAID5 and RAID1 configurations you can lose any single drive in each RAID set and still have data integrity In the event of a drive failure you can quickly and easily replace drives without shutting down the Storage Server The Cutting Edge Storage Server system will automatically recognize and prepare new drives Selecting Force Continue Selecting the Force Continue on Source Errors check box before rebuilding the array ensures rebuilds are not terminated if ECC errors are detected on the source disk Uncorrectable blocks will be rewritten but the data may be incorrect It is recommended that a file system check be executed when the rebuild completes By default this function is disabled Select this option only if the initial rebuild has failed Why cant I rebuild my hard drive in the Cutting Edge Storage Server Rebuilding a hard drive may fail for the following reasons e The hard drive being used for the rebuild is damaged Try using another hard drive CAUTION Use Cutting Edge replacement drives only Failure to do so will void your warranty e More than one drive is damaged Contact Cutting Edge for repairs FIBRE CHANNEL ARRAYS XYRATEX RAID Page 39
252. vices that can fail over from one EdgeWare Storage Server to another are e iSCSI e Windows Common Internet File Services CIFS e UNIX Network File Service NFS e File Transfer Protocol FTP e AppleTalk AFP Page 149 e Remote Replication rsync The current version of EdgeWare does not fail over Remote Mirroring drbd The steps below insure a smooth IP failover configuration between two EW3 servers that support failover services They are listed by Roman numerals because this is a super deployment whose steps e g Handshake can themselves have complex deployments described above in this manual I Set up the handshake II Set up and activate but do not connect the HARs III Set up the Primary and Secondary IV Configure the heartbeat service V Register the shared IP address VI Start the heartbeat monitor and get resources VII Activate the shared HARs I SET UP THE HANDSHAKE The two EW3 IP failover servers should have the following configuration prior to more advanced options to verify usage e The first network port ethO on the back of each EW3 server should be connected to the networking switch preferably a 1 Gb s capable switch e The second network port eth1 should be connected directly to the ethi port on the other EW3 server and by pass any network switch NOTE a crossover cable is not necessary as the hardware auto detects the direct connection The handshake should be set up as descr
253. vsttpd log Nel Se XferLog File varlogderlog SS Replication Save gt Return to index Figure 79 Networking Options for FTP Anonymous User Controls whether anonymous logins are permitted or not If enabled both the usernames ftp and anonymous are recognized as anonymous logins Default No Zf you enable this check whether the defaults on the next three variables are what you really want Anonymous Upload If set to Yes anonymous users will be permitted to upload files under certain conditions For this to work the option write_enable must be activated and the anonymous ftp user must have write permission on desired upload locations Default Yes Anonymous Make Dir If set to Yes anonymous users will be permitted to create new directories under certain conditions For this to work the option Write Enable must be activated and the anonymous ftp user must have write permission on the parent directory Default Yes No Anonymous Password When enabled this prevents vsftpd from asking for an anonymous password the anonymous user will log straight in Default Yes Idle Session Time Specifies the maximum amount of time between commands from a remote client Once triggered the connection to the remote client is closed Default value 600 Page 104 Data Connection Timeout Specifies maximum amount of time data transfers are allowed to stall in seconds Once triggered the connection to the remote client is closed Defa
254. with bug fixes as soon as they are reasonably available If Customer has not selected either the Gold Support Package bug fixes will only provide temporary or permanent workarounds as available LIMITATIONS ON SUPPORT AND ERROR CORRECTIONS The Services to be provided are limited to addressing problems that are demonstrable and reproducible Manufacturer makes no commitment representation or guaranty regarding the amount of time it will take to diagnose or resolve a problem once it is brought to Manufacturer s attention Except as expressly stated in this Agreement Manufacturer will have no obligation to correct errors in or failures of any Covered Products 3 Utilization Limitations Grant of License a LICENSE GRANT Subject to the terms of this License Manufacturer grants to Licensee a non exclusive nontransferable license to use the Product in object code form This License may be terminated by manufacturer effective upon notice to Licensee if Licensee fails to pay in full all fees for the Product and or any Manufacturer hardware on which this Product is installed Other than as specifically described herein no right or license is granted to Licensee to any of Manufacturer s trademarks copyrights or other intellectual property rights The Product incorporates certain third party software which is used subject to licenses from the respective owners The protections given to Manufacturer under this License also apply to the suppliers
255. ze for the logical volume Click on Save 10 A confirmation page appears Figure 32 confirming this file system can be resized Oi Oy AE go The filesystem on this logical volume can be resized After the logical volume is resized the file system will also resize to the new size of the logical volume Resize Logical Volume HM Return to index Figure 32 Resizing confirmation page 11 Click on Resize Logical Volume A result announcement page appears 12 Click on Return to Index on the result announcement 13 Click Mount the logical volume The logical volume is now resized and available for use 14 Recreate the shares from a configuration backup or other technique 7he data is still there The resizing procedure for an iSCSI volume requires all initiators be disconnected and the iSCSI target deleted see documentation below for Storage gt iSCSI module Then steps 3 through 12 are done except 6 is unnecessary as the volume will be found Not in use After that restore the iSCSI target in Storage gt iSCSI module and reconnect via the initiator s Unfortunately the data does mot reappear in this case and must be restored from backup SNAPSHOT VOLUMES The snapshot facility of the EdgeWare OS allows the administrator to create a new block device which is an exact copy of a logical volume frozen at a point in time One typical use is when some batch Page 49 processing a backup for instance

EdgeWare OS User Manual - Cutting Edge Networked Storage

Contents

Download Pdf Manuals

Related Search

Related Contents