Home
        Oracle Optimized Solution for JD Edwards EnterpriseOne (low cost
         Contents
1.   Ensure that the group and    user IDs are the same as those used in the other zones     groupadd  g 98194050 oinstall   groupadd  g 98194051 dba   useradd  u 237445  g oinstall  G dba  d  export home oracle  m  s     usr bin bash oracle    passwd oracle oracle    Note  these settings configure default  insecure  passwords purely as examples  more secure passwords    should be provided     Log in to the oracle user account and add the following paths to the oracle uset   s  profile    file  which is found in the oracle users home directory     ORACLE_BASE  db app oracle  export ORACLE_BASE  ORACLE_HOME SORACLE_BASE product 11 1 0 db_1  export ORACLE_HOME  ORACLE_SID orcl  export ORACLE_SID  JDE_HOME  db jdedwards JDE_HOME bin  export JDE_HOME   PATH SPATH   opt  SUNWspro bin  SORACLE_HOME bin   db usr local bin   usr ucb   export  PATH     for Verity software   VDK612_HOME  db  jdedwards e900 verity612 solaris  export VDK612_HOME  VDK612_PATH SVDK612_HOME _ssol26 bin  export VDK612_PATH  PATH SVDK612_PATH   PATH  LD_LIBRARY_PATH S VDK612_PATH  LD_LIBRARY_PATH  export LD_LIBRARY_PATH    In the DBServer domain  change the following project settings     Increase    jds zonel gt  projmod  s  K  project  max shm ids   privileged  1024 deny   default    the maximum size of a shared memory segment  Default      of physical memory   jds zonel gt  projmod  s  K  project  max shm memory   privileged  12gb  deny      default    34    Oracle Optimized Solution for JD Edwards EnterpriseO
2.   workload is the key to driving the solution to obtain the best possible results that most closely provide  a parallel to real world workloads  In order to optimize the workload  specific configurations of load  generator and test scripts where prepared  The result of this tuning exercise is documented in the next  paragraphs    The interactive portions of the workload ate generated during testing through the load generation  server  To allow the 1200 user load level to be achieved  changes were made to the default    configuration supplied as part of the DIL package to increase the number of users who are able to  execute each task     It was important to test all aspects of the SPARC T4 1 server configuration  so considerable effort was  put into generating appropriate UBE batch workloads during the testing of the interactive response of  the system  Several variations of batch workloads were prepared and tested     22    Oracle Optimized Solution for JD Edwards EnterpriseOne       Short Running UBE Workloads    The short running UBE workloads  as discussed previously  were also tested  These workloads are  created and executed to simulate the actions of users initiating non interactive tasks on the system   such as initiating the preparation and printing of a PDF document such as an invoice  These typically  are initiated and run for 15   30 seconds to run to completion  These jobs are used in these tests to  generate consistent    background    workloads on the database 
3.  JD Edwards EnterpriseOne is a cost effective and practical way of ensuring  proper tracking and compliance  The suite of software consists of three core functional blocks  on top  of which application modules can be loaded  These core functions are listed in Table 1     TABLE 1  CORE FUNTIONAL BLOCKS IN JD EDWARDS ENTERPRISEONE SOFTWARE SUITE       SERVER FUNCTION DESCRIPTION       A database server A conventional Oracle Database 11g Release 2 Standard Edition  server for transaction storage and asset tracking       A Web server A conventional Oracle WebLogic Server 10 3 2 Standard Edition  Web server providing the user interface to the optional modules       The JD Edwards EnterpriseOne application server and The JD Edwards EnterpriseOne 8 98 9 0 application server core  optional modules installation and optional business functionality modules       The JD Edwards EnterpriseOne core installation requires only Standard Edition versions of Oracle  Database and Oracle WebLogic server software  Upgrades to Enterprise Edition versions are possible  to take advantage of more advanced capabilities  However  these are not necessary for the  configurations discussed in this document     Taking advantage of the Oracle Solaris virtualization features  each of these    servers    is hosted inside an  Oracle Solaris Container hosted on a single SPARC T4 1 server from Oracle  This approach provides  several advantages to the user     e Separation of the applications    working environmen
4.  Solution for JD Edwards EnterpriseOne       zonecfg   zonecfg   zonecfg   zonecfg     zonecfg        zonecfg     E  E  E  E  E  E       1Server fs gt  set type zfs  1Server fs gt  end   1Server gt  verify   1Server gt  commit   1Server gt  info    1Server gt  exit    If Oracle Solaris DTrace will later be used to investigate the behavior of the zone  this additional    setting should be issued     zonecfg  z ElServer    zonecfg  zonecfg  zonecfg    zonecfg     ElServer gt      ElServer gt  verify   ElServer gt  commit     ElServer gt  exit    set limitpriv default dtrace_proc dtrace_user    Now the zone is ready to be booted and further configurations can be carried out inside the zone OS    image     zoneadm  z ElServer install    zoneadm list  cv    zoneadm  z ElServer boot    Log in to the server and carry out the following operations to create the required user profiles     zlogin  C ElServer    usermod  s  usr bin bash root    If you root should be allowed to directly log in to the zone  make the following changes     Edit the file  etc ssh sshd_config  and change the line     PermitRootLogin no    To read     PermitRootLogin yes    And then restart the ssh daemon with the command     svcadm restart ssh    28    Oracle Optimized Solution for JD Edwards EnterpriseOne       Create the users requited for the JD Edwards EnterpriseOne installation  and ensure that the group  and user IDs are the same as those used in the Oracle Database zone    groupadd  g 98194049 jde900   
5.  a  system such as the SPARC T4 1 server  To stress the system to determine its true capabilities   batch type workloads must also be used  Batch processes are non interactive and are often used for  functions such as General Ledger processing and Sales Invoice Print Runs for which user interaction is  not required  Such batch tasks can be broadly separated into short running  less than a few tens of    seconds to complete  and long running  many minutes to hours to complete  tasks     The JD Edwards EnterpriseOne terminology for these batch workloads is    Universal Batch Engine  processes     UBEs   UBEs are a common occurrence in the day to day operation of an enterprise   encompassing anything from short batch jobs  such as lightweight database queries to prepare and  print a PDF document  to much longer running jobs  such as Inventory or Payroll reconciliation   which may run for hours and heavily exercise the database and  consequently  the disk storage system     With the goal of demonstrating a real system workload  scripts have been developed to ensure that a  significant level of load is generated on the system by running significant levels of both short and long  batch tasks during testing of the interactive response times to the queries of the target 1200 interactive  users of the system  In order to mote realistically gauge overall system performance  a mixture of   long  and short running UBE workloads are simultaneously launched throughout the testing of the  D
6.  are either very  expensive  especially in the case of SAN storage  or wasteful in the sense that both capacity and power  consumption ate excessive when large numbers of sparsely populated    short stroked    disks atrays ate  deployed     The traditional solutions described above have long been used to address the problem  but the  increasing cost pressures of administering these large disk farms  for example  floor space  increasing  energy consumption  and heat generation  make it increasingly difficult to justify the cost performance  trade off     Oracle has been able to achieve optimal performance from the onboard disk storage of the SPARC  T4 1 server through careful selection of an appropriate disk controller Host Bus Adapter  HBA  that  provides performance acceleration via and onboard battery backed write cache  This offers a  cost effective alternative to achieving sufficiently high performance in this application  with the added  benefit of radically reducing the size  power  and management overhead often associated with external  storage arrays  and it is an excellent way to accelerate databases and disk I O intensive applications     Solution Software Components   The primary software components that comprise this Oracle Optimized Solution are described in the  subsections that follow    Oracle Solaris 10 10 09 Operating System    Oracle Solaris provides industry leading performance and numerous advanced features  including     Virtualization   Oracle Solaris C
7.  capability to scale the JD Edwards EnterpriseOne capacity up to an  anticipated 4 000 to 5 000 users by offloading the database portion of the workload to an additional  dedicated SPARC T4 1 server with a dedicated Fibre Channel storage array  The choice of putting the  database on an additional server is indicated by the maximum I O and memory utilization data  captured in the medium configuration when running with UBE workloads during testing     Oracle Optimized Solution for JD Edwards EnterpriseOne       If a customer environment were to be challenged by the ultimate performance capabilities of a medium  configuration  the upgrade path is incremental and does not requite a complete replacement of already  acquired assets  improving ROI for this solution     Conclusion    Testing shows that a single SPARC T4 1 server from Oracle can provide extremely competitive  performance for small and medium JD Edwards EnterpriseOne deployments  During testing   performance optimization was achieved through careful arrangement of the I O subsystem  which  involved separating and optimizing the placement and use of the various data resources on the servet   s  local storage system  By optimizing the use of the local storage  it is possible to provide high  performance while consuming a very small footprint  This enables a very cost effective solution   maximizing customer ROI as well as minimizing TCO     The solution leverages conventional storage in the most beneficial modes of opera
8.  the ZFS storage subsystem  enhancing its  performance        High capacity disk and I O subsystem An eight   disk capacity  local storage subsystem within the  chassis allows up to 4 8 TB of raw storage capacity that can  be managed through onboard or enhanced hardware RAID  controllers for ultimate data protection        Oracle Optimized Solution for JD Edwards EnterpriseOne       Sun Storage 6 Gb SAS PCle RAID HBA    The Sun Storage 6 Gb SAS PCIe RAID HBA from Oracle provides very high data reliability   availability  and performance to businesses that are facing storage challenges driven by unprecedented  data growth  With data transfer rates of up to 6 Gb sec per port  the RAID adapter brings users  improved performance and capabilities compared to onboard controllers  Designed for Oracle   s latest  high speed PCI Express based servers  Sun Storage 6 Gb SAS PCle RAID HBAs couple advanced  RAID data protection with 8 port  6 Gb sec SAS connectivity and 512 MB of Battery Backed Write  Cache  BBWC  to provide secure  scalable  performance optimized  enterprise class data storage and    retrieval solutions     Storage Architecture    The underlying hardware RAID disk configuration is optimized to enhance both speed of access and  reliability of data on disk  A 6 Gb SAS PCIe RAID HBA is used  This card includes a 512 MB write  cache and battery back up features that combine to improve data integrity and significantly accelerate  write transactions by reducing latency and all
9. Back Write Cache  This provides a key benefit in that implementing a large write cache  allows the OS to receive acknowledgement of a confirmed write to disk operation very rapidly  The  card accepts writes to cache before the data is committed to disk  The data is written to disk at some  point in the near future as time allows  Very rapid confirmation of writes significantly reduces write  latency and  therefore  increases application response times and throughput considerably  In the event  of a power failure before the acknowledged data is written to disk  the battery backed cache will be  maintained until power returns  and the data will be immediately committed to disk following the    restoration of power to the server     The use of the RAID HBA also allows more advanced RAID data protection configurations  in  addition to providing performance benefits  particularly on data reads  due to data striping across    multiple devices     Storage Logical Volume Configurations    The RAID HBA is initialized and partitioned into three independent logical volumes to allow different  RAID configurations for different uses and to separate the different storage areas to allow for more    granular backup and restote processes     paisa a oe    Logical Volume    WA       ZFS Database ZPOOL    ig Volume                  lognodl     ZFS Database Redo Log POOL              rootnodl      WAAAY  ZFS Root installation ZPOOL via ZFS     E1pool    filesystem        om Volume    Hardware Ma
10. E   S COMPLETED 0 2730 7349 8522 0 1741 2024    LONG UBE   S 0 0 0 0 8 4 2  RUNNING COMPLETED          AVERAGE INTERACTIVE 0 270 0 312 0 391 0 519 0 503 0 583 0 588  RESPONSE TIME  SECS                                Test Run 1  Best Achievable Response Time During DIL Testing    The purpose of this test was to determine the best possible response time that can be achieved with  the SPARC T4 1 server under 1200 user load levels  This test is cartied out with no UBE load running  and demonstrates a 0 27 response time average for all interactive queries  Note that no system  resources where exhausted during this test     Oracle Optimized Solution for JD Edwards EnterpriseOne       Test Runs 2 and 3  Extended Response Time Testing with Increased Short UBE Load    These test cases extend the testing of the response time to interactive queries during the addition of  considerable short running UBE job loads  This test case demonstrates that the server can achieve  linear growth in short running UBEs with corresponding small increases in overall interactive response  times since the I O subsystem of the server is not saturated during testing     Testing demonstrated the ability to scale from 2730 short UBEs with 0 312 second response times up  to 7349 short UBEs with a 0 391 second response time     Test Run 4  Maximizing Short UBE Throughput with 0 5 Second Response Times    This test case finds the maximum capability of the system to cope with short UBE loads with  acceptable int
11. Edwards EnterpriseOne suite of Enterprise Resource  Planning  ERP  tools from Oracle  The paper demonstrates how Oracle   s servers  enhanced with carefully tuned conventional and advanced flash based storage  technology provide an excellent price performance combination for small and medium  enterprise deployments     Server consolidation through virtualization is key to this solution  allowing a very dense  and yet high performance solution at low cost and low data center impact  The result is  significant reductions in both initial acquisition and ongoing operating costs to the  enterprise     The JD Edwards EnterpriseOne stack consists of a core set of tools supplemented by  more than seventy application modules addressing industry specific enterprise resource  planning needs  In the testing carried out as part of this Oracle Optimized Solution   commonly used modules are included in a standardized    Day in the life     DIL  workload  that is used to obtain performance metrics for particular system configurations to  characterize their behavior     As part of this paper  a sizing guide and tuning best practices section are provided as  well as information regarding next steps should customers need to scale within and  beyond the capabilities of the chosen platform     Oracle Optimized Solution for JD Edwards EnterpriseOne       A Brief Overview    Today   s enterprises are faced with the need to track assets and ensure legal compliance in ever more  complex environments 
12. EnterpriseOne workload conditions that are possible in the actual deployment to ensure that  acceptable performance under worst case conditions can be maintained before additional consolidation  is performed  If such worst case situations can be satisfied  then considerable further value add and  ROI benefits are readily available to customers choosing to deploy and consolidate on this platform     Oracle Optimized Solution for JD Edwards EnterpriseOne       Sizing Guidelines    For sizing guidelines  hardware configurations are provided for the workloads previously described as a  starting point for average  medium  and large deployments  For more detailed sizing information     contacting a local Oracle representative is encouraged and recommended     The configurations outlined below refer to the systems and configurations available at the time of the  test  It is recommended that the most recent options and system configurations available at the time of    implementation be deployed in the customer environment     Average Configuration   Up to 300 Real World Users    Table 6 provides the recommended hardware components for the medium configuration     TABLE 6  RECOMMENDED HARDWARE COMPONENTS FOR THE MEDIUM CONFIGURATION       SERVER VARIANT COMPONENT CONFIGURATION       Processor  SPARC T4  2 85 GHz  8 core  Factory pre installed   8 threads core             Memory  64 GB RAM  utilizing 16 x 4 GB As indicated in service manual   DIMMs  SPARC T4 1 Server Disks  8 x 300 GB 
13. IL workload by some custom scripts running on the JD Edwards EnterpriseOne server itself  During  testing  these may launch thousands of short UBEs hr and a configurable number of long running  UBEs throughout the period of the benchmark run     Demonstrating the batch UBE workload allows Oracle to ascertain the achievable level of interactive  users while simultaneously maintaining a more realistic level of batch workload  giving a more realistic  metric for overall system capability     Further details about the DIL and UBE workloads are provided in Appendix D     Oracle Optimized Solution for JD Edwards EnterpriseOne       Test Environment    The test environment was implemented as described previously in the    Solution Architecture  Overview    section of this document along with the following two additional servers that were required  for the benchmark configuration     e Deployment server   A small configuration x86 server running the JD Edwards EnterpriseOne  Deployment Server for controlling software updates to the JD Edwards EnterpriseOne Server  This  server does not take an active part in the testing of this configuration  but is required for the initial  deployment and application code maintenance of the JD Edwards EnterpriseOne environment     e Load generation server   An x86 server acting as a load generation device  load testing tool  and  performance evaluation server load     In depth descriptions of solution component configurations are provided in the sub
14. ORACLE    OPTIMIZED SOLUTIONS    An Oracle Technical White Paper  October 2011    Oracle Optimized Solution for JD Edwards  EnterpriseOne    Delivering Performance and Efficiency Using Oracle   s SPARC  T4 1 Server    ORACLE    Oracle Optimized Solution for JD Edwards EnterpriseOne       VV EPOCIMG IONE AeneA E E EE E E E ET 1  A BST AD VENI W ceatra ae a a a ETRS 2  Solution Architecture OVervieW              cccccccceeeeeeeesseneeeeeeeeeeeeeeeeeaeees 3  A New Approach to Optimizing Storage            cccccceeeeeeesssteeeeeeeees 4  Solution Software Components                    cccceeeeeeeeeeeeeeeeeeeeeeeeeeees 4  Licensing of the Software Stack           ccc cccceeeesssssscceeeeeeeeeeeeeseeees 5  Solution Hardware Components                 cc eeceeeeeeeeeeeeeneeeeeeeeeeeeeeee 6  Storage Architecture eoira aa eR 7  Workload DescriPti   M meriiri n EE a G 8  Test Environment eea a a E AE A NEAT RAE 9  SPARC T4 1 Server Configuration                  cccceeeeeeseseseceeeeeeeeeenees 9  Test Results and Analysis    aaeiiecontieosiene needa 13    Test Run 1  Best Achievable Response Time During DIL Testing 13  Test Runs 2 and 3  Extended Response Time Testing with    Increased Short UBE Load aise et se Oe Oss oth Pascal 14  Test Run 4  Maximizing Short UBE Throughput with 0 5 Second  Response TIMeSs  2 t5ci Secs tania Mesh ieccss etree be iatieereante 14  Test Run 5  Maximum Long UBE Throughput with Acceptable  Response MMOS so  Sicance Sack ote ace Renn eas canted mirsanteetea
15. SAS2 2 5 inch disk drives Install disks in all available slots   HBA  Sun Storage 6Gb SAS PCle RAID As indicated in service manual and  HBA  internal  8 port with 512 MB memory installation guide  PCI slot location and cable    routing options must be followed  All eight  ports must be cabled for maximum  performance           1 LEGAL DISCLAIMER   Expected performance characteristics are based on laboratory test implementations and are  provided as is without warranty of any kind  The entire risk of the use of information provided herein remains with the  recipient and in no event shall Oracle be liable for any direct  consequential  incidental  special  punitive  or other  damages including  without limitation  damages for loss of business profits  business interruption  or loss of business  information     Oracle Optimized Solution for JD Edwards EnterpriseOne       For the purposes of the DIL testing at the 1200 user level  an 8 disk chassis fully populated with   300 GB disks was found to provide sufficient capacity  However  for actual customer deployments   larger capacity disk storage may be required  It should be noted that is not possible to re allocate disks  to vary storage in the various logical volumes provided by the RAID HBA since all eight disks are  required in the indicated configurations to provide the minimum required number of disks present in  each of the RAID volume types   RAID 1 requires a minimum of two disks  and each RAID 5 volume  requires a minim
16. This document is provided for information purposes only and the  contents hereof are subject to change without notice  This document is not warranted to be error free  nor subject to any other  warranties or conditions  whether expressed orally or implied in law  including implied warranties and conditions of merchantability or  fitness for a particular purpose  We specifically disclaim any liability with respect to this document and no contractual obligations are  formed either directly or indirectly by this document  This document may not be reproduced or transmitted in any form or by any  means  electronic or mechanical  for any purpose  without our prior written permission     Oracle and Java are registered trademarks of Oracle and or its affiliates  Other names may be trademarks of their respective owners   Intel and Intel Xeon are trademarks or registered trademarks of Intel Corporation  All SPARC trademarks are used under license and  are trademarks or registered trademarks of SPARC International  Inc  AMD  Opteron  the AMD logo  and the AMD Opteron logo are    trademarks or registered trademarks of Advanced Micro Devices  UNIX is a registered trademark licensed through X Open  Company  Ltd  0611    Hardware and Software  Engineered to Work Together    
17. anagement purposes  these groupings have been  adopted  For planning purposes  it is recommended that disks in the same group be kept physically    close to each other  though this is not required for performance reasons     Oracle Optimized Solution for JD Edwards EnterpriseOne       Test Results and Analysis    Multiple test cases where run to test several dimensions of server performance with varying load  configurations of JD Edwards EnterpriseOne  Seven tests where completed and raw data was  collected  as shown in Table 5        TABLE 5  DATA COLLECTED FROM SEVEN TESTS    DESCRIPTION NO UBE   S LOW MEDIUM HIGHEST 8 LONG 4 LONG 2 LONG          SHORT SHORT SHORT UBE   S UBE   S UBE   S  UBE   S UBE   S UBES ONLY PLUS PLUS HIGH  MEDIUM SHORT  SHORT UBE   S  UBE   S  TEST RUN 1 2 3 4 5 6 7  TEST DURATION  H MM SS  1 11 56 1 11 46 1 11 45 1 11 55 1 11 59 1 11 49 1 11 50  JD EDWARDS ENTERPRISEONE 2 5 13 26 7 15 17    SERVER   CPU UTILIZATION       JD EDWARDS ENTERPRISEONE 20 20 21 23 19 19 19  SERVER   MEMORY UTILIZATION             DATABASE SERVER   CPU 0 1 2 5 2 4 5  UTILIZATION   DATABASE SERVER   MEMORY 30 31 31 31 31 32 32  UTILIZATION   ORACLE WEBLOGIC SERVER 2 3 5 6 3 4 4      CPU UTILIZATION    ORACLE WEBLOGIC SERVER 22 21 22 21 21 21 21    MEMORY UTILIZATION       DATABASE SERVER DISK I O 5 78 6 23 9 86 20 100 100 100 100 100 81 100  UTILIZATION   AVG MAX       DATABASE SERVER REDO LOG 1 2 2 17 2 4 4 34 14 83 17 86 16 81  DISK I O UTILIZATION   AVG MAX       SHORT UB
18. and JD Edwards EnterpriseOne server to  ensure that normal daily processing tasks do not have an adverse affect on users    interactive response    times     In order to simulate ongoing creation and execution of these tasks  a script was created in the  ElServer zone that resubmits batches of short UBE tasks every 15   20 seconds to ensure a       consistent level of short UBE load     Long Running UBE Workloads    The long running UBE jobs are started once at the beginning of a test cycle and allowed to run  throughout the testing  Some batch jobs complete within the approximately hour long processing of  the DIL workload  so they are restarted  if necessary  to maintain the target number of long UBE jobs  running during particular tests  To provide further information on the behavior of the system under  long running UBE batch workloads  all the DIL long running UBE tasks were started and allowed to  run to completion to determine runtimes with no other load  interactive or short UBE  on the system   The details of the operations carried out by these tasks are also provided to allow users to determine  applicability of these results to theit own scripts and UBE tasks     TABLE 9  UBE WORKLOAD COMPONENTS          T4 1 T3 1  PLATFORM PLATFORM  RUNTIME RUNTIME  WITHOUT WITHOUT  INTERACTIVE INTERACTIVE  LOAD LOAD  UBE NAME  MINUTES   MINUTES  DESCRIPTION  R42565 36 2 91 71 Sales Order Invoicing Batch Job       The R42565 batch process was configured to initiate document invoic
19. ase orders  work orders  and manufacturing processes such as inventory tracking   These ate sometimes generically referred to by industry acronyms such as SCM  CRM  HCM  SRM   and FMS     The DIL workload uses the JD Edwards EnterpriseOne interactive transaction types described in  Table 10     36    Oracle Optimized Solution for JD Edwards EnterpriseOne       TABLE 10  JD EDWARDS ENTERPRISEONE INTERACTIVE TRANSACTIONS       TRANSACTION TYPE DESCRIPTION VIRTUAL USERS PER  TRANSACTION TYPE AT  1200 USER LEVEL                                                       1 P03B102 Apply Receipts 110 users  2 P04111 Supplier Ledger Inquiry 110 users  3 P051191 Daily Time Entry 24 users  4 P17500 Case Management Add 60 users  5 P31114 Work Order Completion 36 users  6 P3411 MRP Messages  WO Orders  24 users  7 P3411 MRP Messages  OP Orders  24 users  8 P3411 MRP Messages  OT Orders  24 users  9 P4113 Inventory Transfer 60 users  10 P42101 Sales Order Entry     10 Line Items 340 users  11 P42101 Sales Order Update 60 users  12 P4310 Purchase Order Entry     25 Line Items 220 users  13 P4312 Purchase Order Receipts 24 users  14 P4314 Voucher Match 24 users  15 P4915 Ship Confirmation     Approval only 36 users  16 P4915 Ship Confirmation     Confirm Ship only 12 users  17 P4915 Ship Confirmation     Confirm and Change Entry 12 users       Note that the    Users    column indicates the number of users who may simultaneously be carrying out  that particular kind of transaction  The DIL test 
20. bServer net gt  set address 10 6 162 24 22  zonecfg WebServer net gt  set defrouter 10 6 163 254  zonecfg WebServer net gt  end  zonecfg WebServer gt  add net  zonecfg WebServer net gt  set physical igb1  zonecfg WebServer net gt  set address 10 6 222 84 24    zonecfg WebServer net gt  set defrouter 10 6 222 254       zonecfg WebServer net gt  end  zonecfg WebServer gt  verify  zonecfg WebServer gt  commit    zonecfg WebServer gt  info       zonecfg WebServer gt  exit    30    Oracle Optimized Solution for JD Edwards EnterpriseOne       If DTrace is required to investigate the performance of this zone  this additional setting should also be    issued     zonecfg    zonecfg   zonecfg   zonecfg     zonecfg      z WebServer   WeBServer gt  set limitpriv default dtrace_proc dtrace_user  WeBServer gt  verify   WebServer gt  commit    WebServer gt  exit    The zone can now be installed and booted     zoneadm  z WebServer install    zoneadm  z WebServer boot    Log in to the server to create and modify its OS specific settings     zlogin WebServer    usermod  s  usr bin bash root    If root should be allowed to directly log in to the zone  make the following changes     Edit the    file  etc ssh sshd_config  and change the line     PermitRootLogin no    To tead     PermitRootLogin yes    And then restart the ssh daemon with the command     svcadm    restart ssh    Add the required users and groups to the OS image     groupadd  g 98194050 oinstall  groupadd  g 98194051 dba  useradd  
21. build  and  configured with 64 GB of swap      The OS should be installed on c1t0d0  which is a RAID 1 mirtored pair of disks  The output of the  zpool status rootpool command can be used to confirm this  The rootpool should be  hosted on the c1t0d0s0 device    bash 3 00  zpool status rootpool  pool  rootpool  state  ONLINE  scrub  none requested    config     NAME STATE READ WRITE CKSUM  rootpool ONLINE 0 0 0  c1lt0d0s0O ONLINE 0 0 0    errors  No known data errors     bash 3 00     Now create the additional zpools to hold the data for the various zones that will be created   zpool create datapool clt1d0  zpool create logpool cl1t2d0    The Global Zone has been installed on ZFS using rootpool     26    Oracle Optimized Solution for JD Edwards EnterpriseOne       Create a file system on the root pool to house the JD Edwards EnterpriseOne server data     zfs create rootpool Elserverdata  chmod 700  rootpool Elserverdata  Create file systems to host the zone root images that will be used for each zone   zfs create rootpool DBserverroot  chmod 700  rootpool DBserverroot  zfs create rootpool Elserverroot  chmod 700  rootpool Elserverroot    zfs create rootpool Webserverroot             chmod 700  rootpool Webserverroot    Each zone now has a defined serverroot directory and also a data storage pool constructed on top    of appropriate types of storage with suitable performance characteristics based on the type of data and  access patterns that are most likely     In Oracle Solari
22. creases are easily achievable with the SPARC T4 1  server     Real World Sizing Based on the DIL Test Workload    To account for the fact that the DIL workload cannot accurately predict real world customer use cases  and performance requirements  conservative hardware sizing guidelines are outlined in this section     The hardware sizing guides target substantially lower real world user counts to avoid any potential for  performance issues once the solution is deployed and to allow reasonable accommodation for growth  and peak demands that cannot be accurately predicted     Real world sizing  therefore  covers 300 user configurations based on testing carried out at the  1200 users level under the DIL workload  Such conservative actual deployment sizing is common in  JD Edwards EnterpriseOne deployments with sizing guides based on testing using the DIL kit     Room for Consolidation    As indicated in the previous discussion regarding real world sizing  considerable de rating has been  built into the proposed real world deployment sizes based on the performance seen under DIL testing  workloads     One of the primary benefits that may be realized as a result of this conservative sizing is that  substantial headroom may be available for further unrelated workload consolidation or for integration  of connected applications that can augment the capabilities of JD Edwards EnterpriseOne     Such additional consolidation  if undertaken  should consider the full range of JD Edwards  
23. e  by the HBA   cache   rootpool    2x300 GB SAS2 drives managed by the The logical volume Holds the OS installation and    Sun Storage 6 Gb SAS PCle RAID HBA    Configured as a single 300 GB  RAID 1 protected logical volume     Write performance is significantly enhanced  through the NVRAM write cache provided  by the HBA     is used to directly  host the rootpool  ZFS pool  No  additional data  protection is  required     The E1pool is also  housed in a file  system on this  volume     virtual machine container disk  images and OS specific  configuration files  Benefits  significantly from the rapid write  acknowledgements provided by the  NVRAM write cache     The E1pool is used to store JD  Edwards EnterpriseOne server  data  This data is characterized as     read mostly           The physical layout of the storage of the SPARC T4 1 server is illustrated in Figure 3  Note that the    RAID HBA controller is not seen in this image  It is installed in the rear of the server and does not    connect externally     Oracle Optimized Solution for JD Edwards EnterpriseOne                 Sen gn reese coral OME     rootpool   logpool   datapool     Figure 3  Physical layout of storage for an 8 disk SPARC T4 1 server     As can be seen in the diagram  three groups of disks have been configured  Physical location is not  critical since all disks are    equally close    to the disk controller and no disk slots are    favored    in terms of  performance  For convention  planning  and m
24. e  orders located in the F4211 table  Sales Order Detail File      There were 150 documents processed in the testing and it included  47 436 records in the F42119 table  Sales Order History File         23    Oracle Optimized Solution for JD Edwards EnterpriseOne       R42800    20 8    0 52    Sales Order Invoicing       The R42800 batch process was configured to initiate selected document  numbers located in the F4211 table  Sales Order Detail File  with a next  status code of 620 and within a single business unit     There were 12 900 records processed with this data selection criterion        R43500    62 36    155 23    Purchase Order Print       The R43500 batch process was configured to initiate document invoice  numbers located in the F4311 table  Purchase Order Detail File  with a  next status code of 280 and within a single business unit     There were 255 481 records processed with this data selection criterion        R4891    25 44    45 81    Freight Update       The R4981 batch process was configured to initiate shipment numbers  located in the F4215 table  Shipment Header  with a next status of 70 and  within a single business unit     There were 800 records processed with this data selection criterion        R09801    29 7    78 16    General Ledger Post       The R09801 batch process was configured to initiate batch numbers  located in the F0011 table  Batch Control Records  with a batch status of   A  and batch type of  G      There were 990 099 records proc
25. e underlying capabilities upon which three Oracle Solaris  Containers are created to carry out the operations required of the full JD Edwards EnterpriseOne  stack     In the base platform  the disk configuration is managed and configured in such as way as to optimize  performance of the storage subsystems for each of the Oracle Solaris Containers that provide the  application functionality  In this base platform  the base operating system is installed  which consists of  Oracle Solaris 10 10 09  and all relevant patches and any required additional platform wide facilities are  configured and managed  In addition  the required security profile is applied to lock down the base  installation  automatically adding security to the virtualized Oracle Solaris Containers installations     Oracle Optimized Solution for JD Edwards EnterpriseOne       A New Approach to Optimizing Storage    The key area requiring optimization in this solution is the disk I O performance  This can be seen in  analysis of the performance results later in this document  The classic solution to this problem has  been to utilize external high performance storage devices  such as SAN arrays  or to over provision  storage and under utilize the capacity  often called    short stroking    disk drives  to achieve a high spindle  count and wastefully stripe data across the higher speed portion of these disk to boost aggregate I O  operations per second  IOPS   These classical approaches to increasing storage speed
26. ed for Oracle Database Redo Log files Server data       Oracle Optimized Solution for JD Edwards EnterpriseOne       Workload Description    In order for performance and scalability to be adequately characterized  a standard workload was used  to stress the system under test and achieve the optimum utilization of the architecture under peak  loads  The results provide a suitable basis for performance comparisons with alternative competitive  architectures and allow customers to draw accurate comparisons based on their specific workloads    during    real life    operations     For this purpose  the JD Edwards EnterpriseOne product team provides a    Day in the Life     DIL   workload that uses a cross section of JD Edwards EnterpriseOne features in a typical mid scale  manufacturing enterprise environment to simulate a typical day   s workflow  The workload can be  configured to allow the maximum capability of the platform under test to be ascertained  while at the  same time ensuring that vatious system performance metrics are captured  A key metric is the response  time to user queties issued via the Web based GUI as the number of simulated interactive users is  increased during the testing  The workload can be throttled to maintain maximum performance with  acceptable response times to allow tuning for the highest possible user count with an acceptable    response time     Interactive transactions  such as those in the DIL workload  typically generate only moderate load on
27. ed to the previous generation server     C  Tuning and Best Practices    Setting Up the Three Application Zones    Zones ate an effective and efficient means to provide virtual OS image isolation between application  stacks  and are a standard feature of the Oracle Solaris OS  For the Oracle Optimized Solution for JD  Edwards EnterpriseOne  three zones are configured based on the global zone  the initial base OS  image installed on the SPARC T4 1 server   Instructions for the creation and configuration of each  zone ate provided in this section of the document     Global Zone Configuration    Prior to configuring the OS on the global zone  the low level disk configuration must be completed in  the OBP environment  The procedures for this are described in detail in Appendix A  Once the disk  hardware configuration is completed  a conventional installation of Oracle Solaris can be carried out on  the RAID 1 mirrored logical volume and the remaining portions of the tuning and best practices  catried out     The global zone represents the original base OS installation on the platform  and it must be configured  appropriately to host the three application zones that will be used  the Oracle Database zone  the JD  Edwards EnterptiseOne zone  and the Oracle WebLogic Web server zone     The ZFS file system was used throughout this configuration because it provides considerable ease of  configuration and portability of zone images once configured   It is also recommended that the foll
28. eractive response times  During testing  the maximum short UBE throughput achieved  was 8522 jobs completed during the test while maintaining a 0 519 second average interactive response  time  This result demonstrates a maximum achievable result because the disk I O subsystem is  reaching peaks of 100  utilization and the response becomes non linear near this data point     Test Run 5  Maximum Long UBE Throughput with Acceptable Response Times    The purpose of this test was to test the platform   s capability to process a maximum number of long  UBE loads during testing while maintaining an acceptable 0 503 second interactive response time   Eight long running UBE jobs where maintained during the test period  Some jobs where completed  and restarted during the test  and others ran throughout the duration of the testing  During this time   no short UBE loads where tested     It can be seen that the long UBE loads heavily taxed the database and  consequently  the disk  subsystem of the server  maintaining peak and average disk utilization of 100  throughout the test  duration  This demonstrates that the maximum UBE load that can realistically be maintained is no  higher than eight jobs concurrently     Test Runs 6 and 7  Best Compromise Performance for Long and Short UBEs with  Acceptable Response Times    In real world environments  it is expected that users will generate a mix of both long and short UBE  loads during operation and the system will need to retain acceptable in
29. essed with this data selection criterion        R31410    9 94    17 13    Work Order Processing       The R31410 batch process was configured to initiate document invoice  numbers in the F4801  Work Order Master File      There were 28 751 records processed with this data selection criterion        R31802A    29 07    12 06    Manufacturing Account Journal       The R31802A batch process was configured to initiate document invoice  numbers in the F4801  Work Order Master File  with a status code of  95    There were 1 501 records processed with this data selection criterion        R42520    16 61    49 11    Print pick Slips Batch       The R42520 batch process was configured to initiate shipment numbers in  tables F4201  Sales Order Header File   F4211  Sales Order Detail File    and F49T211  Sales Order Detail File   Secondary Tag Table  and a next  status code of  520     There were 37 154 records processed with this data selection criterion        24    Oracle Optimized Solution for JD Edwards EnterpriseOne       Note  The status codes are defined by the activity rules of the JD Edwards EnterpriseOne  environment  Default status codes were used whenever possible     Note  Some database layout changes where implemented between testing on SPARC T3 1 and SPARC  T4 1 servers  This is expected to be the reason for the increased processing times for jobs R42800 and  R31802A  Even so  the average performance gain seen when using the SPARC T4 1 server is  approximately 2x compar
30. groupadd  g 98194050 oinstall   groupadd  g 98194051 dba   useradd  u 237422  g jde900  d  JDE  m  s  bin ksh jde900   passwd jde900 jde900   useradd  u 237445  g oinstall  G dba  d  export oracle  m  s    usr bin bash   oracle    passwd oracle oracle    Note these example settings configure insecure passwords  more secute passwords should be provided     Log in to the oracle user account and add the following paths to the   profile file found in the  oracle user   s home directory   ORACLE_BASE  db app oracle  export ORACLE_BASE  ORACLE_HOME SORACLE_BASE product 11 1 0 db_1  export ORACLE_HOME  ORACLE_SID orcl  export ORACLE_SID  JDE_HOME  db jdedwards JDE_HOME bin  export JDE_HOME  PATH S PATH   opt  SUNWspro bin  SORACLE_HOME bin   db usr local bin   usr ucb      export PATH    for Verity software  VDK612_HOME  db  jdedwards e900 verity612 solaris  export VDK612_HOME  VDK612_PATH SVDK612_HOME _ssol26 bin  export VDK612_PATH  PATH SVDK612_PATH   PATH  LD_LIBRARY_PATH SVDK612_PATH   LD_LIBRARY_PATH  export LD_LIBRARY_PATH       As root in the E1Server domain  change the following project settings to increase the maximum    number of shared memory segments  The default is 100   jds zonel gt  projmod  s  K  project  max shm ids   privileged  1024  deny     default    Increase the maximum size of a shared memory segment  Default   1   4 of physical memory   jds zonel gt  projmod  s  K  project  max shm memory   privileged  12gb  deny      default    Increase the maximum number 
31. hat are rapidly evolving and changing to keep up with the  demands of the business  Oracle Database 11g provides efficient  reliable  and secure data management  for mission critical transactional applications  query intensive data warehouses  and mixed workloads   With new self managing capabilities  Oracle Database 11g also eliminates time consuming  error prone  administrative tasks  so database administrators can focus on strategic business objectives instead of on    performance and availability fire drills     Oracle WebLogic Server 10 3 2 Standard Edition    Oracle WebLogic Server Standard Edition is an application server designed for building and running  enterprise applications and services  It fully implements the latest Java Platform  Enterprise Edition   Java EE  standards and offers choice in development frameworks and tooling  Comprehensive and  accessible management capabilities enable administration of sophisticated systems via a well designed  graphical console and or automation  All users benefit from Oracle WebLogic Server s reliability and  performance  which has been tested over years of enterprise grade production use in demanding    customer environments the world over     Licensing of the Software Stack    Software licensing has a material impact on the total cost to deploy integrated systems  In addition to  the end user applications themselves  the applications often carry prerequisites  such as database  management systems and Web middleware  This i
32. is provided with an expectation of running 1000 users  in total  This must be increased to 1200 users to fully test this platform  To reach this maximum level   the number of users running each transaction was increased by 20  in all cases compared to default  DIL configurations     37    Oracle Optimized Solution for JD Edwards EnterpriseOne       The short running UBEs generate consistent load on the JD Edwards EnterptiseOne server and are  continuously re submitted upon completion  Long running UBEs tend to generate substantial load on  the database server since they typically involve more complex SQL queries  Once started  these  long running UBEs continue to run throughout the test runs  Tests runs typically last one hour and  some long running UBEs run for more than three hours before completion  representing further  continuous load on the database server     Such a combination of interactive  short UBE and long UBE workloads provide a good workload   stressing all aspects of the solution   s configuration  Tables 11 and 12 list the transaction types for the  long running and short running UBE jobs  respectively     TABLE 11  LONG RUNNING UBE JOBS       TRANSACTION TYPE DESCRIPTION                         R09801 General Ledger Post   R31802A Manufacturing Account Journal  R31410 Work Order Processing   R3483 MRP Processing   R42520 Print Pick Slips Batch   R42800 Sales Order Update   R43500 Purchase Order Print       38    Oracle Optimized Solution for JD Edwards Enter
33. isk LSIDe11MR9261 812 0  1167966208 Blocks  597 GB    21    Oracle Optimized Solution for JD Edwards EnterpriseOne       In this case  only three drives ate visible since there are now only three logical volumes visible from the  RAID card     The OBP disk device paths that correspond to each of the individual logical volumes are therefore easy  to derive and in this case would be as follows    pci 400 pci 2 pcit 0 pci c Megaraid  sas 0 disk   pci 400 pci 2 pcit0 pci c Megaraid  sas 1 disk   pci 400 pci 2 pcit 0 pci c Megaraid  sas 2 disk    Target 0 will be the boot device  Create a device alias of disk that can be more easily used as the  default boot device    0  ok nvalias disk  pci 400 pci 2 pci 0 pci c Megaraid  sas 0 disk    Next set the server to automatically boot from the newly created device      0  ok setenv boot device disk    Note that when if setting up an Oracle Solaris JumpStart client for this machine  the logical volumes  appear as  in this case  c2t0d0  and c2t2d0   c2t3d0      Once completed  Oracle Solaris can be installed with and it should see only three disks  as have been  configured above  It is strongly encouraged that ZFS be used as the file system on all three disks   including the root disk  because future management of the server will be greatly simplified through the  extensive feature set offered by ZFS     B  DIL Benchmark Tools Configuration    The JD Edwards EnterpriseOne DIL workload is flexible and configurable  Optimizing the use of the
34. naged Battery  Backed Write Cache Accelerated and  Protected    Figure 2  Logical and physical volume configurations for the storage infrastructure     Oracle Optimized Solution for JD Edwards EnterpriseOne       TABLE 4  ZFS STORAGE POOL CONFIGURATIONS                ZFS POOL ZFS  NAME HARDWARE CONFIGURATION CONFIGURATION DESCRIPTION  datapool    3x 300 GB SAS2 drives managed by the The logical volume Holds all Oracle Database files and  Sun Storage 6 Gb SAS PCle RAID HBA  is used to directly working data  Database  e Configured as a single 550 GB host the datapool performance benefits significantly  RAID 5 protected logical volume  ZFS pool  No from the rapid write  N aye dha es additional data acknowledgements provided by the  e Write performance is significantly enhanced    ey         protection is NVRAM write cache   through the NVRAM write cache provided i  required   by the HBA   logpool e 3 x 300 GB SAS2 drives managed by the The logical volume Holds all Oracle Database redo log  Sun Storage 6 Gb SAS PCle RAID HBA is used to directly files  Database performance  e Configured as a single 550 GB host the logpool benefits significantly from the  RAID 5 protected logical volume  ZFS pool  No separation of the log files onto a    anaes additional data separate file system and also from  e Write performance is significantly enhanced Ane f        protection is the rapid write acknowledgements  through the NVRAM write cache provided y i    required  provided by the NVRAM writ
35. ne       Increase the maximum number of semaphores allowed per semaphore set  Default   512   jds zonel gt  projmod  s  K  process max    sem nsems   privileged  4096 deny        default    Increase the maximum number of message queue identifiers  Default   128     jds zonel gt  projmod  s  K  project  max msg ids   privileged  2048 deny   default    Increase the maximum number of semaphores identifiers  Default   128   jds zonel gt  projmod  s  K  project  max sem ids   privileged  15320  deny        default    Increase the number of bytes for a message on the message queue  Default   64k   jds zonel gt  projmod  s  K  process  max msg qbytes   privileged  16834  deny      default    Increase the maximum number of message on a message queue  Default   8192   jds zonel gt  projmod  s  K  process  max msg messages   privileged  16384  deny        default    For security  add the following parameter to the  etc system file     set noexec_user_stack 1    And finally reboot the zone to make these changes occur     reboot  Once the domain is rebooted  the Oracle Database software can be installed     Note   The following changes are stated  but must be performed as a user that is appropriate to the  customer   s Oracle Database security requirements     Please make the following changes to the Oracle Database configuration through sqlplus as the  sysdba uset     alter system set processes   3000 scope spfile    alter system set open_cursors 2000 scope spfile    alter system set se
36. nfiguration through the Server Manager GUI  The  configuration changes will be reflected in the files jas ini and jdbj  ini     JAS   INI changes     PathCodes   PD900      MAXUser default is 150  MAXUser 1000    UserSession 36000000     PerfMON     transactionTimeout 60000    Time out value for requests sent to the Enterprise Server    default is 90000    enterpriseServerTimeout 180000      maxPoolSize  default value is 100    maxPoolSize 35    JDBJ INI changes          maxConnection default is 150  maxConnection 1000    connectionTimeout 1800000    resultSetTimeout 60000    32    Oracle Optimized Solution for JD Edwards EnterpriseOne       Creating and Configuring the Oracle Database Server Zone    Create a standard  default  sparse  zone based on the global  In this case  two network devices are used     one for external access  on the 10 6 162 x subnet  and another for internal private network access  between zones  on the 10 6 222 x subnet      zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg  zonecfg    zonecfg    ZOnec    Q    zonec    Q    zonecfg    zonecf    Q          zonecfg        z DBServer    DBServer gt  create    DBServer gt  set zonepath  rootpool DBserverroot DBServer   DBServer gt  add net    DBServer net gt  set physical igb0O   DBServer net gt  set address 10 6 162 22 22   DBServer net gt  set defrouter 10 6 163 254   DBServer net gt  end    DBServer gt  add net    DBSer
37. o create the required three logical volumes as described earlier  on an 8 disk server using all 8 disks   the commands required would be     Make a RAID 1 volume out of drives 0 and 1   0  ok eli  CfgLdAdd  r1 252 0 252 1   Make a RAID5 volume out of drives 2 3 4     0  ok cli  CfgLdAdd  r5 252 2 252 3 252 4     Make a RAID5 volume out of drives 5 6 7   0  ok cli  CfgLdAdd  r5 252 5 252 6 252 7           Note that it is important to create the logical volumes in this order since when Oracle Solaris is  installed  it will  by default  install on the first disk it sees  unless configured otherwise   and that should  be the RAID 1 mirrored root volume     List the details and status of the logical drives by running this command    0  ok cli  LdPdInfo    Next create an alias to the bootable drive that was made  This is not strictly necessary but makes future  administration easier by reducing the amount of typing and possibilities for mistakes that accompany    long path names     First look to see what drives OBP currently sees  This may take a few seconds on systems with lots of  drives and especially if Fibre Channel drives have been connected      0  probe scsi all    Drives that are under the control of the RAID card will be listed under a path similar to this               pci 400 pci 2 pcit0 pcic Megaraid  sas 0  Target 0   Unit 0 Disk LSIDe11MR9261 81i2 0  583983104 Blocks  298 GB  Target 1   Unit 0 Disk LSIDe1l1MR9261 812 0  1167966208 Blocks  597 GB  Target 2   Unit 0 D
38. of semaphores allowed per semaphore set  Default   512   jds zonel gt  projmod  s  K  process  max sem nsems   privileged  4096 deny        default    Increase the maximum number of message queue identifiers  Default   128     jds zonel gt  projmod  s  K  project  max msg ids   privileged  2048 deny   default    29    Oracle Optimized Solution for JD Edwards EnterpriseOne       Increase the maximum number of semaphores identifiers  Default   128     jds zonel gt  projmod  s  K  project  max sem ids   privileged  15320  deny     default    Increase the number of bytes for a message on the message queue  Default   64k   jds zonel gt  projmod  s  K  process  max msg qbytes   privileged  16834  deny      default    Increase the maximum number of message on a message queue  Default   8192   jds zonel gt  projmod  s  K  process  max msg messages   privileged  16384  deny        default    For security  add the following parameter to the  etc system file     set noexec_user_stack 1    And finally reboot the zone to make these changes occur     reboot    Once the domain is rebooted  the JD Edwards EnterpriseOne software can be installed     Creating and Configuring the WebLogic Web Server Zone    Create the sparse zone based on the global zone OS  Do this in the global zone   zonecfg  z WebServer  zonecfg WebServer gt  create  b  zonecfg WebServer gt  set zonepath  rootpool Webserverroot WebServer  zonecfg WebServer gt  add net  zonecfg WebServer net gt  set physical igb0  zonecfg We
39. ontainers enable optimized resource utilization to deliver predictable  service levels     ZFS file system   ZFS radically simplifies the management of complex storage subsystems and  provides seamless data integrity and portability functions at no cost to the user  Performance  optimizations leverage flash based storage to give extreme performance with minimal administrative  overhead     Networking   The open  virtualizable  and programmable Oracle Solaris networking stack delivers  high throughput and the option to utilize integrated hardware encryption at zero cost     Oracle Optimized Solution for JD Edwards EnterpriseOne       JD Edwards EnierpriseOne 8 98 9 0    JD Edwards EnterpriseOne is an integrated application suite of comprehensive enterprise resource  planning software that combines business value  standards based technology  and deep industry  experience into a business solution with a low total cost of ownership  A collection of prebuilt  Business Intelligence  BI  applications available as modules for JD Edwards EnterpriseOne provide  organizations the ability to implement and integrate more quickly  with less tisk  and at a fraction of the  cost required for building traditional BI solutions  JD Edwards EnterpriseOne provides organizations  with the ability to transform information into business actions and enjoy a strategic advantage over    less nimble competitors     Oracle Database 11g Release 2 Standard Edition    Designed for data center environments t
40. owing setting is placed in the  etc system file in the global  zone  as a secutity feature    set noexec_user_stack 1  To simplify network communications between the zones  host names  on private subnets  should be  populated into the global zone  etc hosts  Aliases to the name of the zone alongside the actual    host name adopted by the zone will make management easier     10 6 222 241 jds zonel ElServer  10 6 222 84 jds zone2 WebServer  10 6 222 79 jds02 DBServer    Mount additional storage devices for use in the zones as data storage     25    Oracle Optimized Solution for JD Edwards EnterpriseOne       From the output of the format command  the three logical volumes can be determined  as required   to allow proper configuration of the required ZFS pools and subsequent ZFS file systems that will be  used by the zones     bash 3 2  format    Searching for disks   done    AVAILABLE DISK SELECTIONS   0  clt0Od0  lt LSI MR9261 81 2 50 cyl 65533 alt 2 hd 287 sec 31 gt    pci 400 pci 2 pcit0 pci c LSI mrsas 0 sd 0 0  Ts GLtla0  lt LSI MR9261 81 2   12 556  93GR gt    pci 400 pci 2 pcit0 pci c LSI mrsas 0 sd 1 0  2  c1t2d0  lt LSI MR9261 8i 2 12 556 93GB gt    pci 400 pci 2 pci 0 pci c LSI mrsas 0 sd 2 0  Specify disk  enter its number    C     bash 3 00     In this listing  the three devices containing the LSI MR9261 naming are the RAID configured  Logical Volumes provided by the SAS RAID card     The SPARC T4 1 test system was installed with Oracle Solaris 10 8 11  pre release 
41. owing very rapid  battery backed  write completion  acknowledgement through NVRAM write cache  This HBA is used to provide three logical disk  volumes to the platform optimized for write performance and data protection     The logical volumes provided by the hardware RAID controller are managed by Oracle Solaris under  the control of ZFS  providing greatly enhanced features for rapid import  export  and check  pointing rollback of any file system  which provides rapid paths to recovery in the event of failure and  implicit data integrity checking and correction     By appropriate positioning of data on these logical volumes  considerable performance gains have been  achieved compared to standard disk configurations  Performance gains are such that the use of the  in chassis SAS2 storage devices provides sufficient performance without the need for the added  expense and complication of external storage arrays     TABLE 3  STORAGE CONFIGURATION       VOLUME CONFIGURATION DESCRIPTION         Exposed as a single logical drive  provides the Global Zone root  2 x 300 GB SAS2 disks RAID 1    VOLUME 1   4 OS boot disk for the platform and root images for the containers for  mirrore  the other virtual servers          VOLUME 2 3 x 300 GB SAS2 disks RAID 5 Exposed to the operating system as a single logical drive  used by  mirrored ZFS as a storage pool for Oracle Database  3 x 300 GB SAS2 disks RAID 5 Exposed as a single logical drive  used by ZFS as a storage pool  VOLUME 3    mirror
42. priseOne       TABLE 12  SHORT RUNNING UBE JOBS                                                                                     TRANSACTION TYPE DESCRIPTION   R0004P UDC Records Types Print  ROOO6P Business Unit Report   R00067 Business Unit Translation Report  R0008P Date Patterns Report   R0010P Company Constants Report  R0012P1 AAI Report   R0014 Payment terms Report  R0018P Tax Detail Report   R00425 Organization Structure Report  R014021 One Line per Address Report  R01402W Who   s Who Report   R01403 Full Address Report with Codes  R03B155 A R Summary Analysis  R03B31 Activity Log Report   R03B461 Collection Report   R04413 Open A P Summary Report  R04602 Supplier Analysis Report  R09017 Account Translation Report  R09205P Account Master Report  R09410 Trial Balance Report   R098201 Annual Close Report   R41411 Select Items Cost Count  R41543 Item Ledger Account Integrity  R42072 Price Category Print   R46990 Warehouse Location Integrity  R48096 Cost Plus Markup Listing       39    ORACLE       Oracle Optimized Solution for JD Edwards  EnterpriseOne   October 2011  Version 1 2   Authors  Peter Wilson  Laura Verdibello    Oracle Corporation   World Headquarters   500 Oracle Parkway  Redwood Shores  CA 94065  U S A     Worldwide Inquiries   Phone   1 650 506 7000  Fax   1 650 506 7200    oracle com    as   Oracle is committed to developing practices and products that help protect the environment    Copyright    2011  Oracle and or its affiliates  All rights reserved  
43. racle Database or Oracle WebLogic Server products  or if the broader implementation calls  for use of these products by applications beyond JD Edwards EnterpriseOne  customers have the  option of acquiring normal  full use licenses instead     Solution Hardware Components  The subsections below describe the server and storage components that were selected for this Oracle    Optimized Solution     SPARC T4 1 Server    Oracle   s SPARC T4 processor based servers are the platform of choice for a full range of data center  applications and workloads  The SPARC T4 1 server was selected for the Oracle Optimized Solution  for JD Edwards EnterpriseOne because of the key characteristics described in Table 2     TABLE 2  KEY CHARACTERISTICS OF THE SPARC T4 1 SERVER THAT MAKE IT THE RIGHT CHOICE       SPARC T4 1 SERVER SYSTEM CHARACTERISTIC KEY BENEFITS FOR THE SOLUTION       SPARC T4 processor A single processor with 8 cores and 64 threads based on the  latest SPARC core with enhanced single threaded  performance allows very effective provisioning of resources  and enables the SPARC T4 1 server to run many virtual  servers  helping drive up utilization  lower IT costs  and keep  server sprawl to a minimum        High memory capacity and bandwidth Deployed with typically 64 GB of memory  but capable of up  to 256 GB   the SPARC T4 1 server provides the capability of  holding large volumes of in memory data  optimizing Web  and database response times  and providing a large  workspace to
44. s of the card  there is a large amount of output to collect  Ensure you  have a large history or scroll back buffer on the terminal you are using  You will refer back to this  information later     0  ok cli  AdpAllInfo    Check to see that all the physically installed drives are visible to the controller and reporting correctly   All drives should report as 6 Gbps SAS2 devices and 300 GB drives will report with a real world  capacity of 279 GB     0  ok cli  PdList    Check to see if any logical volumes have already been configured on the server  Check the manual for  the required procedures needed to destroy any pre existing logical volumes      0  ok cli  LdPdInfo    You can now create the required set of logical volumes  There will be three volumes created   1  RAID1 mirrored Root disk volume  2  RAID5 mirrored Data volume for the database    3  RAID5 mirrored Log volume for the Database Server redo log files    20    Oracle Optimized Solution for JD Edwards EnterpriseOne       To create a logical volume  the syntax is similar to this example     0  ok cli  CfgLdAdd  r1 252 0  252 1   r1 means we are creating a RAID1 volume  252 is the enclosure number  which can be determined  from the output of the earlier PAList command  and 0 and 1 are the disk slots in the enclosure that    are to be used to create a mirrored pair of drives  This command RAID1 mirrors drives 0 and 1 into a  logical drive that will then be visible to the operating system that will be loaded later    T
45. s true of JD Edwards EnterpriseOne  which at a  minimum requires a database management system and a Java  Web  server     In the context of the Oracle Optimized Solution for JD Edwards EnterpriseOne  the database is  Oracle Database and the Java server is Oracle WebLogic Server  To provide    complete solution  licensing    that mimics the    Optimized Solution engineering     Oracle offers the licensable product  called Oracle Technology Foundation for JD Edwards EnterpriseOne  This product provides  limited use licensing of Oracle Database Standard Edition  Oracle WebLogic Server Standard Edition   and several additional components of Oracle Fusion Middleware for use with JD Edwards    EnterpriseOne deployments     Oracle Optimized Solution for JD Edwards EnterpriseOne       Because Oracle Technology Foundation for JD Edwards EnterpriseOne is licensed per application  user  not per CPU  these components can be installed on an unlimited number of processors   providing some protection of software licensing costs due to CPU sprawl  An additional benefit is the  ability to implement multiple  identical  hardware deployments for no additional license fees  This  allows customers to provide development  QA  and testing instances of their deployments in an  extremely cost effective manner while further reducing development times and enhancing  speed to deployment     In cases where customers find value in the extended feature set offered by the Enterprise Edition of  either the O
46. s zones  it is possible to assign CPU resources to each zone  but this was not found to  be necessary in this configuration since during testing  CPU resources where not a limiting factor to  performance     To take advantage of the performance enhancements available in JDK1 6  the system level link   usr java should be checked to ensure that it points to jdk1   6 0  by default   usr  java  points to jdk1 5 0 for compatibility with some older applications      Creating and Configuring the JD Edwards EniterpriseOne Server Zone    Create a standard  default  sparse  zone based on the global zone  In this case  two network devices are  used  one for external access  on the 10 6 162 x subnet  and another for internal private network access  between zones  on the 10 6 222 x subnet      zonecfg  z ElServer             zonecfg ElServer gt  create   zonecfg ElServer gt  set zonepath  rootpool Elserverroot ElServer  zonecfg ElServer gt  add net   zonecfg ElServer net gt  set physical igb0  zonecfg ElServer net gt  set address 10 6 162 23 22  zonecfg ElServer net gt  set defrouter 10 6 163 254  zonecfg ElServer net gt  end   zonecfg ElServer gt  add net   zonecfg ElServer net gt  set physical igb1  zonecfg ElServer net gt  set address 10 6 222 241 24  zonecfg ElServer net gt  set defrouter 10 6 222 254  zonecfg ElServer net gt  end   zonecfg ElServer gt  add fs   zonecfg ElServer fs gt  set dir  db   zonecfg ElServer fs gt  set special rootpool Elserverdata       27    Oracle Optimized
47. sections that  follow     SPARC T4 1 Server Configuration    The system under test represented a mid range SPARC T4 1 server configuration with an 8 disk  internal storage subsystem  A larger storage capacity configuration may utilize 600 GB disk drives in  place of the tested 300 GB drives     Memory Configuration    The SPARC T4 1 server has 16 memory DIMM slots  For maximum performance  it is recommended  that all slots be filled with identical density memory DIMMS  In this configuration  4 GB DIMMs  where chosen providing a total of 64 GB of RAM  To further increase RAM  it is recommended that  8 GB or 16 GB DIMMs be substituted for a total capacity of up to 256 GB of memory  If a small  configuration system is chosen with only 32 GB of memory  it is recommended that only eight 4 GB  DIMMs be installed allowing additional DIMMs to be installed at a later time  if needed     Storage Configuration    The SPARC T41 server incorporates an 8 disk storage chassis  A doubling of available storage  capacity can be achieved by adopting the larger 600 GB capacity disk drives available for this platform   however  for the DIL workload  the 300 GB disk drives provide more than enough capacity for these  tests     Oracle Optimized Solution for JD Edwards EnterpriseOne       The SPARC T4 1 server is provided with two onboard disk controllers  however  in this configuration   these are not used and instead they are replaced by a single higher performance advanced RAID HBA  with Battery 
48. ssion_cached_cursors 1000 scope spfile   alter system set pga_aggregate_target 0 scope spfile   alter system set sga_target 6G scope spfile    alter system set sga_max_size 7G scope spfile    alter system set memory_target 11G scope spfile     alter system set memory_max_target 11G scope spfile                 alter system set recyclebin OFF scope spfile     35    Oracle Optimized Solution for JD Edwards EnterpriseOne       Default sizes for the standard database files were increased in the following ways     e The system and sysaux tablespaces were resized to 1500 M and set to autoextend on next  10 M     e The tempfile tablespace was resized to 7000 M     e The number of redo logs was increased to 6 of 250 M sizes each     D  Workload Details  There ate two primary workloads provided by the DIL Workload used for the benchmark tests   e Interactive Transactions   These are user transactions that reflect interactive Web based operations     e UBE batch jobs   Short  and long running UBE batch job types are run during the test to increase  the realism of the workload     The details of each of these workloads ate described in the following section     DIL Workload    Oracle s DIL workload consists of a suite of scripts that simulate user interactions through a load  generation system  The simulated transactions exercise common operations of JD Edwards  EnterpriseOne applications  including business processes such as various interactions related to payroll   sales orders  purch
49. t achieves substantially higher performance than a standalone SPARC T4 1  server  providing a range of architectures to choose from depending on the requirements of the  particular deployment  Though not tested as part of this solution  it is discussed here for informational    purposes     Table 7 shows the recommended hardware components for the large configuration     TABLE 7  RECOMMENDED HARDWARE COMPONENTS FOR THE LARGE CONFIGURATION       SERVER VARIANT COMPONENT CONFIGURATION       Processor  1 x SPARC T4  2 85 GHz  8 core     Factory pre installed   8 threads core       Memory  128 GB RAM  utilizing 16 x 8 GB Ea      SPARC T4 1 Server DIMM As indicated in service manual   s       Install disks in slots 0 and 1 and mirror using  the onboard SAS HBA     Disks  2 x 300 GB SAS2 2 5 inch disk drives       Processor  1 x SPARC T4  2 85 GHz      Factory pre installed   8 core  8 threads core       Memory  64 GB  utilizing 16 x 4 GB DIMMs As indicated in service manual        SPARC T4 1 Server        s i i   Install in slots 0 and 1 and mirror for Root disk  Disks  2 x 146 GB SAS2 2 5 inch disk drives    protection        HBA  Emulex Dual Port Fibre Channel aa      As indicated in service manual        FC_AL HBA  Oracle   s Sun Storage Dual RAID Controllers  12 x 300 GB 3 5 inch ers        As indicated in service manual   2540 M2 Array FC Disks       The large configuration is provided to demonstrate that the SPARC T4 1 server remains central to the  solution and provides the
50. tentanede aamneses 14  Test Runs 6 and 7  Best Compromise Performance for Long and  Short UBEs with Acceptable Response Times                 cceeeeee 14  FAN ANA SIS so oe kine a led sa ec eed A Slee aad 15  Real World Sizing Based on the DIL Test Workload                   000 1S  ROOM for CONSOMOAUION  sse erer e ERN seca 15  Sizing Guidelines 37 8  elh nes aeckte din Ser aeere eda  oiediteanes tele mcenecaes 16  OMCIUSION  gic AOL E tet ot E A ROSA he Ad 19  References css seine ea oa anaes See es eG EE ESSE 19  POPENCICES   cus ieta auton ee neues Eea aaa weet IE 20  A  Configuring the Sun Storage 6 Gb SAS PCle RAID HBA Card 20  B  DIL Benchmark Tools Configuration                    ccceeeeeeeeeeeeeeees 22  C  Tuning and Best Practices sii 1 5cxccesnassetiniensthetenenesd seers ete aera 25    D  Workload Deg e a e er een a a N E E 36    Oracle Optimized Solution for JD Edwards EnterpriseOne       Introduction    Whether organizations are deploying a new application service or modernizing existing  business critical services  selecting and implementing a solution architecture that will  perform optimally and meet service level agreements is a challenging process  An  optimized solution must balance technical requirements against business needs such as  budget constraints  performance and reliability requirements  and future scalability needs     This paper describes Oracle   s extensive testing of a hardware and software combination  optimized to efficiently host the JD 
51. teractive user response times   Tests 6 and 7 demonstrate overall system capabilities when four or two long UBEs are maintained  along with a maximal level of short UBEs while maintaining acceptable system response times close to  0 5 seconds     With four long UBEs and a throughput of 1741 short UBEs  response time was maintained at 0 583  seconds and the system was 100  utilizing disk I O capabilities throughout the test run     During test run 7  two long UBEs where maintained with a throughput of 2024 short UBEs and a  response time of 0 588 seconds  The system has vety little spare disk I O capacity during this test   averaging 81  and peaking at 100      Oracle Optimized Solution for JD Edwards EnterpriseOne       Test 7 demonstrates the maximum DIL throughput of the system under the expected conditions of  mixed long and short UBE loads     Final Analysis    The SPARC T4 1 server is extremely capable of handing the workload at these levels  the single  limiting factor being disk I O capacity on the integrated disk sub system  To achieve higher results  requires additional hardware in the form of additional higher performance disk systems or dedicated  off load of the database server and database storage to a standalone platform     At no time during testing were the CPU or memory of the server more than 27   CPU  or 74    memoty  utilized  which indicates that if the disk I O capacity limitations where improved  as  discussed earlier  overall significant performance in
52. tion and avoids the  need to use more expensive flash based or external storage devices  such as Fibre Channel arrays     References    TABLE 8  REFERENCES       DESCRIPTION URL       JD Edwards EnterpriseOne Product page http   www oracle com us products applications jd edwards        enterpriseone index html          SPARC T4 1 server page http   www oracle com us products servers   storage servers sparc enterprise t series sparc t4 1 ds   487858 pdf          Oracle   s flash technology page http   www oracle com us products servers        storage storage flash storage index html                Oracle JD Edwards Partner Portal http   www go9withide com   Oracle JD Edwards EnterpriseOne Documentation http   www oracle com technetwork documentation jdedent   098169 html    Oracle Optimized Solution for JD Edwards EnterpriseOne       Appendices    A  Configuring the Sun Storage 6 Gb SAS PCle RAID HBA Card    The initial configuration of the Sun Storage 6 Gb SAS PCIe RAID HBA card requires that the server  be initially configured at the OBP prompt  before an operating system is booted  or installed      From the OBP ok prompt        0  ok show devs    check the output for a line that looks like   pci 400 pcit 2 pcit0 pci c Megaraid  sas 0  This is the root path to the controller        Select that device  note that the initial space at the front of the quoted text is important    0  ok    pci 400 pci 2 pcit 0 pci c Megaraid  sas 0  select    dev    Now get a report on the statu
53. ts simplifies many aspects of security and    management   e Portability and backup options are improved through integration with ZFS     e Oracle Solaris Containers enable fine grained control over access to system resources for tuning  purposes     e Systems consolidation and interconnect elimination significantly reduces data center footprint and  simplifies management     Compared to competitive configurations consisting of multiple servers and storage devices  the  consolidated approach offered by the Oracle Optimized Solution for JD Edwards EnterpriseOne  provides greatly simplified deployment and management     Oracle Optimized Solution for JD Edwards EnterpriseOne       Solution Architecture Overview    The solution architecture takes advantage of Oracle Solaris Containers and optimizes the use of  onboard storage to take best advantage of the resources and bandwidth available within the platform   Figure 1 shows how the storage devices are mapped to specific containers within the SPARC T4 1    server     Oracle Container Environments         Container 3       Host Container 1 Container 2                 JD Edwards  EnterpriseOne    Solaris wos Oracle 11g R2 Oracle  Image an    aa Database WebLogic    Management 10 3 2          Figure 1  Technical architecture for the Oracle Optimized Solution for JD Edwards EnterpriseOne     The SPARC T4 1 server performs four primary roles  as identified by the four functional areas listed in  Figure 1  The base platform provides th
54. u 237445  g oinstall  G dba  d  export home oracle  m       s  usr bin bash oracle    passwd    oracle oracle    Log in to the oracle users home directory using su   oracle and add the following paths to    the  profile    export  export  export  export    export    file in the home directory    PATH  usr bin   usr sbin   usr ucb  etc  usr openwin bin     WL_HOME  export home weblogic wlserver_10 3   MSLOGS  export  home weblogic user_projects domains wll1_base_domain  NODEMGR_HOME     WL_HOME    common nodemanager    DOMAIN_HOME  export home weblogic user_projects      domains wl1_base_domain    For security  add    the following parameter to the  etc systen file     set noexec_user_stack 1    Next reboot the domain to let these settings take effect     reboot    31    Oracle Optimized Solution for JD Edwards EnterpriseOne       And finally install the Oracle WebLogic server software according to the paths and user names that  have been established above     Once the Oracle WebLogic software is installed  some further tuning should be carried out     Five JVMs were used per instance on the Oracle WebLogic server during the 1200 user test  Each JVM  used a minimum and maximum heap size of 4 096 MB  NewSize and MaxNewSize where set to  1344m  These values where modified through the Oracle Enterprise Manager Console  which can be  accessed via the following URL     http     lt WebLogicServerHostname_or_IP gt  7001 console login    Further modifications are made to the server co
55. um of three disks      The server configuration suggests 64 GB RAM be installed as sixteen 4 GB DIMMs  This  configuration fully populates the DIMM slots in the server with memory and provides maximum  performance at this memory density  If room for memory expansion is required  it is suggested that the  server initially be populated with only eight 8 GB DIMMs  half populating the system  leaving easy  incremental space for an additional eight DIMMS when required  This will impose a small performance  penalty due to the non fully populated memory system not achieving its full bandwidth  but this is not  likely to be an issue since memory is not a limiting factor in system performance     For storage management  the Sun Storage 6GB SAS RAID PCIe RAID HBA is selected and should be  installed according to standard service or installation documentation     Oracle Optimized Solution for JD Edwards EnterpriseOne       Larger Configurations   Above 300 Real World Users    Larger configurations of the system were not directly part of the testing for the Oracle Optimized  Solution for JD Edwards EnterpriseOne  However  the SPARC T4 2 server has been tested in  higher capability configurations and has demonstrated capacities as high as 10 000 interactive users     Further testing and performance enhancements described here will be carried out in a subsequent  version of this solution that mirrors this configuration using SPARC T4 1 servers to provide a stepping  stone configuration tha
56. ver net gt  set physical igbl   DBServer net gt  set address 10 6 222 79 24   DBServer net gt  set defrouter 10 6 222 254   DBServer net gt  end    DBServer gt  add fs    DBServer fs gt  set dir  db    DBServer fs gt  set special datapool DBserverdata   DBServer fs gt  set type zfs    DBServer fs gt  end    DBServer gt  verify    DBServer gt  commit    DBServer gt  info    DBServer gt  exit    If DTrace will later be used to investigate the behavior of the zone  this additional setting  should be issued     zonecfg    zonecfg   zonecfg   zonecfg     zonecfg      z DBServer    DBServer gt   DBServer gt   DBServer gt     DBServer gt     set limitpriv default  dtrace_proc dtrace_user  verify  commit    exit    Now the zone is ready to be booted and further configurations can be carried out inside the OS image     zoneadm  z DBServer install    zoneadm    list  cv    zoneadm  z DBServer boot    Log in to the server and carry out the following operations to create the required user profiles     zlogin  C DBServer    usermod  s  usr bin bash root    33    Oracle Optimized Solution for JD Edwards EnterpriseOne       If root    changes     should be allowed to directly log in to the zone  this can be enabled with the following    Edit the file  etc ssh sshd_config  and change the line   PermitRootLogin no   To read    PermitRootLogin yes   And then restart the ssh daemon with the command     svcadm restart ssh    Create the users required for the JD Edwards EnterpriseOne installation
    
Download Pdf Manuals
 
 
    
Related Search
    
Related Contents
  User`s Manual - J Flow Controls  precaución - Brunswick Marine in EMEA Center    BF5A/BF4.5B MANUAL DO PROPRIETÁRIO  veget pro nature    様式第 7号(第 6条関係) 岩 崎 信 様 延選第 号 平成 26年 2月 26日  LE-15/30 - interhospitalar  Finlux 50F6090 User's Manual    Copyright © All rights reserved. 
   Failed to retrieve file