Categories

JAVA DATEBASE
Technology Network Community
Oracle Database
Fusion Middleware
Development Tools
Java
Desktop
Server & Storage Systems
Enterprise Management
Berkeley DB Family
Cloud Computing
Big Data
Business Intelligence
Architecture
Migration and Modernization
E-Business Suite
Siebel
PeopleSoft Enterprise
JD Edwards World
Industries
JD Edwards EnterpriseOne
User Productivity Kit Pro (UPK) and Tutor
Governance, Risk & Compliance (GRC)
Master Data Management (MDM)
Oracle CRM On Demand
On Demand: SaaS and Managed Applications
AutoVue Enterprise Visualization
Primavera
ATG
Agile PLM
Endeca Experience Management
Fusion Applications
Archived Forums

 



Tags

EBS General Discussion


File Copy operation and R12.1.3 Instance performance is very slow


Hi All,        We are using sun solaris SPARC 10 OS 64bit on Sun sparc server T3-2 server with 256 threads and 64GB RAM 113 GB Swap size. We have configured R12.1.3 Instance through OBA setup tool.  Now the thing is currently we are in implementation stage for our client. Our client is saying instance operations is very slow. Forms are opening slowly and the instance response is totally slow.  we analyzed instance there is no errors except the cpu error (Database will not support more than 64 cpu in server). But any how with 64 Cpu instance should work normally am i right?  Then we analysed in os level and we found the Operating system responses is very poor. Especially copying 10GB files it is taking more than 1 hour. We raised about this to hardware vendor he is saying the problem is with your instance like this.  We configured RAID Level 5 for oracle application the OS mount point configured with RAID 1 + 0.  We had run Gather schema statistics concurrent and run adstats.sql script from sys user. But the performance is same. While installation all the pre-req are passed with out fails. but installation was taken more time.  Please help us to ensure the things whether it is hardware issue or Instance issues.

We are using sun solaris SPARC 10 OS 64bit on Sun sparc server T3-2 server with 256 threads and 64GB RAM 113 GB Swap size. We have configured R12.1.3 Instance through OBA setup tool.  Please see if (Migration from fast single threaded CPU machine to CMT UltraSPARC T1 and T2 results in increased CPU reporting and diminished performance [ID 781763.1]) is applicable.  Now the thing is currently we are in implementation stage for our client. Our client is saying instance operations is very slow. Forms are opening slowly and the instance response is totally slow.  Enable trace to find out why it takes that long -- https://forums.oracle.com/forums/search.jspa?threadID=&q=Enable+AND+Trace&objID=c3&dateRange=all&userID=&numResults=15&rankBy=10001  Also, please see the docs referenced in old threads which discuss the same topic.  https://forums.oracle.com/forums/search.jspa?threadID=&q=Performance+Tuning+AND+R12&objID=c3&dateRange=all&userID=&numResults=15&rankBy=10001 https://forums.oracle.com/forums/search.jspa?threadID=&q=Performance+AND+Tuning+&objID=c3&dateRange=all&userID=&numResults=15&rankBy=10001  we analyzed instance there is no errors except the cpu error (Database will not support more than 64 cpu in server). But any how with 64 Cpu instance should work normally am i right?  Typically yes, but not necessarily -- Depends on other factors.  Then we analysed in os level and we found the Operating system responses is very poor. Especially copying 10GB files it is taking more than 1 hour. We raised about this to hardware vendor he is saying the problem is with your instance like this.  Can you reproduce the issue if you shutdown Oracle? If yes, then ask the vendor to fix the issue.  We configured RAID Level 5 for oracle application the OS mount point configured with RAID 1 + 0.  We had run Gather schema statistics concurrent and run adstats.sql script from sys user. But the performance is same. While installation all the pre-req are passed with out fails. but installation was taken more time.  Please help us to ensure the things whether it is hardware issue or Instance issues.  Looks like the issue is not with Oracle EBS application/database and it is something to do with your OS/Network/Storage configuration, so I would suggest you shutdown all the services on this server and check the performance -- Report it to the vendor (or to your network/storage/admin) team.  Thanks, Hussein  Thanks, Hussein

We put a test clone of 11.5.10 on a T2 Niagra server and the load average frequently went above 200 running MRP/DRP.  The machine really shined doing the hammerora or TPCC benchmarks, but was barely able to handle having eBusiness Suite installed and running.

Hi Hussain,  Please see if (Migration from fast single threaded CPU machine to CMT UltraSPARC T1 and T2 results in increased CPU reporting and diminished performance [ID 781763.1]) is applicable.  The docs well noted.   Enable trace to find out why it takes that long   We already enabled the trace and sent the report to oracle support and we are analysing. This is for your kind information.  Typically yes, but not necessarily -- Depends on other factors  We daily starting database with 64threads. I understood there are some alternativie solutions for T3 cpu issues from sun hardware / oracle support.  Please give step by step to apply this cpu issue.  Can you reproduce the issue if you shutdown Oracle? If yes, then ask the vendor to fix the issue.  Yes we reproduced but same performance for copy operations. for ex: To copy 2.5Gb data it is taking 18Mins to complete this operation.  Thanks  Hemanth

Hi Hemanth,  Please give step by step to apply this cpu issue.  What CPU issue you are referring to? Please elaborate more.  Yes we reproduced but same performance for copy operations. for ex: To copy 2.5Gb data it is taking 18Mins to complete this operation.  Do you have any other server to test the same and see how long it takes so you can compare?  Thanks, Hussein

Hi Hussein   What CPU issue you are referring to? Please elaborate more.  We are having SUN T3-2 Server with 2 SPARC T3 16-core 1.65 GHz Processor (256 CPU). While installation of R12.1.3 we faced the below issue   Error : ORA-00821: Specified value of sga_target 1024M is too small, needs to be at least 1452M  Cause: OBA Database not supporting more than 64 core CPUs  Solution: Users can turn off some of the hardware threads  To disable core processors - # psradm -f 63-255  As per the document Doc ID 1116525.1 we just disabled the cups and started the new installation. finally we completed the installation and executed the oba diagnostic script and uploaded the script to oracle OBA Team. they also certified to proceed forward.  So the thing is now we are starting our Application with only 64 CPU. So if there is any possibilities like  Database patch or OS patch or any alternative available can you please guide me?. our client is feeling that this only the cause for instance performance issue.  We are having 16GB shm_max value in /etc/project file. And we are having 8GB sga_target memory. Our total RAM size is 64GB, swap size is 113GB. This is for your information. Please advice Memory parameters for this configuration EBS R12.1.3 environment on sun sparc OS 10  5.10.  Is the SGA target size and shm_max_memory ok for us? We followed installation doc only for setting kernal parameters value.  I would like to highlight this point if we are accessing instance and opening forms it is responding slowly some small forms opening with in the seconds . But each forms are taking Min 6seconds to open. From the login page itself it is taking long time to respond.  Do you have any other server to test the same and see how long it takes so you can compare?  I understood that you mean for copy operation. We are having one sun generic server with sun os 10 not sparc. We tested the copy operation in that server it is taking only 42Mintes to copy 22GB data with in the same server mount points. that machine ount points configured in RAID 5.  There is no test server in our client. We are using Prod server for our implementation. we are using clone instance in the server. Prod instance is down. Any how client released PO to buy new test server.  You already highlighted that the issues is not in our EBS and database right? could you once again ensure that with the current informations.  Due to some Technical reasons i can't able to update earlier.  Thanks Hemanth

Hi Hemanth,  Error : ORA-00821: Specified value of sga_target 1024M is too small, needs to be at least 1452M  Cause: OBA Database not supporting more than 64 core CPUs  Solution: Users can turn off some of the hardware threads  To disable core processors - # psradm -f 63-255  As per the document Doc ID 1116525.1 we just disabled the cups and started the new installation. finally we completed the installation and executed the oba diagnostic script and uploaded the script to oracle OBA Team. they also certified to proceed forward.  So the thing is now we are starting our Application with only 64 CPU. So if there is any possibilities like  Database patch or OS patch or any alternative available can you please guide me?. our client is feeling that this only the cause for instance performance issue.  This error is report in the installation doc, under "Known Issues" section -- Oracle E-Business Suite Installation and Upgrade Notes Release 12 (12.1.1) for Oracle Solaris on SPARC (64-bit) [ID 761568.1]  We are having 16GB shm_max value in /etc/project file. And we are having 8GB sga_target memory. Our total RAM size is 64GB, swap size is 113GB. This is for your information. Please advice Memory parameters for this configuration EBS R12.1.3 environment on sun sparc OS 10  5.10.  Is the SGA target size and shm_max_memory ok for us? We followed installation doc only for setting kernal parameters value.  Database Initialization Parameters for Oracle Applications Release 12 [ID 396009.1] bde_chk_cbo.sql - EBS initialization parameters - Healthcheck [ID 174605.1] Tuning performance on eBusiness suite [ID 744143.1]  I understood that you mean for copy operation. We are having one sun generic server with sun os 10 not sparc. We tested the copy operation in that server it is taking only 42Mintes to copy 22GB data with in the same server mount points. that machine ount points configured in RAID 5.  Any reason you are using RAID 5?  There is no test server in our client. We are using Prod server for our implementation. we are using clone instance in the server. Prod instance is down. Any how client released PO to buy new test server.  Any feedback from the hardware and OS vendor about the performance issue when copying the files?  You already highlighted that the issues is not in our EBS and database right? could you once again ensure that with the current informations.  Please go through the docs referenced above and make sure you have all the initialization parameters set properly. Also, log a SR and ask for an update about the database patch that Oracle is releasing as mentioned in R12 installation doc (under Known Issues section).  If you encounter performance issues at the OS level, then I would say that this is not an Oracle EBS issue. Once this issue is fixed, and your application/database still not acting as expected from the performance perspective then we need to do more investigation about the performance.  Thanks, Hussein

Hi Hussein,   Thanks a lot'  I will check pathces with the help of SR.  Raid level desired by our client.  Can you say perfect raid level for oracle EBS with data redundancy.  No feedback from hardware vendor. We asked to reinstall OS nad Format the /oracle Mount point in that mount point we are running our apps and DB.  Thanks Hemanth

Can you say perfect raid level for oracle EBS with data redundancy.  Please see the links referenced in this thread -- https://forums.oracle.com/forums/thread.jspa?threadID=952709  Thanks, Hussein

Hi hussain,   Can you confirm that file copying operation in SPARC T3-2  server with sun sparc 10 OS is normal or abnormal case?  It's taking 1hour to complete 10Gb data copy operation.Please advice.  Thanks Hemanth.C

Hi hussain,  Thanks a lot hussain !   For copying performance issue hardware vendor is suggesting one OS patch.  We are awaiting for OS patch from hardware Er. After we will check our application performance.  This is for your information.  Thanks Hemanth

Hi Hemanth,  Keep us updated!  Thanks, Hussein

Hi Hussain,   Now hardware vendor reconfigured the RAID level in /oracle Mount point from RAID 5 to RAID 1+0. Now copy operation working fine it is taking 45Min to complete 100GB data copy operation.    But R12.1.3 is not working fine. Same performance is there. Oracle Updated the SR as it's our OBA product bug. This is the last update from oracle. Oracle Dev team is working on this.  Thanks Hemanth.C


Related Links

Post Installation error in r12
Loading of Java error with java.io.IOException from E-Business Suite 12
HRMS_PF.K
use of scripts in $OAD_TOP/admin/scripts/CONTEXT
_SQLEXEC_PROGRESSION_COST to 0
Question regarding workflow
mtl_serial_numbers & mtl_serial_numbers_temp
Import Discoverer Content for E-Business Suite Release 12 using adupdeul.sh
xml report for excel output
frm-40502: oracle error: unable to read list of values
ORA-04031:
Apps 11i clone RAC to Non RAC
Runaway Processes in our EBS 11i Server
Interoperability Notes Oracle EBS 11i with Oracle Database 11gR2 (11.2.0.1)
How to cancel all alert email in WF_NOTIFICATIONS_OUT Queue.
Drive only ouput of a single Concurrent Request to a specific Location .