cancel
Showing results for 
Search instead for 
Did you mean: 

Performance & communication issues after applying GRC 5.3 SP09 on Oracle10G

Former Member
0 Kudos

Hello,

After applying the SP9 on GRC 5.3, RAR slowed down badly while doing user risk analysis and it is giving the following log:

Mar 2, 2010 5:34:59 PM com.virsa.cc.xsys.meng.ObjAuthMatcher <init>

FINEST: ObjAuthMatcher constructed: 175ms, #singles=13903, #ranges=4, #super=1

Mar 2, 2010 5:35:00 PM com.virsa.cc.xsys.util.RiskLoader loadAuthMapCache

INFO: Auth Map cache reloaded successfully

Mar 2, 2010 5:35:01 PM com.virsa.cc.dataextractor.bo.DataExtractorSAP getObjPermissions

FINEST: getObjPermissions: elapsed time=2263ms

Mar 2, 2010 6:02:10 PM com.virsa.cc.common.util.ConnectionUtil getConnection

FINEST: Data Source not ready, will wait 30 seconds and try again: ResourceException in method ConnectionFactoryImpl.getConnection(): com.sap.engine.services.connector.exceptions.BaseResourceException: Cannot get connection for 120 seconds. Possible reasons: 1) Connections are cached within SystemThread(can be any server service or any code invoked within SystemThread in the SAP J2EE Engine), 2) The pool size of adapter "SAPSR3DB" is not enough according to the current load of the system or 3) The specified time to wait for connection is not enough according to the pool size and current load of the system. In case 1) the solution is to check for cached connections using the Connector Service list-conns command, in case 2) to increase the size of the pool and in case 3) to increase the time to wait for connection property. In case of application thread, there is an automatic mechanism which detects unclosed connections and unfinished transactions.

The JCOs have the following pool configuration:

Maximum Pool Size 50

Maximum connections 100

Connection Time out 10,000

Maximum Waiting time 60,000

It is also giving this log on the second server node:

Mar 1, 2010 5:48:02 AM com.virsa.cc.common.util.ConnectionUtil getConnection

FINEST: retry: 10ResourceException in method ConnectionFactoryImpl.getConnection(): com.sap.engine.services.dbpool.exceptions.BaseResourceException: SQLException thrown by the physical connection: java.sql.SQLException: Io exception: Got minus one from a read call

com.sap.engine.services.dbpool.exceptions.BaseSQLException: ResourceException in method ConnectionFactoryImpl.getConnection(): com.sap.engine.services.dbpool.exceptions.BaseResourceException: SQLException thrown by the physical connection: java.sql.SQLException: Io exception: Got minus one from a read call

at com.sap.engine.services.dbpool.cci.ConnectionFactoryImpl.getConnection(ConnectionFactoryImpl.java:59)

at com.virsa.cc.common.util.ConnectionUtil.getConnection(ConnectionUtil.java:33)

at com.virsa.cc.xsys.riskanalysis.AnalysisDaemon.getServerCount(AnalysisDaemon.java:296)

Caused by: com.sap.engine.services.dbpool.exceptions.BaseResourceException: SQLException thrown by the physical connection: java.sql.SQLException: Io exception: Got minus one from a read call

Filesystems and DB have plenty of space available.

Your thoughts will be appreciated!

GM

Accepted Solutions (0)

Answers (1)

Answers (1)

Former Member
0 Kudos

Hi,

Did you set Java Connector values as given in section "4.6 Optimize System Connectors" in the following link?

http://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/90aa3190-8386-2b10-c4ba-ced67322e...

Thanks

Himadama

Former Member
0 Kudos

Hi Himadama,

Yes, with the exception of the connection time out and maximum waiting time which have been extended.

GM

Former Member
0 Kudos

Hi,

Try restarting the complete SAP GRC system once.

Thanks

Himadama

Former Member
0 Kudos

The JCOs have the following pool configuration:

Maximum Pool Size 50

Maximum connections 100

Connection Time out 10,000

Maximum Waiting time 60,000

Oracle is 10.0.2.4 opatch version 16

GM

Former Member
0 Kudos

Hi,

After bouncing the system the jobs were triggered as normal but after few seconds, they are getting stuck again generating the same logs:

Mar 2, 2010 6:02:10 PM com.virsa.cc.common.util.ConnectionUtil getConnection

FINEST: Data Source not ready, will wait 30 seconds and try again: ResourceException in method ConnectionFactoryImpl.getConnection(): com.sap.engine.services.connector.exceptions.BaseResourceException: Cannot get connection for 120 seconds. Possible reasons: 1) Connections are cached within SystemThread(can be any server service or any code invoked within SystemThread in the SAP J2EE Engine), 2) The pool size of adapter "SAPSR3DB" is not enough according to the current load of the system or 3) The specified time to wait for connection is not enough according to the pool size and current load of the system. In case 1) the solution is to check for cached connections using the Connector Service list-conns command, in case 2) to increase the size of the pool and in case 3) to increase the time to wait for connection property. In case of application thread, there is an automatic mechanism which detects unclosed connections and unfinished transactions.

Former Member
0 Kudos

Hi,

Mar 2, 2010 6:02:10 PM com.virsa.cc.common.util.ConnectionUtil getConnection

FINEST: Data Source not ready, will wait 30 seconds and try again: ResourceException in method ConnectionFactoryImpl.getConnection(): com.sap.engine.services.connector.exceptions.BaseResourceException: Cannot get connection for 120 seconds

Did you test the connection in RAR and in the GRC j2ee server --> Webdynpro --> Webdynpro console --> Maintain Jco destination?

Thanks

Himadama

Former Member
0 Kudos

Hi,

Yes, JCo connections are tested successfully. The problem comes once the risk analysis triggers the jobs, then you will see that log.

On the other hand, when performing an offline analysis the process hangs until the point where the system is not able to report the outcome, the user ID used on the test has only a few SOD so the offline analysis should be performed on no more than 10 secs.

Former Member
0 Kudos

Hi,

Did you able to solve the issue?

Did you complete user, role and profile sync(incremental) jobs before you start the risk analysis job? did you try risk analysis for any other userid?

Thanks

Himadama

Former Member
0 Kudos

Hi Gustova,

One of our customer has got similar issue and was able to fix it by installing ORACLE Patch 18 and rebooting the system.

Hope this helps.

Best Regards,

Sirish Gullapalli.

Edited by: Sirish Gullapalli on Mar 3, 2010 9:39 PM

Former Member
0 Kudos

Hi Sirish,

Thanks for the advice, we are going to try it out and will let you know the outcome.

GM

Former Member
0 Kudos

Hi Sirish,

We are currently up to 10.2.0.4 patchset 6810189 -which is the latest patch for 10gR2, there is no such patch 18. Any other ideas?

GM

Former Member
0 Kudos

Hi Martinez,

Please open an OSS message on this issue, so that SAP support can look into the issue.

Thanks & Regargards,

Sirish Gullapalli.

Former Member
0 Kudos

Hi Sirish and all,

We put the message and they found the problem. The file CC5.3_MESSAGES.txt had to be reapplied after the upgrade. So didn't matter if it was uploaded during the initial configuration you have to do it again after the upgrade. Still I don't understand how this affected performance or the ability to execute jobs at the point that we had to bounce the system and the logs gave nothing else but messages on pool size, connation time out or even DB issues. Maybe the process was looking for the message in the DB and the connection was hanging until getting it rather performing an exit call... I don't know, just a guess...

GM.

former_member771067
Active Participant
0 Kudos

Hi GM,

I am facing similar problems with AC 5.3 Support pack 10.2. To perform risk analysis for any user it is taking 6-10 minutes time. Was this the issue in your case (performance issue). or it was not able to perform the risk anlysis and finally shown any error. It would be really helpful for me to analyze my issue.

Thanks in advance.

Regards,

Gurugobinda

Former Member
0 Kudos

Hi Gurugobinda,

The problem we faced was during the execution of the jobs never been able to finish. For yours it seems you can get better performance by re-indexing the DB, see note 1044174 Recommendation for CC 5.x running on Oracle 10G Database.

GM