Menu
▾
▴
Re: [Sequoiadb-discuss] [Sequoia] Bug ?
|
From: Emmanuel C. <ma...@fr...> - 2009-05-20 17:49:34
|
Hi Pierre, It might well be that the problem comes from an infinite retransmission from JGroups. You should just get a newer version of JGroups and replace the jar file in the lib directory. I don't remember which version shipped with Sequoia 2.10.9 but it's probably quite old. JGroups 2.6.10.GA should be compatible with Sequoia 2.10.*. Thanks for the feedback, Emmanuel > This is not reproduced every time, it's the first time I explicitly observed this behavior. > > I can see in this log that StackOverflowError is thrown for each query from user1 (vdb user=oracle user), queries for user2 work perfectly. > Time between log "Load balancing read request" and "Runtime error during command execution (null)" is always between 1s and 2s. > > 5 minutes after first error, we can see strange behavior : > - c1 see c2 quitting group (29 Apr 2009 13:00:22,684 | sequoia.controller.recoverylog | Storing checkpoint Member(address=/172.20.1.81:65201, uid=db) quit group db-172.20.1.77:65209-20090429130022678+0000 at request id 1679), but c2 is still alive and does not see other quit ... This cause c1 is in a 1 member group and c2 in a 2 members group !! > > After a restart of the 2 controllers, all seems to be fine. > > Is it possible that StackOverflow is thrown by jGroups (we see it even when a query is select on local backend) ? > After bug on c1 see c2 quit, how can it be possible to have an update on c2 ? > > I join a log which shows that c2 send update request to all members, c1 does not receive anything, but at the end the request seems to be executed (I can not check database) > > > I hope at least 1 symptom could help to find problem source ... > > > > > Pierre Besson-Deblon > > -----Message d'origine----- > De : seq...@li... > [mailto:seq...@li...]De la part de > Emmanuel Cecchet > Envoyé : 18 May 2009 22:39 > À : Sequoia general mailing list; > seq...@li... > Objet : Re: [Sequoia] Bug ? > > > Hi Pierre, > > Note that the mailing list has now moved to > seq...@li... > The java.lang.StackOverflowError error probably comes from an infinite > retry loop somewhere (hard to tell where since the exception is null). > Could you reproduce the issue with Sequoia 2.10.10? I don't have the > source code of 2.10.9 readily available and a number of fixes (as well > as a JGroups update) have been performed in 2.10.10. > Does the problem happen systematically? > > Thanks for your feedback, > Emmanuel > > >> continue ... I made an edit at the end of previous message >> >> >> -----Message d'origine----- >> De : seq...@li... >> [mailto:seq...@li...]De la part de >> BESSON-DEBLON, Pierre (SOGETI HIGH TECH) >> Envoyé : 18 May 2009 19:32 >> À : Sequoia general mailing list >> Objet : [Sequoia] Bug ? >> >> >> Hello, >> >> We use sequoia with 2 controllers on RAIDb1, sequoia is 2.10.9 with jgroups. >> >> We experienced this weird behavior : >> >> # on controller 1 (c1) backend deactivated, on c2 backend activated >> c2.log -> 29 Apr 2009 12:56:12,400 | sequoia.controller.RequestManager.db | Database backend db2 is now enabled >> >> # c1 receive a read request and forward to c2 >> c1.log -> >> 29 Apr 2009 12:56:16,271 | sequoia.controller.virtualdatabase.request.db | S 844424930131983 0 SELECT ... >> 29 Apr 2009 12:56:16,273 | sequoia.controller.RequestManager.db |Sending request SELECT ... to Member c2 >> >> c2.log -> >> 29 Apr 2009 12:56:16,353 | sequoia.controller.virtualdatabase.db | S 844424930131983 0 SELECT ... >> 29 Apr 2009 12:56:16,355 | sequoia.controller.virtualdatabase.db | 172.20.1.81:65209: Scheduling distributedRequest 844424930131983 from Member c1 >> 29 Apr 2009 12:56:16,355 | sequoia.controller.distributedvirtualdatabase.request.db | S 844424930131983 0 SELECT ... >> 29 Apr 2009 12:56:16,356 | sequoia.controller.virtualdatabase.db | handleMessageMultiThreaded (class org.continuent.sequoia.controller.virtualdatabase.protocol.ExecRemoteStatementExecuteQuery): S 844424930131983 0 SELECT ... >> 29 Apr 2009 12:56:16,356 | sequoia.controller.virtualdatabase.db | 172.20.1.81:65209: Executing distributedRequest 844424930131983 from Member c1 >> 29 Apr 2009 12:56:16,356 | sequoia.controller.RequestManager.db | Database schema is dirty, refreshing it >> 29 Apr 2009 12:56:16,359 | sequoia.controller.RequestManager.db | Scheduling read Request (844424930131983): SELECT ... >> 29 Apr 2009 12:56:16,359 | sequoia.controller.RequestManager.db | Load balancing read request (844424930131983): SELECT ... >> >> and right after in c2.log file >> 29 Apr 2009 12:56:21,268 | sequoia.controller.virtualdatabase.VirtualDatabaseWorkerThread.db | StatementExecuteQuery command >> 29 Apr 2009 12:56:21,268 | sequoia.controller.virtualdatabase.request.db | S 562949953430692 0 SELECT ... >> >> c1.log >> 29 Apr 2009 12:56:21,233 | sequoia.controller.RequestManager.db | Request SELECT ... completed. >> 29 Apr 2009 12:56:21,234 | sequoia.controller.RequestManager.db | An error occured while executing remote select request 844424930131983 >> org.continuent.sequoia.common.exceptions.NoMoreBackendException >> at org.continuent.sequoia.controller.requestmanager.distributed.RAIDb1DistributedRequestManager.execRemoteStatementExecuteQuery(Unknown Source) >> at org.continuent.sequoia.controller.requestmanager.distributed.DistributedRequestManager.statementExecuteQuery(Unknown Source) >> at org.continuent.sequoia.controller.virtualdatabase.VirtualDatabase.statementExecuteQuery(Unknown Source) >> at org.continuent.sequoia.controller.virtualdatabase.VirtualDatabaseWorkerThread.statementExecuteQuery(Unknown Source) >> at org.continuent.sequoia.controller.virtualdatabase.VirtualDatabaseWorkerThread.run(Unknown Source) >> 29 Apr 2009 12:56:21,234 | sequoia.controller.virtualdatabase.db | Request '844424930131983' failed (null) >> 29 Apr 2009 12:56:21,234 | sequoia.controller.virtualdatabase.VirtualDatabaseWorkerThread.db | Error during command execution (null) >> >> >> So, How can request fail with NoMoreBackendException despite db2 on c2 is enabled ? >> Is it normal to have no more log in c2.log to indicate Exception ?, is it possible that request is hanging on c2 and a timeout raise the exception ? >> >> edit >> found this in c2.log >> 29 Apr 2009 12:56:24,553 | sequoia.controller.virtualdatabase.VirtualDatabaseWorkerThread.db | Runtime error during command execution (null) >> java.lang.StackOverflowError >> 29 Apr 2009 12:56:24,554 | sequoia.controller.virtualdatabase.VirtualDatabaseWorkerThread.db | Reset command >> I think it could help >> >> >> Thanks for any help. >> >> >> Pierre Besson-Deblon >> >> >> The information in this e-mail is confidential. The contents may not be disclosed or used by anyone other then the addressee. Access to this e-mail by anyone else is unauthorised. >> If you are not the intended recipient, please notify Airbus immediately and delete this e-mail. >> Airbus cannot accept any responsibility for the accuracy or completeness of this e-mail as it has been sent over public networks. If you have any concerns over the content of this message or its Accuracy or Integrity, please contact Airbus immediately. >> All outgoing e-mails from Airbus are checked using regularly updated virus scanning software but you should take whatever measures you deem to be appropriate to ensure that this message and any attachments are virus free. >> >> _______________________________________________ >> Sequoia mailing list >> Se...@li... >> http://forge.continuent.org/mailman/listinfo/sequoia >> >> >> > > > > ------------------------------------------------------------------------ > > ------------------------------------------------------------------------------ > Crystal Reports - New Free Runtime and 30 Day Trial > Check out the new simplified licensing option that enables > unlimited royalty-free distribution of the report engine > for externally facing server and web deployment. > http://p.sf.net/sfu/businessobjects > ------------------------------------------------------------------------ > > _______________________________________________ > Sequoiadb-discuss mailing list > Seq...@li... > https://lists.sourceforge.net/lists/listinfo/sequoiadb-discuss > -- Emmanuel Cecchet FTO @ Frog Thinker Open Source Development & Consulting -- Web: http://www.frogthinker.org email: ma...@fr... Skype: emmanuel_cecchet |
