Emmanuel, Have you been able to see something strange in our configuration, or any suggestion that we can do ?
Toni. ---------- Forwarded message ---------- From: Toni Menendez Lopez <[EMAIL PROTECTED]> Date: 11-jul-2008 16:21 Subject: Fwd: [Sequoia] Problems inserting in DATABASE ( only DISPLAY is allowed ) To: Sequoia general mailing list <[email protected]>, Emmanuel Cecchet <[EMAIL PROTECTED]> Cc: [EMAIL PROTECTED], Antonio Menendez Lopez < [EMAIL PROTECTED]> Emmanuel , Just to sumarize and, may be, clarify our problem. We have a running environment composed of 3 nodes. Two BE and one FE. Data Distribution corresponds to a Full Repplication model (RAIDb-1). Our applications are running in a JBOSS AS. Untill now we had two different Data Bases in both BE. Access through these DB was performed from a J2E service developed with Spring and Hibernate. These two services doesn't have any problem interacting with DB through Sequoia interface. We have now new services attaching the new DB. All of them accesses DB by using Hibernate. There are J2E services and a GUI developed with JSF. The new J2E services interact DB from a J2E container. We have tried to configure Hibernate.cfg.xml module by using JNDI and JTA transaction factory class. We have been geting exceptions in commit() calls untill we changed the DS definition to a <no-tx-datasource>. But we must use a <local-tx-datasource>, so we need face again the problems regarding transaction handling. All the traces we have send to you are related to problems we have in the GUI module. We have tried with JNDI connection and we don't have any problem while trying to access to our DB when we call SELECT statements. But whenever we call a DELETE, UPDATE or INSERT statement, we alwasys get an Exception which origin seems to be a Sequoia Timeout. The value of the timeout seems to match the values we have declared in our DB DS description file: <blocking-timeout-millis>12000</blocking-timeout-millis> as well as the timer values declared in the virtualdabase description file. For your info, whenever we use a JDBC connection from the J2E services, we get also a Sequoia Timeout exception while executing update DB commands. The amazing thing for us is that whenever play INSERT, UPDATE or DELETE commands from the Sequoia console, these commands are always successfuly exectued. So that, it seems the Sequoia description files are right. Since the services interacting the other two DBs through Sequoia are running, we don't think the problem regards to a communication problem, since (according Sequoia documentantion) the transport configuration is not based on single DB but for all Sequoia processes. So that, this is our situation. We have followed all the Sequoia documentation fro clientes accessing DB through Hibernate in JBOSS environments by using JDBC connections, but we have problems in update DB statements thrown by Sequoia modules. ---------- Forwarded message ---------- From: Toni Menendez Lopez <[EMAIL PROTECTED]> Date: 11-jul-2008 10:01 Subject: Re: [Sequoia] Problems inserting in DATABASE ( only DISPLAY is allowed ) To: Emmanuel Cecchet <[EMAIL PROTECTED]>, [email protected] Cc: [EMAIL PROTECTED], Antonio Menendez Lopez < [EMAIL PROTECTED]> Emmanuel, About UDP, the thing is that we have another 2 databases working OK with this TCP configuration. Anyway, I have added jgroup logs in order to see if something is missed. About the other parameter we have changed and still same problem !!! Look traces now, the transaccion with problems is "delete from iplat.uafamily" at this time : 11:44:25,027 And it looks is replicated in both BackEnds. Do you know if I can do any other test to see where is the problem ? I have trace mysql in both backends and it seems it try to do the delete in both mysqls of both backends. 2008/7/10, Emmanuel Cecchet <[EMAIL PROTECTED]>: > > Toni, > > It looks like there is something with your new JGroups configuration based > on TCP. > Could you try the original configuration on UDP shipped with Sequoia to > confirm that this is just a group communication configuration issue? It > looks like messages are never coming back. > Looking at the logs on the second controller should also help you see if > messages are really delivered or not. > I can't recall if 'loopback=true' has an influence on the Hedera message > buffers but you might also want to try to tweak that parameter. > > Hope this helps, > Emmanuel > > I have modified log4j with high level of traces and here are the new >> traces, I hope you can find something.... >> Toni. >> >> 2008/7/8, Emmanuel Cecchet <[EMAIL PROTECTED] <mailto: >> [EMAIL PROTECTED]>>: >> >> Hi Toni, >> >> The timeouts you see are due to a configuration problem with the >> group communication (SELECTs don't go through the group >> communication). >> Don't hesitate to post your config files if you need help. >> >> Hope this helps, >> Emmanuel >> >> We have a problem in our configuration in sequoia and we >> don´t know how to follow the problem, the thing is the following : >> - From our application we are able to see data in oaur >> DATABASE but not to CREATE ( INSERT ) new data, this are logs >> in sequoia.log. >> We have the following situation: >> - all SELECT transactions are successfully executed from >> client applications interacting DB through Sequoia >> - all INSERT, UPDATE, DELETE statements end with the following >> exception exception >> org.continuent.hedera.adapters.TimeoutException >> If we check Sequoia logs, we can see something like... >> [08-07-08 14:00:31,664] WARN continuent.hedera.adapters >> Timeout while waiting for responses >> [08-07-08 14:00:31,665] WARN controller.RequestManager.iplat >> Request 44 failed on controller >> Member(address=/10.10.10.7:7830 <http://10.10.10.7:7830/> >> <http://10.10.10.7:7830 <http://10.10.10.7:7830/>>, >> id=mibeas_0d0d0d03_sequoia_iplat) >> (org.continuent.hedera.adapters.TimeoutException) >> [08-07-08 14:00:31,665] WARN controller.RequestManager.iplat >> Request 44 failed on controller >> Member(address=mibeas01_data/10.10.10.6:7830 >> <http://10.10.10.6:7830/> <http://10.10.10.6:7830 >> <http://10.10.10.6:7830/>>, uid=mibeas_0d0d0d03_sequoia_iplat) >> (org.continuent.hedera.adapters.TimeoutException) >> [08-07-08 14:00:31,666] INFO controller.RequestManager.iplat >> Unable to find request insert into iplat.uafamily >> (UAFAMILY_NAME) values (?)/<!%S|MiFamilia|!%> in list of >> requests that failed on all backends. >> [08-07-08 14:00:31,731] WARN controller.RequestManager.iplat >> Request insert into iplat.uafamily (UAFAMILY_NAME) values >> (?)/<!%S|MiFamilia|!%> failed (Internal exception >> org.continuent.hedera.adapters.TimeoutException) >> [08-07-08 14:00:31,798] WARN controller.virtualdatabase.iplat >> Request insert into iplat.uafamily (UAFAMILY_NAME) values >> (?)/<!%S|MiFamilia|!%> failed, transaction has been aborted >> (Internal exception >> org.continuent.hedera.adapters.TimeoutException) >> [08-07-08 14:00:31,799] WARN >> virtualdatabase.VirtualDatabaseWorkerThread.iplat Error >> during command execution (Internal exception >> org.continuent.hedera.adapters.TimeoutException) >> [08-07-08 14:00:31,801] WARN >> virtualdatabase.VirtualDatabaseWorkerThread.iplat Transaction >> 29 was aborted by database >> >> -- Emmanuel Cecchet >> FTO @ Frog Thinker Open Source Development & Consulting >> -- >> Web: http://www.frogthinker.org <http://www.frogthinker.org/> >> email: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]> >> Skype: emmanuel_cecchet >> >> _______________________________________________ >> Sequoia mailing list >> [email protected] >> <mailto:[email protected]> >> https://forge.continuent.org/mailman/listinfo/sequoia >> >> >> > > -- > Emmanuel Cecchet > FTO @ Frog Thinker Open Source Development & Consulting > -- > Web: http://www.frogthinker.org > email: [EMAIL PROTECTED] > Skype: emmanuel_cecchet > >
_______________________________________________ Sequoia mailing list [email protected] https://forge.continuent.org/mailman/listinfo/sequoia
