tag:blogger.com,1999:blog-1304066656993695443.post4179046184105573801..comments2023-12-17T11:32:00.020+00:00Comments on Paul Done's Technical Blog: Anatomy of WebLogic TLOGs and considerations for transaction recoveryPaul Donehttp://www.blogger.com/profile/09556312012162376804noreply@blogger.comBlogger18125tag:blogger.com,1999:blog-1304066656993695443.post-63314215488370365372019-03-26T19:17:40.212+00:002019-03-26T19:17:40.212+00:00Hi Paul,
Very useful write up and effort !
Is th...Hi Paul, <br /><br />Very useful write up and effort !<br />Is there a way to query this uncommitted data from the XA transaction using an sql query? <br /><br />Thanks !<br />RRhellohttps://www.blogger.com/profile/00731368730910356232noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-61195827723694626952018-08-21T17:52:23.902+01:002018-08-21T17:52:23.902+01:00Hey Paul,
We have a serious issue in one of our en...Hey Paul,<br />We have a serious issue in one of our environment.<br />The transactions are getting blocked on the TLOG, we tried to check many thing but no luck, can you provide some insight what can be done to resolve it ?<br /><br /><br />Relevant Information Collection <br />--------------------------------------- <br />"[ACTIVE] ExecuteThread: '17' for queue: 'weblogic.kernel.Default (self-tuning)'" #110 daemon prio=5 os_prio=0 tid=0x00007f280c00b800 nid=0xf696 waiting for monitor entry [0x00007f27aa4f4000] <br />java.lang.Thread.State: BLOCKED (on object monitor) <br />at weblogic.transaction.internal.StoreTransactionLoggerImpl.store(StoreTransactionLoggerImpl.java:265) <br />- waiting to lock <0x00000006d388b4c0> (a weblogic.transaction.internal.StoreTransactionLoggerImpl) <br />at weblogic.transaction.internal.ResourceCheckpoint.blockingStore(ResourceCheckpoint.java:437) <br />- locked <0x000000076539aed8> (a weblogic.transaction.internal.ResourceCheckpoint) <br />at weblogic.transaction.internal.ResourceDescriptor.checkpointResourceDiscovery(ResourceDescriptor.java:482) <br />at weblogic.transaction.internal.ResourceDescriptor.checkpointIfNecessary(ResourceDescriptor.java:469) <br />at weblogic.transaction.internal.ServerTransactionImpl.checkpointIfNecessary(ServerTransactionImpl.java:2674) <br />at weblogic.transaction.internal.ServerTransactionImpl.globalPrepare(ServerTransactionImpl.java:2521) <br />at weblogic.transaction.internal.ServerTransactionImpl.internalCommit(ServerTransactionImpl.java:297) <br />at weblogic.transaction.internal.ServerTransactionImpl.commit(ServerTransactionImpl.java:257) <br />at weblogic.ejb.container.internal.BaseRemoteObject.postInvoke1(BaseRemoteObject.java:422) <br />at weblogic.ejb.container.internal.StatelessRemoteObject.postInvoke1(StatelessRemoteObject.java:20) <br />at weblogic.ejb.container.internal.BaseRemoteObject.__WL_postInvokeTxRetry(BaseRemoteObject.java:287) <br />at weblogic.ejb.container.internal.SessionRemoteMethodInvoker.invokeInternal(SessionRemoteMethodInvoker.java:67) <br />at weblogic.ejb.container.internal.SessionRemoteMethodInvoker.invoke(SessionRemoteMethodInvoker.java:21) <br />at com.amdocs.cih.services.customerprofile.ejbs.CustomerProfileServiceBean_sbey4r_EOImpl.createCustomerProfile(Unknown Source) <br />at sun.reflect.GeneratedMethodAccessor2551.invoke(Unknown Source) <br />at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) <br />at java.lang.reflect.Method.invoke(Method.java:498) <br />at com.amdocs.aif.j2ee.AbstractJ2eeConnector$ExecuteMethod.run(AbstractJ2eeConnector.java:116)Anonymoushttps://www.blogger.com/profile/13327148419364939691noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-22473368902512234132015-03-05T21:49:26.799+00:002015-03-05T21:49:26.799+00:00Excellent Article Paul...Thank you for the in dept...Excellent Article Paul...Thank you for the in depth write up..Gayatrihttps://www.blogger.com/profile/14833412952470880863noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-17466403214024449612014-07-30T15:02:47.112+01:002014-07-30T15:02:47.112+01:00Hi Paul,
do you still have files related to this ...Hi Paul,<br /><br />do you still have files related to this test? I've got doubts related to datasource and JNDI name. I want to believe that not datasource but jndi name should be used by TXN recovery... Let me know what was the JNDI name for a data source used in test.<br /><br />Thanks,<br />Ryszardryszard.styczynskihttps://www.blogger.com/profile/13421472021906942132noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-53552996797335236532013-11-15T09:35:58.758+00:002013-11-15T09:35:58.758+00:00Jean-Francois, I sometimes see that ignorance from...Jean-Francois, I sometimes see that ignorance from certain Oracle DBAs (usually the ones that arrogantly claim they know everything about the Oracle DB, which I for one, don't). Invariably, they've never come across client applications that use XA transactions against the DB before, and assume they work like 'normal' db transactions. 'Normal' db transactions do not outlive the connection/session, but XA transactions can and often do. You need to tell your DBA to become more informed by googling "Oracle XA".Paul Donehttps://www.blogger.com/profile/09556312012162376804noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-57449063149635407272013-10-17T10:10:58.956+01:002013-10-17T10:10:58.956+01:00Hello, an Oracle DBA told me that the DataBase tra...Hello, an Oracle DBA told me that the DataBase transaction was linked to the session associated to the JDBC connexion. He also told me that a broken connexion emplie an inflight transaction purged by the DataBase (new JDBC connexion produce a new DataBase Session) (cannot be continued). So there is no way to continue the transaction after a WLS restart (only retry the transaction branch). Any idea on this point ?<br /><br />(Excelent post on that subject ...)Anonymoushttps://www.blogger.com/profile/04310097003998859412noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-11479630768123517742013-07-14T08:48:01.126+01:002013-07-14T08:48:01.126+01:00
ini my cluster when my managed server fail ,i ge...<br /><br />ini my cluster when my managed server fail ,i get warning :<br />Warning: Fail-back retry of Transaction Recovery Service for server [serverName] failed.Anonymoushttps://www.blogger.com/profile/01115954774771569170noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-38511013866696629392013-07-14T08:06:32.096+01:002013-07-14T08:06:32.096+01:00why my weblogic never update tlog/ wls_admin.dat a...why my weblogic never update tlog/ wls_admin.dat about transaction, i use jta transaction but tlog always empty tlog ? thnx you Anonymoushttps://www.blogger.com/profile/01115954774771569170noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-69414257274498151702013-05-23T13:16:46.797+01:002013-05-23T13:16:46.797+01:00incredible useful, thanksincredible useful, thanksfarukhttps://www.blogger.com/profile/11818987475049642901noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-20721003943907951132013-03-07T11:06:17.500+00:002013-03-07T11:06:17.500+00:00Apologies, but for various reasons I'm not abl...Apologies, but for various reasons I'm not able to share the code for BoxBurner, so that is left as an exercise for the reader ;)Paul Donehttps://www.blogger.com/profile/09556312012162376804noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-42030221232651701782013-02-05T06:52:04.579+00:002013-02-05T06:52:04.579+00:00Hi! Thank you so much, very in-depth testing! Woul...Hi! Thank you so much, very in-depth testing! Would you mind sharing the test application with us? NelsonUnknownhttps://www.blogger.com/profile/16385261310163783750noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-90379739832125801732013-01-24T01:02:33.238+00:002013-01-24T01:02:33.238+00:00Paul, excellent job, by far the best article read ...Paul, excellent job, by far the best article read so far on this topic. Keep posting please!vernettohttps://www.blogger.com/profile/00289134594764720847noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-91364308826275282772013-01-09T01:10:08.433+00:002013-01-09T01:10:08.433+00:00Dear Paul,
I met a very slow startup of the weblog...Dear Paul,<br />I met a very slow startup of the weblogic server and by reading the logs, it seems to be related to the TLOG recovery where the persistent store scan its log and data files in order to reconstruct its state at the last time it was running.<br /><br />Nov 28, 2012 3:01:05 PM HKT ..BEA-280008..Opening the persistent file store "xxx" for recovery: directory=/abc/xxx requestedWritePolicy="Direct-Write" fileLockingEnabled=true driver="wlfileio3".<br /><br />Nov 28, 2012 3:16:25 PM HKT ..BEA-280009..The persistent file store "xxx" (956e4aef-e7ef-41ac-bc6b-afcf4afc1383) has been opened: blockSize=512 actualWritePolicy="Direct-Write(read-buffered)" explicitIOEnforced=false records=37. <br /><br />As u can see, it took around 15min. We met this problem twice. The policy is set to Direct-Write and blocksize set to 512, wondering if it would help by changing those settings.<br /><br />Wondering if you have any idea?<br />Thanks a lot.<br />Anonymoushttps://www.blogger.com/profile/16905996069326120012noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-32152407833257932722012-12-09T16:02:15.015+00:002012-12-09T16:02:15.015+00:00An interesting article Paul - thanks for taking th...An interesting article Paul - thanks for taking the trouble to publish your results.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-1063324717700729902012-11-05T10:41:37.095+00:002012-11-05T10:41:37.095+00:00Good question. I definitely would not try to manua...Good question. I definitely would not try to manually commit or roll back anything. Ideally, if the Tlogs and JMS are persisted on shared file-system, then some sort of file-system replication is in place to enable it to be restored in DR site with DR copy of domain that can then be started to push these thru to completion. Likewise if Tlogs and JMS are in DB, use something like Oracle Data Guard to ensure the data is recoverable from DR site.Paul Donehttps://www.blogger.com/profile/09556312012162376804noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-71288999380544362752012-11-05T09:59:27.625+00:002012-11-05T09:59:27.625+00:00Thx Paul for your reply!
For Disaster Recovery, sh...Thx Paul for your reply!<br />For Disaster Recovery, should it be safer to force commit all the pending transactions manually for the case where things going wrong (like using IP instead of domains or anything else) preventing the commit; and then, on another XA resource (JMS for example), all related JMS messages have been committed that implies Lost Messages since the related data in DB are lost?Anonymoushttps://www.blogger.com/profile/16905996069326120012noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-3273146080016138802012-11-01T12:25:47.673+00:002012-11-01T12:25:47.673+00:00Good point and yes the JMS persistent stores have ...Good point and yes the JMS persistent stores have to be considered too, for reliably storing and then restoring after a failure. In the example tests, the WebLogic Transaction Service will also have liaised with the the JMS Server (treated like an XA resource manager, like the database was) to push through their XA pending messages. In hindsight, I should have also showed the output from monitoring the JMS servers for listing their pending messages too. Paul Donehttps://www.blogger.com/profile/09556312012162376804noreply@blogger.comtag:blogger.com,1999:blog-1304066656993695443.post-80472537249346066532012-10-31T09:37:41.898+00:002012-10-31T09:37:41.898+00:00Hi Paul, only pending transactions related to the ...Hi Paul, only pending transactions related to the Database are mentioned. Just wondering what would happen for the JMS Resource and the message that has been sent or read.<br />does JMS have "XA transactions pending commit" similar to Oracle database?Anonymoushttps://www.blogger.com/profile/16905996069326120012noreply@blogger.com