Atomikos Forum

Error during checkpoint

Over the last 24 hours we have periodically received the following exception in our logs:

2016-03-23 10:11:39,771 [pool-4587-thread-1] ERROR: exception thrown by invokeLater() action com.clj2.ui.app.callmonitor.ActiveCallTakerTable$4$1@62a13bd9
org.springframework.transaction.TransactionSystemException: JTA failure on commit; nested exception is com.atomikos.icatch.jta.ExtendedSystemException: Error in commit: could not flush state image Error during checkpointing com.atomikos.persistence.LogException
        at org.springframework.transaction.jta.JtaTransactionManager.doCommit(JtaTransactionManager.java:1037)
        at org.springframework.transaction.support.AbstractPlatformTransactionManager.processCommit(AbstractPlatformTransactionManager.java:757)
        at org.springframework.transaction.support.AbstractPlatformTransactionManager.commit(AbstractPlatformTransactionManager.java:726)
        at org.springframework.transaction.interceptor.TransactionAspectSupport.commitTransactionAfterReturning(TransactionAspectSupport.java:478)
        at org.springframework.transaction.interceptor.TransactionAspectSupport.invokeWithinTransaction(TransactionAspectSupport.java:272)
        at org.springframework.transaction.aspectj.AbstractTransactionAspect.ajc$around$org_springframework_transaction_aspectj_AbstractTransactionAspect$1$2a73e96c(AbstractTransactionAspect.aj:63)

This just started happening at run-time with no known cause or explanation.  I have checked the disk space and all is well there.  I read an older post about possible startup ordering issues but this did not occur on startup; in fact, the app has been running for about a month before this error started yesterday.  Also, this does not happen on every commit just on some.  The TX log continues to grow and currently stands at around 60MB.  Is there possibly some limit on TX log size getting breached?  If so, how/why would this occur? Any ideas anyone?
Mark Thomas Send private email
Wednesday, March 23, 2016
 
 
Looks like the root cause of this was a ConcurrentModificationException in XAResourceTransaction.writeData(DataOutput out) while iterating over the vector of HeuristicMessage (line 969). I've manually patched this bug.
Mark Thomas Send private email
Friday, April 08, 2016
 
 
Thanks for sharing!

Cheers,
Guy Pardon Send private email
Friday, April 22, 2016
 
 
Note: in the new 4.0 release we have a new logging mechanism that should be much more performant.
Guy Pardon Send private email
Friday, April 22, 2016
 
 

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics