LockAcquisitionException while adding content in repository.

rupeshsawaliya1 — Fri, 28 Aug 2009 09:00:41 GMT

Hi,I am facing one issue while running test script through Jmeter, issue is related to dead lock in Alfresco.Following is the environment I am havingerver:Websphere v7.0DatabaseB2Alfresco Version:2.1.1Testing tool: JmeterI have developed one test script, which tests Alfresco’s Add Content Functi

Re: LockAcquisitionException while adding content in repository.

dward — Fri, 28 Aug 2009 11:03:40 GMT

Your error message is documented here

http://publib.boulder.ibm.com/infocenter/db2luw/v8/topic/com.ibm.db2.udb.doc/core/rsql0900.htm

SQL0911N The current transaction has been rolled back because of a deadlock or timeout. Reason code 2

Reason code 2 is

transaction rolled back due to deadlock.

This is a perfectly normal occurrence when there are multiple competing transactions. In a transaction collision, one transaction will win and the others will have to be retried. Normally Alfresco's RetryingTransactionHelper will handle such LockAcquisitionExceptions by retrying the transaction after a random 'back off' interval, up to the number of times configured in

server.transaction.max-retries

Alfresco will only give up and log the exception if this number of retries is exceeded. The default is 40.

That is why, in the default Alfresco log4j configuration we have the following setting, which stops Hibernate from logging exceptions which are handled and retried by Alfresco.

log4j.logger.org.hibernate.util.JDBCExceptionReporter=fatal

Now the first thing to check is that your log configuration isn't allowing org.hibernate.util.JDBCExceptionReporter to report INFO, WARN or ERROR messages. I am not sure how this is controlled in Websphere, and I have a feeling that the default Websphere settings might be letting these through. So it may well be that you can ignore these messages if you can confirm they are coming from JDBCExceptionReporter.

If the transactions are actually getting retried 40 times and still failing then this suggests that the server is seriously overloaded. I would be surprised if this is the case.

You could experiment with different values for these parameters to see how it affects the frequency of these errors

server.transaction.max-retries
server.transaction.min-retry-wait-ms
server.transaction.max-retry-wait-ms
server.transaction.wait-increment-ms

Regards

Dave

topic Re: LockAcquisitionException while adding content in repository. in Alfresco Archive

LockAcquisitionException while adding content in repository.

Re: LockAcquisitionException while adding content in repository.