We observed that under the heavy load, the SendTransactedMsg scheduler job hangs waiting for the MQ connection, this waiting has in turn consumed all the free WCS DB2 connections in the pool. This is because the job's DB2 connection is still open while it's waiting for the MQ connection.

In the log file, we get error messages like this:

[1/20/11 3:56:30:049 MST] 00000ed9 CommerceSrvr E com.ibm.commerce.messaging.outboundservice.SendTransactedMsgCmdImpl assembleMsg() CMN8216E: An error com.ibm.ejs.persitence.EnumeratorException
original exception:
COM.ibm.db2.jdbc.DB2Exception: [IBM][CLI Driver][DB2/AIX64] SQL0911N The current transaction has been rolled back because of a deadlock or timeout. Reason code "68". SQLSTATE=40001 occurred while sending the message.


So we increased the DB2 connection pool size and the MQ pool size. It had no effect, we still got this error sometimes under the heavy load.

After some investigation and consulting with IBM WCS support team. We finally found out why.

When we use DB2 with Websphere Commerce Developer, we sometimes have to face the situation that DB2 password is expired or has to be changed. Because DB2 uses native operating system user accounts, and these accounts have to obey company security regulations.

When this happens, Websphere Commerce Developer breaks down because all of its EJB enterprise beans still reference the old password.

Here's how I fix it.

I'm getting this error after installing DB2 v8. Here's the solution I found that works.

Open a DB2 command window.
1. db2set DB2_GRP_LOOKUP=LOCAL,TOKENLOCAL
2. db2stop
3. db2start