vCenter

 View Only

  • 1.  Issue with Hyperic server log

    Posted Oct 27, 2009 07:01 PM
    I'm running hyperic server 4.1.2 and currently out of nowhere the log files have started grow. I'm seeing the log grow as be as 1GB for 1 days worth. This was never the case before. In the log it seems that this event is what's eating up space:

    WARN [Thread-6] [com.arjuna.ats.arjuna.logging.arjLoggerI18N@113] [com.arjuna.ats.arjuna.recovery.RecoverAtomicAction_4] - RecoverAtomicAction: transaction not activated, unable to replay phase 2 commit


    Any thoughts?


  • 2.  RE: Issue with Hyperic server log

    Posted Oct 29, 2009 06:58 AM
    Hi,

    I just did a quick search on google and it came up with the following posting:
    http://www.jboss.org/community/wiki/DiskFullWarnMessage

    If the size of the logfiles is the problem, you might consider to change the loglevel. You should find a jboss-log4j.xml in your HQ server directory. At the bottom of the file the general loglevel is defined, which is INFO by default. I think you can change it to ERROR and after a restart your log output is very, very low. Just in case you need to debug a setting you can change it back to INFO.

    Cheers,
    Mirko


  • 3.  RE: Issue with Hyperic server log

    Posted Dec 06, 2009 05:09 PM
    excowboy nailed it.

    These messages are being generated because when the disk filled up and corrupted transaction states were written to the disk. As the HQ server tries to recover it is spitting out these messages.

    We need to clean up the corrupted states. You should be able to do this by stopping the HQ server, deleting the contents of the below directory, and restarting.

    <hq-install-path>/hq-engine/server/default/data/tx-object-store/HashedActionStore/defaultStore/

    Please allow about 5 minutes after starting the HQ server and if it continues to fail to start, attach an updated server.log.

    http://www.jboss.org/community/wiki/DiskFullWarnMessage