Thanks for the pointers. I'll take a look. There is one more piece of information: The purgeOldJobs service is in a "crashed" status. Do you think that is significant?
Thanks, On Wed, Jul 13, 2011 at 4:32 PM, BJ Freeman <bjf...@free-man.net> wrote: > You now know why I don't recommend cloud configuration for realtime > operations, unless your running over dedicate lines not part of the > internet. > to summarize you environment caused the problem not ofbiz > Now you have jobs cued that should have been run but have piled up. > you need a way to get the job run so they don;t time out the system. > I recommend you look at the purge old jobs service, copy and modify it > to run your jobs, maybe by time group. > > Josh Jacobson sent the following on 7/13/2011 12:48 PM: >> Currently I am running: >> >> Red Hat Enterprise Linux Server release 5.5 >> 6 CPUs, 16384MB RAM >> >> It was very recently upgraded from 2 CPUs and 8GB of RAM because we >> were having performance issues (lots of swap memory being used). It's >> on one of those cloud servers. Now it's running without using any >> swap. >> >> On Wed, Jul 13, 2011 at 12:22 PM, BJ Freeman <bjf...@free-man.net> wrote: >>> Ok so you have the latest code. >>> what is the eviorment you working with. >>> OS >>> Memory >>> CPU speed >>> >>> Josh Jacobson sent the following on 7/13/2011 12:12 PM: >>>> BJ, >>>> >>>> I am running 10.04. >>>> >>>> On Wed, Jul 13, 2011 at 12:00 PM, BJ Freeman <bjf...@free-man.net> wrote: >>>>> the key is Transaction timeout >>>>> this could be the job length >>>>> could be the database connection >>>>> >>>>> please specify the version of ofbiz since earlier transaction problems >>>>> were taken care of by changing code that deals with transactions. >>>>> >>>>> Josh Jacobson sent the following on 7/13/2011 11:48 AM: >>>>>> Hello Everyone, >>>>>> >>>>>> I have an ofbiz instance in production where none of the jobs are >>>>>> being performed. I have about 160K jobs in pending status, but they >>>>>> are never being schedule. >>>>>> I can see the following in the log: >>>>>> >>>>>> 2011-07-13 13:32:01,959 (org.ofbiz.service.job.JobPoller@2599930b) [ >>>>>> JobManager.java:201:ERROR] ---- exception report >>>>>> ---------------------------------------------------------- Transaction >>>>>> error trying to commit when polling and updating the JobSandbox: >>>>>> org.ofbiz.entity.transaction.GenericTransactionException: Roll back >>>>>> error (with no rollbackOnly cause found), could not commit >>>>>> transaction, was rolled back instead: >>>>>> javax.transaction.RollbackException: Transaction timeout (Transaction >>>>>> timeout) Exception: >>>>>> org.ofbiz.entity.transaction.GenericTransactionException Message: Roll >>>>>> back error (with no rollbackOnly cause found), could not commit >>>>>> transaction, was rolled back instead: >>>>>> javax.transaction.RollbackException: Transaction timeout (Transaction >>>>>> timeout) ---- cause >>>>>> --------------------------------------------------------------------- >>>>>> Exception: javax.transaction.RollbackException Message: Transaction >>>>>> timeout ---- stack trace >>>>>> --------------------------------------------------------------- >>>>>> javax.transaction.RollbackException: Transaction timeout >>>>>> org.apache.geronimo.transaction.manager.TransactionImpl.commit(TransactionImpl.java:269) >>>>>> org.apache.geronimo.transaction.manager.TransactionManagerImpl.commit(TransactionManagerImpl.java:245) >>>>>> org.ofbiz.entity.transaction.TransactionUtil.commit(TransactionUtil.java:259) >>>>>> org.ofbiz.entity.transaction.TransactionUtil.commit(TransactionUtil.java:245) >>>>>> org.ofbiz.service.job.JobManager.poll(JobManager.java:197) >>>>>> org.ofbiz.service.job.JobPoller.run(JobPoller.java:90) >>>>>> java.lang.Thread.run(Thread.java:619) >>>>>> -------------------------------------------------------------------------------- >>>>>> >>>>>> I believe that the JobManager is not being able to handle all those >>>>>> jobs to schedule them, so nothing is being scheduled, which of course >>>>>> make the job list longer. >>>>>> >>>>>> Can anyone think of how to make the jobs run? >>>>>> >>>>>> All help much appreciated, >>>>>> >>>>> >>>> >>> >> >