Re: JobManager.reloadCrashedJobs() fails to complete

2009-08-29 Thread Scott Gray
Thanks David, I'll go ahead and make the change. My main doubt was that aside from UI queries the statusId fields doesn't seem to be used for the actual processing of any jobs, reloadCrashedJobs, poll and purgeOldJobs all use the date fields only. The poller is fine because it's querying

Re: JobManager.reloadCrashedJobs() fails to complete

2009-08-29 Thread David E Jones
That would make sense to do. I guess when the statudId field was added the query was never adjusted. On a side note, my guess is that the performance difference is due to indexed versus non-indexed fields, and that statudId is indexed automatically because the reverse-foreign- key indexing

JobManager.reloadCrashedJobs() fails to complete

2009-08-29 Thread Scott Gray
As the JobSandbox table grows, the time taken to query the database for crashed jobs also grows and eventually causes timeouts meaning any crashed jobs are no longer loaded. The query looks like this: SELECT * FROM JOB_SANDBOX WHERE (FINISH_DATE_TIME IS NULL AND CANCEL_DATE_TIME IS NULL AND