Hi Paul, and thank you for your feedback on Quartz. As regards the current job scheduler, there are a few issues that I often have to cope with in production, when the number of records in the JobSandbox grows over a certain limit (due to high load or to some failing job that is executed over and over); you can end up with db lock errors on the JobSandbox and the implementation of the service that cleans the old job (purgeOldJobs) is not ideal; that service also makes use of the ServiceSemaphore entity in order to guarantee that no more than one job at the time is run and I have some concerns about the reliability of this approach (race conditions, stale data); some interesting threads are:
http://ofbiz.135035.n4.nabble.com/Replace-JobManager-with-Quartz-Scheduler-td2999416.html http://ofbiz.135035.n4.nabble.com/Job-Manager-td4635496.html http://permalink.gmane.org/gmane.comp.java.ofbiz.user/34131 On the other hand, the OFBiz job scheduler is used a lot and so it is "tested" every day in many production instances and, despite of some problems (some of them mentioned above), as you said, it just works. For this reason I agree that we should evaluate this (and similar other proposed in my first email) changes carefully before taking a decision. Kind regards, Jacopo On Mar 22, 2013, at 12:26 AM, Paul Piper <[email protected]> wrote: > I have worked with Quartz before - it is a fantastic solution, so I'd also > second that opinion of implementing Quartz as a replacement for the job > scheduler. The only point that would speak against it, is that I haven't > really had any issues with the current scheduler, so a replacement may not > be necessary and should be discussed whether it is a worthwhile venture. > > > > -- > View this message in context: > http://ofbiz.135035.n4.nabble.com/Some-ideas-for-the-future-of-the-OFBiz-tp4639965p4639968.html > Sent from the OFBiz - Dev mailing list archive at Nabble.com.
