There is another config 'job.cancel.threshold.minutes' (defaulted to 60 mins) based on which async job is forcefully cancelled. Can you check if this is cancelling the job?
Also if you want to debug - checkout the GC code present in AsyncJobManagerImpl.java -Koushik -----Original Message----- From: Marcus [mailto:shadow...@gmail.com] Sent: Friday, 6 February 2015 11:56 To: dev@cloudstack.apache.org Subject: Re: long running async job I believe the job.expire.minutes is the cleanup (the only thing that *should* be removing jobs from the table), that's 1440 min. On Thu, Feb 5, 2015 at 11:19 AM, Daan Hoogland <daan.hoogl...@gmail.com> wrote: > there is a global setting called something like > job.timeout.threshold.* . there also is a cleanup value. The timeout > or cleanu don't kill the backecnd process! > > On Thu, Feb 5, 2015 at 6:53 PM, Marcus <shadow...@gmail.com> wrote: >> Hi, >> I'm wondering if someone can refresh me on the limitations around >> async jobs and how to tune them. How long will they run for and >> which global settings control that? As an aside, in one of my dev >> environments running 4.3 I've found that a long running job will >> simply disappear from the async_jobs table after about an hour. For >> example if I make an async job that just does a shell 'sleep 7200' >> and then poll for the async job the queryAsyncJobResult api call >> works and then eventually throws an error saying that the job is not >> found, around about the hour mark (within 10 minutes, haven't timed exactly). >> When this happens, I can clearly see that the job is still running by >> doing a process list on the server, but the job entry has been purged >> from the mysql table. My job.expire.minutes is 1440. > > > > -- > Daan
