The Doctor
The Doctor
I don't think so, no. This would best be done with multiple Commander Agents and conditional actions.
The output page says "check var/log/prod.log for additional information." Have you done so? What's in there?
There are other tickets about this with a number of fixes: * Make sure the job scheduler hasn't crashed. * Comment out the TIMEZONE variable in .env and restart everything....
Shell into the container and do `ps ax | grep schedule`. If you see an instance of schedule.rb then the job scheduler is up and running. Then do a `ps...
Okay. To kill the stuck Website Agent you'll have to go to the Job Management page and hit the "delete" button for that task. What resources to do you have...
@convexshiba The multi-process Docker setup does use the delayed_job runners, yes. Basically, the threaded job runner can deadlock itself because all of it runs in a single instance of Ruby...
@jtullett > Nothing in ps. But huginn is running happily right now with nothing in the queue - I'll wait for it to get stuck then try again. That might...
@jtullett The lack of scheduler.rb in the process table is what concerned me. If you're running threaded.rb, however, maybe it carries out the same task (vis a vis, updating the...
@jtullett The thing is, it's a threading deadlock. Too many agents running are causing the running Ruby interpreter to lock up. A few of us have tried to find it...
@jtullett That is a similar issue, and I think with the same root cause. As far as I know, RSS Agents can't. Then again, I haven't tried. The only thing...