I vaguely remember seeing a thread about time limits on job events, but I couldn’t find it. I wanted to find out what the plan was for this, as there is currently no limit to the amount of time a slave can waste if an event hangs.
I don’t recall this being brought up before. Maybe it was brought up in a thread in conjunction with other things and was overlooked…
There currently is no plan in place for this, but this might be something we could consider in Deadline 8. If the triggering of events is sandboxed from the rest of Deadline, then a timeout could be set, at which point the sandboxed process would get killed if the event is taking too long.
Cheers,
Ryan
OK. In the meantime, since I’m calling my own child process to do all of the actual event work, I’ll just set a reasonable process timeout myself.