Been seeing slaves on our farm not respecting the task timeouts of a job.
I’ve attached screenshots of an example, it’s been “rendering” for over a day, but the task timeout on the job is set to 6 hours. In the live slave log, it looks like it tried to cancel the current task, then just got hung up.
This is using Deadline 8.0.0.69
Looks like the cancel did get issued, but the Sandbox must have been in some kind of hung/unresponsive state. So it seems like the root cause of this is likely related to the other sandbox issues you guys are experiencing, we’re looking into that currently.
But either way, we could probably be more aggressive about killing the sandbox process when the time out is detected, if the process isn’t responding at all.
