I can’t imagine I’m the first one to ever bring this up, but would it be possible to change the job suspend behavior so that it did not kill running tasks by default? I realize there are probably some people who prefer the current behavior, so maybe there’s a way to make the “kill” behavior optional without requiring a popup every time you want to suspend a job.
Combined with the fact that individual tasks cannot currently be suspended or otherwise “skipped” without marking them as done or failed, the current behavior can be fairly inconvenient with a large-ish farm and long-running jobs when highly strategic wrangling is required.
A hacky workaround I’ve used is to change the pool to something like “none” if you have slaves configured to ignore the none pool. They’ll finish their task and then no new machines will pick up the job. You can also just set the whitelist to a blank list. Both hacky and not very good since they don’t give any indication why they deliberately aren’t rendering. I too would like this feature.
This feature is coming in Deadline 7, so you’ll be able to just suspend the non-rendering tasks. Might be handy to have an option when suspending jobs to only suspend the non-rendering tasks as well.
That’s great to hear. I would still vote for adding a way to suspend a job “nicely” though, as it’s much faster than having to select all of the queued tasks and suspending them.
Yup, that’s what I was referring to when I mentioned making it an option when suspending jobs. It could ask if you want to suspend all tasks or all non-rending tasks. It’s been added to the todo list.
