Failing a Job in Different Ways

Hey guys,

I’ve been wanting to implement a new way of failing tasks from our current method, which is to use the FailRender call. The behaviour we have set up is to error a task 6 times before it fails. This works well most of the time, however if we have a job that runs overnight that is long running and times out 6 times on every frame, we’re essentially losing upwards of 10 hours per task render time which could be better spent elsewhere.

I don’t want to get rid of the timeout function, and I don’t want to turn down/off the task failure either. What I want is a way to call an option that fails a task other than the FailRender call, so I can keep it as a separate function that immediately kills the task rather than letting it re-run. I saw something about a FailJobTask when I was digging around, but I couldn’t work out where that was getting called from, or the functionality.

If anyone could shed some light on a nice way to go about this, that would be really appreciated.

Thanks!
Courtenay

Hello Courtenay,

I am guessing by some of the stuff you are saying that you have a fair bit of custom written stuff going on in your farm, am I right? We have task and job based timeouts built into deadline as of at least 6.1, so those should work for you if you choose to. If those won’t work, I can consult with our coders to see if I can get more information on this.

Instead of FailRender, you can use the AbortRender function, which takes an AbortLevel value as the second parameter. Just search for the AbortRender function in the Deadline Scripting Reference:
thinkboxsoftware.com/deadlin … mentation/

If you use AbortLevel.Fatal, the task will be set to Failed so that it won’t re-render again.

Cheers,
Ryan