Hi, I have some jobs type that are automated scene creation instead of render jobs. For these type of job if the job fail due to a bug in the code or some other reason, the job will always fail
I would like to be able to override the error handling for these job so that the job will become failed after only 1 error. I don;t want to change my repository settings (I currently have 30 task fails before the job turns failed)
Is there a setting (hidden maybe) that I could set at submission that would do this?
Unfortunately at the moment there isn’t a way to override the Failure Detection on a per-job basis. It definitely seems like a useful idea though, so we’ll add it to the wish list.
While you are adding to the wish list, it would be interesting to be able to have different error types with different settings.
Example: All the errors related to the plugin not finding the render executable would be of type A, all other errors of type B
In the settings I could say that if a slave fails with error type A, add the slave to the list (I know that this error will appear every time this slave try to render this task)
But the slave fail with error type B, it can retry rendering it, the error may be caused by network issue or any other issue that may or may not re-appear.
So in short, separating the errors that I know will append every time from errors that are not. Too often a slave with a wrong setup will fail complaining about the render executable not being found and this slave will fail on a bunch of jobs…
We would have to add different FailRender functions to the script API that force different behavior (ie: regular error, error and blacklist slave, etc). I’ll add it to the wish list.