Weird behavior with error cleaning

Deadline
1

Hi,

I am having some weird issues with deadline.

Lets say by exemple, i have a job with 7 error. My maximum error allowed before making a task failed is 10.

I do a clean error report on the job, and do some changes to fix the problem (by exemple create a missing folder).

Then trough the job, i got 3 random error that are not fatal to me. Then if i pause the job, and resume it a bit later, the job wont pick up on the farm. I go modified proprieties to put the job on the pool having the highest priority, put the priority to a 100, but still the job wont go on the node. I also tried to restart the node with still no success.

The problem come up also if i tried to put another job dependant of this one (the buggy one).

The only solution i found yet is to resubmit the job, but then i need to redo all of my dependencies.

Note that the bug might not be cause by the error cleaning. It is the only correlation i found yet with the problem tought…

Do you have any idea of what beahavior could cause such a random bug ?

Thanks,

Fred

2

It sounds like the slave(s) the job should be rendering on have been marked as bad for the job. This is configured through the failure detection settings:
thinkboxsoftware.com/deadlin … detection/

Do you have the option enabled to mark a slave as bad? If the slave is bad, you can see this for the job (and clear them) by right-clicking on the job in the Monitor and selecting View Bad Slaves List.

Cheers,

  • Ryan