Hello,
We are running deadline pulse and slave on a linux HPC. From time to time threw out the day we get jobs that go into a stalled state requiring a restart of the service. We are not sure why they are stalling but until things are fixed its become quite time consuming constantly monitoring the GUI and in general dealing with it.
I would like to make a script and run it on the pulse server to watch for jobs that have gone into a stalled state.
is there a method or command i can run which will output job status of all current jobs on in the queue?
I tried ‘deadlinejobmonitor -nogui’ but that didn’t work out…
thanks much,
greg
mono deadlinecommand.exe GetSlaves | awk -F= ‘$1 == “MachineName” {printf($2 " ")} $1 == “SlaveState” {print $2}’
this works… and i think i can work with that. but if there is a better way without using mono, like an API method or similar, I’d love to hear about that.
do we have a python script repo?
take care,
greg
Hey Greg,
Glad you found a solution!
You can’t really avoid using Mono, mainly because Deadline runs on top of Mono (in the same way it runs on top of .NET on Windows). The deadlinecommand shell script that the installer creates is essentially calling “mono deadlinecommand.exe” under the hood.
Our current script API must be accessed in the context of Deadline (ie: deadlinecommand -executescript), but in Deadline 6.x, we will be looking at having a proper python API that can be accessed directly from native python scripts.
Cheers,
thanks Ryan. good news on 6.0 too.
take care,
greg