slurm jobs not entering queue about fyrd

[MikeDacre]

Interesting, I will check on Monday morning and see if I can reproduce this.

[MikeDacre]

Hi @jbloom, I am really sorry this took so long, I have been very busy at work and I just didn't have time to work on this at all.

So, I can't reproduce this on any of my machines, so it is likely unique to your machine, I would like to try and figure it out though. Could you try and run the same test again and then use sacct to see whether it entered the queue? It has been given a job number so my guess is that one of three things are happening:

1. There is a lag in the queue being updated on your machine, and so squeue and sacct are not reporting up to date information when you run job.wait()
2. The queue is being cleared very quickly after job completion.
3. squeue/sacct are returning slightly different information on your system and it isn't being parsed correctly.

So, could you try again and let me know if they do make it into the queue at all when this happens?

Thanks!

[MikeDacre]

@jbloom: I tried to add a workaround that should solve possibility 1. by adding more attempts to query the queue before failure. I also made some changes to slurm queue parsing that make it more robust, so there is a good chance that those combined changes should solve this issue for you. Please let me know.

[MikeDacre]

Closing now as I think this issue is fixed, let me know if it isn't