Good morning,
I’ve been having trouble all weekend to run interactive jobs on Yggdrasil, they stay in pending status even when nodes are available, and they don’t appear in the queue list. Also, the sacct and spart commands are not working, I don’t know if it is related, but I wanted to mention it.
Hope you can help me with this,
Cristina
We had a connection issue between Baobab and Yggdrasil a couple of days ago. This prevent commands such as sacct to work. This is fixed now. But the issue you had with pending job even with available nodes is not normal of course. If Slurm is quite busy this may takes some time for it to allocate a resource.
Can you please let us know the full command you used to request an interactive job and how long did you wait?
Hi @Yann.Sagon ,
Yes, I don’t know if you saw my last email, but if not, I attach here the screenshots again.
I had two screens open, in one I was submitting the interactive job, on the second I was tracking my queue. What I noticed is that the job seems to run but die after 10s, without really giving me permission to do anything. The screen submitting the job stays forever in the “Pending” state, it never changes. I have tried with different partitions, including our private nodes, and it is always the same.
Regards,
Cristina
