Uppsala Multidisciplinary Center for Advanced Computational Science

Slurm problems on Rackham -- fixed

2017-04-28

Today, Friday, we had a problem with Slurm on Rackham, that made Slurm stop for a while.

This is a known bug in our current Slurm version and there exists a new version that solves that problem.

We are not sure that we are able to upgrade before the weekend, so we might leave Slurm working in a reduced way, running but with bad priorities for waiting jobs.

Perhaps you have to live with bad priorities over the weekend and we are sorry about that.

Update on Tuesday at 1150 hours

We were lucky to be able to restore the priority system, while still running the old version of Slurm.

Tomorrow, during the maintenance window, we plan to upgrade to the new version, without this known bug, so we now somewhat prematurely declare the problem to be fixed.

Old System News