There were some changes in salloc operation for Cray systems. More
information about the changes is available here:
https://github.com/SchedMD/slurm/blob/slurm-2.2/NEWS
Could you try SLURM v2.2.7? It definitely has some changes that might
fix this.
Moe Jette
SchedMD LLC
Quoting Alexis GÜNST HORN <[email protected]>:
Hello,
I've just upgraded my Debian Cluster from Stable (Squeeze) to Testing
(Wheezy).
In Squeeze, SLURM is 2.1.11
In Wheezy SLURM is 2.2.6
Before, on Squeeze, everything was working fine.
But now, it doesn't work anymore.
In fact, the "salloc" command does not work any more.
For example, if I do this, on the management node :
$ salloc -w node2
salloc: Pending job allocation 9
salloc: job 9 queued and waiting for resources
salloc: job 9 has been allocated resources
salloc: Granted job allocation 9
salloc: Relinquishing job allocation 9
salloc: Job allocation 9 has been revoked.
So, it doesn't work. Moreover, just after that, a process "bash" is
created on the management node, which take 99% of CPU.
Is it a bug ? A changement in config file to do ?
Please help me !
--
Alexis GÜNST HORN
