On 7/1/19 6:11 am, Marcus Wagner wrote:
But that means, the docker container runs outside the cgroup of the
slurm job. Thus there exists no restriction to the container, so it can
use all resources!
[...]
If this is the case, in my opinion docker cannot be used on shared
systems but only
On 7/1/19 4:44 am, Burian, John wrote:
We see the same behavior using pam_slurm_adopt.
My reading of what Tom was saying was that he could SSH into compute
nodes *before* running a job there, and that wouldn't be possible with
pam_slurm_adopt. But yes, you would expect SSH sessions to be
But that means, the docker container runs outside the cgroup of the
slurm job. Thus there exists no restriction to the container, so it can
use all resources!
If e.g. one badly configured job requests one GPU, but uses all four,
the following jobs on the node will all crash, because they canno
On 5/1/19 12:17 am, Tom Smith wrote:
> Novice question: When I use srun, it closes my SSH sessions to compute
> nodes.
>
> Is this intended behaviour by design? If so, I may need need to know
> more about how slurm is intended to be used. If unexpected, how do I
> s