[slurm-users] Simple free for all cluster

2020-10-02 Thread John H
Hi All Hope you are all keeping well in these difficult times. I have setup a small Slurm cluster of 8 compute nodes (4 x 1-core CPUs, 16GB RAM) without scheduling or accounting as it isn't really needed. I'm just looking for confirmation it's configured correctly to allow the controller to 's

[slurm-users] scancel the solver, not MPI

2020-10-02 Thread Gerhard Strangar
Hi, I'm wondering if it's possible to gracefully terminate a solver that is using MPI. If srun starts the MPI for me, can it tell the solver to terminate and then wait n seconds before it tells MPI to terminate? Or is the only way of handling this using scancel -b and trapping the signal?

Re: [slurm-users] Auto-select partition?

2020-10-02 Thread Renfro, Michael
Depends on the version of Slurm. The docs for 17.11 [1] shows using packjob, and the docs for the current version (20.02 as of this writing) [2] shows using hetjob. It's really easy to wind up on documentation later than your running version, since only the top-level documentation page [3] menti

Re: [slurm-users] Simple free for all cluster

2020-10-02 Thread Renfro, Michael
Depending on the users who will be on this cluster, I'd probably adjust the partition to have a defined, non-infinite MaxTime, and maybe a lower DefaultTime. Otherwise, it would be very easy for someone to start a job that reserves all cores until the nodes get rebooted, since all they have to d

Re: [slurm-users] Auto-select partition?

2020-10-02 Thread Diego Zuccato
Il 02/10/20 10:47, Diego Zuccato ha scritto: I found what was wrong. The line > #SBATCH hetjob should have been #SBATCH packjob Is it an error in the docs (strange, since sbatch accepts both but appears to ignore hetjob) or they're two distinct modes? -- Diego Zuccato DIFA - Dip. di Fisica e

Re: [slurm-users] Auto-select partition?

2020-10-02 Thread Diego Zuccato
Il 02/10/20 09:06, Diego Zuccato ha scritto: > But IIUC, even if there's no default partition and the user did not > select one explicitly, slurm can automatically select one containing all > the requested resources, right? I'm also experimenting with heterogeneus jobs https://slurm.schedmd.com/he

Re: [slurm-users] Auto-select partition?

2020-10-02 Thread Diego Zuccato
Il 01/10/20 16:01, Relu Patrascu ha scritto: > > Besides having a separate partition for each type of node, you can also > have a partition which includes all the nodes, and use the Default=yes > option in its definition.This is how it's currently configured, but being > composed of heterogeneous