diff --git a/modules/services/slurm.nix b/modules/services/slurm.nix index 02ec2b17..60bb9d3d 100644 --- a/modules/services/slurm.nix +++ b/modules/services/slurm.nix @@ -134,6 +134,10 @@ inputs: # omit --mpi=pmix MpiDefault=pmix + + # record more info + JobAcctGatherType=jobacct_gather/cgroup + AccountingStorageTRES=gres/gpu ''; extraConfigPaths = let gpus = slurm.node.${inputs.config.nixos.model.hostname}.gpus or null; @@ -144,6 +148,14 @@ inputs: (inputs.localLib.attrsToList gpus)); in [(inputs.pkgs.writeTextDir "gres.conf" "AutoDetect=nvml\n${gpuString}")] ); + extraCgroupConfig = + '' + ConstrainCores=yes + ConstrainDevices=yes + ConstrainRAMSpace=yes + ConstrainSwapSpace=yes + AllowedSwapSpace=20 + ''; }; munge = { enable = true; password = inputs.config.sops.secrets."munge.key".path; }; };