Disable IPMI watchdog in fox

Fixes: #231

m/common/base/users.nix

@@ -108,7 +108,7 @@
         home = "/home/Computational/dbautist";
         description = "Dylan Bautista Cases";
         group = "Computational";
-        hosts = [ "apex" "hut" "tent" "raccoon" ];
+        hosts = [ "apex" "hut" "tent" "raccoon" "owl1" "owl2" ];
         hashedPassword = "$6$a2lpzMRVkG9nSgIm$12G6.ka0sFX1YimqJkBAjbvhRKZ.Hl090B27pdbnQOW0wzyxVWySWhyDDCILjQELky.HKYl9gqOeVXW49nW7q/";
         openssh.authorizedKeys.keys = [
           "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIAb+EQBoS98zrCwnGKkHKwMLdYABMTqv7q9E0+T0QmkS dbautist@bsc-848818791"

m/fox/configuration.nix

@@ -25,6 +25,9 @@
   boot.initrd.availableKernelModules = [ "xhci_pci" "ahci" "nvme" "usbhid" "usb_storage" "sd_mod" ];
   boot.kernelModules = [ "kvm-amd" "amd_uncore" "amd_hsmp" ];
 
+  # Disable IPMI watchdog as the BMC is not stable
+  boot.blacklistedKernelModules = [ "ipmi_watchdog" ];
+
   hardware.cpu.amd.updateMicrocode = lib.mkDefault config.hardware.enableRedistributableFirmware;
   hardware.cpu.intel.updateMicrocode = lib.mkForce false;
 

m/module/slurm-common.nix

@@ -5,8 +5,8 @@
     controlMachine = "apex";
     clusterName = "jungle";
     nodeName = [
-      "owl[1,2]  Sockets=2 CoresPerSocket=14 ThreadsPerCore=2 Feature=owl"
-      "fox       Sockets=8 CoresPerSocket=24 ThreadsPerCore=1"
+      "owl[1,2]  Sockets=2 CoresPerSocket=14 ThreadsPerCore=2 MemSpecLimit=4096 RealMemory=128797"
+      "fox       Sockets=8 CoresPerSocket=24 ThreadsPerCore=1 MemSpecLimit=4096 RealMemory=773659"
     ];
 
     partitionName = [
@@ -41,9 +41,9 @@
       # multiple hardware threads (or CPUs).
       SelectType=select/cons_tres
 
-      # Ignore memory constraints and only use unused cores to share a node with
-      # other jobs.
-      SelectTypeParameters=CR_Core
+      # Both cores and memory are consumable resources, so we can put a limit in
+      # memory as well.
+      SelectTypeParameters=CR_Core_Memory
 
       # Required for pam_slurm_adopt, see https://slurm.schedmd.com/pam_slurm_adopt.html
       # This sets up the "extern" step into which ssh-launched processes will be
@@ -58,7 +58,9 @@
 
     extraCgroupConfig = ''
       CgroupPlugin=cgroup/v2
-      #ConstrainCores=yes
+      ConstrainRAMSpace=yes
+      ConstrainSwapSpace=yes
+      AllowedRAMSpace=99
     '';
   };