Multipass Single-node Slurm Cluster

The slurm-multipass deployment application creates a single-node Slurm environment in a Multipass virtual machine.

Prerequisites

• Multipass installed and running
• Enough local resources for the VM
• v8x installed and authenticated

sudo snap install multipass
multipass version
v8x login

Create the Cloud Account

v8x cloud account create local-multipass --provider on_prem
v8x cloud account list --provider on_prem

Deploy Slurm

v8x cluster create slurm-multipass-dev \
  --cloud-account local-multipass \
  --app slurm-multipass \
  --options operating_system=rockylinux9,cpu=4,mem=8,disk=128G

The --options flag is a comma-separated list for slurm-multipass only:

Key	Purpose
operating_system	One of rockylinux9, rockylinux10, noble, resolute
cpu	Number of VM CPUs
mem	Memory size in GiB
disk	Disk size, such as 128G

Inspect the Deployment

v8x cluster get slurm-multipass-dev
v8x app deployment list
multipass list

Use the deployment ID from v8x app deployment list when you need the local deployment record:

v8x app deployment get <deployment-id>

Access Slurm

multipass shell slurm-multipass-dev
sinfo
squeue
srun --nodes=1 --ntasks=1 hostname

If the VM name differs from the cluster name, check multipass list and use the VM name shown there.

Delete the Deployment

You can delete the recorded deployment and clean up associated local resources:

v8x app deployment delete <deployment-id> --force

Or delete the cluster record and request app cleanup:

v8x cluster delete slurm-multipass-dev --cluster-type slurm --app slurm-multipass --force

For low-level VM cleanup, use Multipass directly:

multipass delete <vm-name> --purge

Troubleshooting

multipass version
multipass list
multipass info <vm-name>
multipass logs <vm-name>
v8x cluster create slurm-multipass-dev --cloud-account local-multipass --app slurm-multipass -v

Common issues are usually local resource pressure, Multipass daemon startup problems, or an unsupported operating_system value.