A step-by-step guide to deploying Slurm HPC workloads on Red Hat OpenShift using the Slinky operator. Covers both web console and Helm CLI installation methods, namespace setup, authentication secrets (JWT and Slurm keys), Controller and NodeSet custom resource deployment, cluster verification, job submission examples (basic, resource-constrained, long-running, Python batch), troubleshooting common issues, and manual/auto scaling of compute pods.
Table of contents
What is Slinky?Why run Slurm on OpenShift?Slurm architecture on OpenShiftThe OpenShift clusterInstall SlurmTest the clusterTroubleshootingRunning jobsScale workloadsFinal thoughtsSort: