Skip to main content

SLURM Cluster Management

Overview​

Slurm (Simple Linux Utility for Resource Management) is a workload manager used for high-performance computing (HPC), job scheduling, and batch processing on allocated compute resources.

In Armada Bridge, you can create and manage Slurm clusters, submit and monitor jobs from the UI, and view cluster resources and the Slurm dashboard.

This guide covers:

  • Creating a Slurm cluster (name, version, worker and master node selection)
  • Monitoring cluster creation and viewing master/worker details
note

Bridge supports Slurm cluster creation using Slurm version 24.05.6. This version is available when you create a cluster (see Step 2: Configure Cluster).

Prerequisites​

  • Tenant Admin access — Log in as a Tenant Admin to create and manage Slurm clusters.
  • Bare Metal or Virtual Machine resources allocated to your tenant
  • Minimum two nodes (Bare Metal or VM) available — one for the master node and at least one for worker nodes

Create Slurm Cluster​

Step 1: Navigate to Slurm​

  1. In the left sidebar, open Compute → Slurm.

  2. Click Create Slurm Cluster.

    Slurm Navigate

Step 2: Configure Cluster​

  1. Enter the Slurm cluster name (e.g., slurm-cluster-01) and Description (e.g., Creating Slurm cluster).

  2. Select the Version (e.g., 24.05.6) from the dropdown.

  3. Click Next.

    Slurm Configuration

Step 3: Worker Node Selection​

  1. Select one or more worker nodes.

  2. Click Next.

    note

    At least one worker node is required to create a Slurm cluster. You can select Bare Metal or Virtual Machines.

    Slurm Worker Node

Step 4: Master Node Selection​

  1. Select the master node.

  2. Click Next.

    note

    Only one master node is required. You can select a Bare Metal or Virtual Machine.

    Slurm Master Node

Step 5: Create Slurm Cluster​

  1. Review the configuration and click Create to start Slurm cluster creation.

    Create Slurm Cluster

Step 6: Monitor Cluster Creation​

  1. After creation starts, the cluster state shows Processing.

    Processing Slurm Cluster

  2. When creation completes successfully, the status shows Success.

    Success Slurm Cluster

note

Cluster creation typically takes around an hour, depending on the compute nodes or virtual machines used.

Step 7: View Master and Worker Details​

  1. Click the ellipsis (three-dot) menu for the slurm cluster.

  2. Click the View (eye icon) to see the Slurm master and worker node information.

    View Slurm Creation Button

    View Slurm Nodes Details

Common Use Cases​

  • Batch processing — Run batch jobs across multiple nodes
  • HPC workloads — Execute high-performance computing tasks
  • Parallel computing — Distributed computing tasks
  • Machine learning training — GPU-accelerated ML jobs

Next Steps​

Last updated on