Skip to content

Worker Nodes

Full build-out of a GPU compute/worker node (e.g. jrcai01).

  • NFS

    Hostname/FQDN and the NFS client mount.

    Read more

  • LDAP

    SSSD client authentication setup.

    Read more

  • MUNGE

    Installing MUNGE and copying the controller's key.

    Read more

  • GPU / CUDA

    CUDA toolkit installation for SLURM jobs.

    Read more

  • SLURM

    Installing the built package and joining the cluster.

    Read more

  • SSH Access

    Denying interactive SSH on compute nodes.

    Read more

  • clush

    Granting the controller's cluster_admin key access to this node.

    Read more