Skip to content

Latest commit

 

History

History
29 lines (13 loc) · 1.29 KB

schedulerprereqs.rst

File metadata and controls

29 lines (13 loc) · 1.29 KB
Raw

Before you build clusters

  • Ensure that all cluster nodes are up and running.

  • Verify that the inventory file is updated as mentioned in the inventory sample file.

    • For Slurm, all the applicable inventory groups are slurm_control_node, slurm_node, and login.
    • For Kubernetes, all the applicable groups are kube_control_plane, kube_node, and etcd.
    • The centralized authentication server inventory group, that is auth_server, is common for both Slurm and Kubernetes.
  • Verify that all nodes are assigned a group. The inventory file is case-sensitive. Follow the format provided in the sample file link.

Note

* The inventory file accepts both IPs and FQDNs as long as they can be resolved by DNS. * In a multi-node setup, IP's cannot be listed as a control plane and a cluster node. That is, don't include the kube_control_plane IP address in the compute group. In a single node setup, the compute node and the kube_control_plane must be the same.

  • Users should also ensure that all repos are available on the cluster nodes.
  • If the cluster requires more than 10 kubernetes nodes, use a docker enterprise account to avoid docker pull limits.