0

I would like to run Kubernetes on a local VM/host and add compute nodes allocated via slurm job allocation on the remote supercomputer. The compute nodes on the remote supercomputer are accessible on the local host by first login to the supercomputer through ssh, which ends up on the login node of the supercomputer, and then ssh to the compute node. My goal is to run an MPI job managed by Kubernetes running on local node but executed on an HPC cluster.

Is there already some tool that can be leveraged for this kind of usage, or are there some relevant directions on how to achieve it.

Amit Ruhela
  • 311
  • 2
  • 8
  • 1
    have you seen this article? https://www.osti.gov/servlets/purl/1642947 – Bazhikov Jan 14 '22 at 23:59
  • The paper mainly discusses Flannel as CNI for K8s. I am not sure if this would work with VMs running K8s on one side and compute nodes accessible through login nodes on another side. We basically need a mechanism to pass through login nodes and make compute node directly accessible to the VM running K8s – Amit Ruhela Jan 21 '22 at 16:40
  • Have you found a solution? – Bazhikov Mar 11 '22 at 12:58
  • Not yet, all other solutions are too complex for our simple requirements. – Amit Ruhela Mar 12 '22 at 13:08

0 Answers0