Fix the failing interactive job submits in knightly OOD
Description
The interactive job submission failed on the knightly OOD UI failed because it couldn't find the munge socket at /var/run/munge/munge.socket.2
. This occurred because the munge service did not start during deployment. The ansible task responsible for it only enabled it, but did not start it. Since it's a deployment task, we need to start the service.
Solution
Start the munge services in the ansible role - ansible/roles/slurm_client/tasks/main.yml
Note: The munge and slurmd services are enabled together in the ansible task. We need to see how we can split it up so that only munge service is started because that is the only service required for this issue.
Edited by Eesaan Atluri