Uploaded image for project: 'Red Hat Workload Availability'
  1. Red Hat Workload Availability
  2. RHWA-285

FAR: Support off Action

XMLWordPrintable

    • True
    • False
    • Hide
      Cause: Sometimes admin don't trust the unhealthy node and don't want it to access the cluster. So the admin prefers to shut down the node and the Fence Agents Remediation (FAR) Operator has supported only the reboot action.
      Consequence: Admins are unable to investigate and troubleshoot the unhealthy node, when FAR is turning the unhealthy node to off followed by on.
      Fix: FAR supports the off action which keep the node shutdown, while the node is fenced but not recovered. The admin will be responsible to turn the node on (or to recover the node) after a possible further investigation.
      Result: Admin can run FAR with the option of shutting down the node and troubleshoot why it has become unhealthy.
      Show
      Cause: Sometimes admin don't trust the unhealthy node and don't want it to access the cluster. So the admin prefers to shut down the node and the Fence Agents Remediation (FAR) Operator has supported only the reboot action. Consequence: Admins are unable to investigate and troubleshoot the unhealthy node, when FAR is turning the unhealthy node to off followed by on. Fix: FAR supports the off action which keep the node shutdown, while the node is fenced but not recovered. The admin will be responsible to turn the node on (or to recover the node) after a possible further investigation. Result: Admin can run FAR with the option of shutting down the node and troubleshoot why it has become unhealthy.
    • Feature
    • Proposed

      After limiting FAR to one action, reboot, we would like to enable it's support to the off action.

      It is required when an admin don't trust the unhealthy node and don't want it to access the cluster. So the admin prefers to shut down the node. The off (shutdown) action ensures fencing just without recovering the node. But then the admin/user will be responsible to turn the node on (or to recover the node) after a possible further investigation.

      Currently, the reboot action is equivalent to run off and then on for a specific node.

      See https://github.com/ClusterLabs/fence-agents/blob/main/doc/FenceAgentAPI.md#agent-operations-and-return-values for more available actions for FAs from ClusterLabs

              kkii@redhat.com Keiichi Kii
              oraz@redhat.com Or Raz
              Votes:
              1 Vote for this issue
              Watchers:
              5 Start watching this issue

                Created:
                Updated:
                Resolved: