LonghornShareManagerOff #
Meaning #
Longhorn share manager count is off. This is likely due to a recent bug in Longhorn. It should be fixed in Longhorn version 1.6.1 and can be mitigated by an earlier hotfix. If this fix failed, this alert can reappear.
Impact #
When not all share-manager Pods are up, this means that RWX volumes are not usable or will stall on IO.
Pods needed these volumes might not start properly.
Diagnosis #
Check the number of expected Longhorn share manager and their status:
kubectl get lhsm -n longhorn-system
Compare them to the actual number of share-manager’s running:
kubectl get pods -n longhorn-system | grep share-manager
Mitigation #
Patch the longhorn share manager object with an error status:
kubectl -n longhorn-system patch lhsm --type=merge --subresource status --patch 'status: {state: error}' <longhorn share name>