You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We have been noticing this when a node dies or we terminate it. The PVC remains bound to lost node and prevents the k8s scheduler from placing the pod on the new node. This requires manual intervention which is not ideal at scale given how common is to lose nodes.
The functionality is triggered by Nodes getting deleted. The exact timing and manner of deletion is quite dependent on the underlying provider. I'm going to close this as "done" but feel free to re-open if you're encountering issues with the node watcher.
We have been noticing this when a node dies or we terminate it. The PVC remains bound to lost node and prevents the k8s scheduler from placing the pod on the new node. This requires manual intervention which is not ideal at scale given how common is to lose nodes.
JIRA Link: K8S-220
The text was updated successfully, but these errors were encountered: