Hacker News new | past | comments | ask | show | jobs | submit login

You may have misunderstood me. The case I'm talking is, the node reports Ready, but the pod itself is not functioning properly.

One common issue we have is the pod gets stuck in a restart loop (for whatever reason, including starvation of resources). k8s just keeps restarting it for days on that node, instead of simply rescheduling it after X restarts or some other condition.

https://github.com/kubernetes/kubernetes/issues/13385




Isn't that what custom liveliness probes are for?


No- liveness probe just restarts the pod if the check fails, doesn't kill it.


I think a failed liveness will restart the container, not the whole pod.


You're right. We always have one container per pod so I didn't fully think about this.


Isn't that because the pod's restart policy is set to restart? What if you set that to off? Does it fail the whole pod?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: