Single node clusters support

[pablo-suazo]

I would like to use descheduler in a single node kubernetes cluster.

It is useful to me to automatically delete pods with too many restart or failed pods.

It is possible to overrride this check in any way?

if len(nodes) <= 1 { klog.V(1).InfoS("The cluster size is 0 or 1 meaning eviction causes service disruption or degradation. So aborting..") return fmt.Errorf("the cluster size is 0 or 1") }

[a7i]

/assign

[k8s-triage-robot]

The Kubernetes project currently lacks enough contributors to adequately respond to all issues.

This bot triages un-triaged issues according to the following rules:

• After 90d of inactivity, lifecycle/stale is applied
• After 30d of inactivity since lifecycle/stale was applied, lifecycle/rotten is applied
• After 30d of inactivity since lifecycle/rotten was applied, the issue is closed

You can:

• Mark this issue as fresh with /remove-lifecycle stale
• Close this issue with /close
• Offer to help out with Issue Triage

Please send feedback to sig-contributor-experience at kubernetes/community.

/lifecycle stale

[vaibhav2107]

/remove-lifecycle stale

[larryrensing]

any movement on this? it would be great to have!

[erikgb]

This issue becomes even more annoying when switching from cronjob to deployment workload. Since the rescheduled aborts, this makes the deployment pod crash-loop.

I suggest adding an option to enable support for single-node clusters. The default can still be to abort in single-node clusters to give users more awareness of the additional disturbance running descheduler with just a single cluster node.

[panteparak]

At work, I scale down replicas down to 1/2 on many of our apps during night times, which reduced down the number of nodes needed to 1 and cause the same error as above. It would be good if some sort of flag such as single node suppress such error