Deleting deployment does not delete pods associated with it about pykube HOT 9 OPEN

+1 to this. this seems a bit "odd" so to speak, because most users are familiar with kubectl, and would expect "delete" to delete pods as well

from pykube.

One workaround is, you can scale a deployment to 0, and then delete the deployment. Scaling it to 0 deletes the associated pods.

from pykube.

I'm using this

kubectl patch deployment some-name-integration --namespace production -p '{"spec":{"replicas":0}}'

before deleting deployment, and it works well :)

from pykube.

Deleting the Job seems not to delete the Pod neither.

May be relate to kubernetes/kubernetes#20497 . But if I try kubectl delete $job and the pods will be cleaned up.

from pykube.

This is the current expected behavior. I can't recall where I've discussed this before. I couldn't find another issue so it must have been on Slack.

The delete method on each object is the equivelent of deleting the object through the HTTP API. kubectl implements additional logic to cascade deletions to dependent objects.

This functionality is something I'd like to see in pykube and even the default behavior, but it is a feature enhancement as opposed to a bug in the current code.

from pykube.

Thanks for @brosner and it makes sense to me.

Does pykube have plan to implement the additional logic to delete all the resources? Or just wait for kubernetes team to provide the full-function APIs for that? I'm not sure if they will do that. It would be great if we can implement the feature by ourselves.

from pykube.

+1

from pykube.

Let's hope this logic will be moved to the server side in the future: kubernetes/kubernetes#12143 (issue is kind of stale)

from pykube.

We can use deletion policy to tell server to delete dependent resources.

#105 (comment)

from pykube.