K
Q

Can't delete a Stateful Set in Kubernetes

8/30/2017

I can't delete this Stateful Set in Kubernetes, even with --cascade=false so it doesn't delete the Pods managed by it.

kubectl get statefulsets

NAME                        DESIRED   CURRENT   AGE
assets-elasticsearch-data   0         1         31m

Then:

kubectl delete statefulsets assets-elasticsearch-data
^C

... hangs for minutes until I give up, then:

kubectl delete statefulsets assets-elasticsearch-data --cascade=false
statefulset "assets-elasticsearch-data" deleted
kubectl get statefulsets
NAME                        DESIRED   CURRENT   AGE
assets-elasticsearch-data   0         1         32m

I'm using Google's GKE.

-- cahen
google-kubernetes-engine
kubernetes
kubernetes-helm

Similar Questions

Kubernetes - PersistentVolumeClaim failed
minikube + Docker Volumes?
How does 'kubectl apply -f' update a deployment behind the scenes?
KVM Disksize&RAM for Minikube

5 Answers

10/2/2019

in my case i was using an old version of kubectl.

i instaled a recent one on centos via yum and the problem is solved, i am able to delete the stalled statefulset

vim /etc/yum.repos.d/kubernetes.repo

[kubernetes]
name=Kubernetes
baseurl=https://packages.cloud.google.com/yum/repos/kubernetes-el7-$basearch
enabled=1
gpgcheck=0
repo_gpgcheck=0
-- jacktrade
Source: StackOverflow
5/13/2019

If the mentioned force flags don't work, I would suggest looking at the yaml metadata block and delete any existing finalizers as well as setting blockOwnerDelete to false, then retry deleting.

-- Timothy Perez
Source: StackOverflow
11/14/2017

Had a similar issue with k8s 1.8. Tried many times and it timedout. Eventually I tried,

kubectl delete statefulsets mariadb -n openstack --force

error: timed out waiting for "mariadb" to be synced

This appears to work :

kubectl delete statefulsets mariadb -n openstack --force --grace-period=0 --cascade=false

warning: Immediate deletion does not wait for confirmation that the running resource has been terminated. The resource may continue to run on the cluster indefinitely.

statefulset "mariadb" deleted

-- jezogwza
Source: StackOverflow
8/31/2017

I could reproduce the bug twice with Kubernetes 1.7.3 and, after destroying the cluster for the 3rd time and downgrading to Kubernetes 1.6.7, I had no problem deleting Stateful Sets or Helm deployments (Elasticsearch Helm chart in my case).

-- cahen
Source: StackOverflow
8/31/2017

Try the delete action again with --grace-period=0 and --force.

-- gogene
Source: StackOverflow