calico icon indicating copy to clipboard operation
calico copied to clipboard

Calico operator installation object stuck on deleting

Open rshiva777 opened this issue 1 year ago • 7 comments

I am trying to delete calico installation object, but it keep on stuck deleting

Expected Behavior

Calico installation object should be deleted

Current Behavior

Calico installation object is not getting deleted

Possible Solution

by removing finalizer from the installation object, it is being deleted

Steps to Reproduce (for bugs)

  1. helm install calico projectcalico/tigera-operator --version v3.27.0
  2. kubectl delete installations.operator.tigera.io/default

Context

Unable to delete the calico installation

Your Environment

  • Calico version: v3.27.0
  • Orchestrator version (e.g. kubernetes, mesos, rkt): 1.28
  • Operating System and version: Amazon Linux
  • Link to your project (optional):

rshiva777 avatar Jul 30 '24 08:07 rshiva777

by removing finalizer from the installation object, it is being deleted

I wouldn't recommend doing this manually, it will result in resources being left in the cluster. The operator removes this itself when it has successfully cleaned up after itself - most likely it is failing to clean up one of its resources and thus is waiting before removing the finalizer.

I'd recommend upgrading to v3.28.0 - we fixed a number of issues related to termination in that release, so it's quite possible you're running in to one of those.

caseydavenport avatar Jul 30 '24 15:07 caseydavenport

We are also facing same issues with Calico 3.27.4. We tried upgrading to Calico 3.28.0, but unfortunately, the issue persists.

Below are the environment details:

Calico versions: v3.27.4 & v3.28.0 Orchestrator version (e.g., Kubernetes, Mesos, rkt): EKS-1.28 Operating System and version: Amazon Linux

ujjwalsoni1903 avatar Aug 01 '24 13:08 ujjwalsoni1903

@ujjwalsoni1903 @rshiva777 Do you have the Calico API server running? i.e., does this show anything?

kubectl get apiservers.operator.tigera.io

If so, you need to delete that as well in order for the Installation to terminate:

kubectl delete apiserver default

If that's not the problem, could you show the output of the following commands both prior to deleting the Installation as well as after?

kubectl get installation -o yaml
kubectl get apiservers -o yaml
kubectl get tigerastatus -o yaml

As well as the logs from the operator pod after deleting the Installation?

caseydavenport avatar Aug 06 '24 18:08 caseydavenport

@caseydavenport , Thanks for your reply !!

I have checked after deleting apiserver. Still we need to remove finalizer from the installation object.

please find attached output of below commands:

kubectl get installation -o yaml
kubectl get apiservers -o yaml
kubectl get tigerastatus -o yaml

output.txt

ujjwalsoni1903 avatar Aug 08 '24 08:08 ujjwalsoni1903

That all seems reasonable to me.

I think output from after deleting might be helpful, as well as logs from the tigera-operator pod.

caseydavenport avatar Aug 19 '24 17:08 caseydavenport

@ujjwalsoni1903 @rshiva777 any new info?

tomastigera avatar Sep 10 '24 16:09 tomastigera

@tomastigera no. Need to circle back on the issue

ujjwalsoni1903 avatar Sep 10 '24 18:09 ujjwalsoni1903