[k8s] Cannot remove model and cluster after trying to manage a cluster with a non-reachable controller

Bug #1828870 reported by Peter Matulis on 2019-05-13
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
juju
High
Christian Muirhead

Bug Description

I tried, perhaps foolishly, to manage a remote GKE cluster with a local LXD controller. Charm deployment failed and I realised my mistake (i.e. the cluster nodes cannot initiate TCP connections with the controller). The real problem is that I cannot remove the model, even with the new `--force` option. I then could not remove the cluster from the controller due to the existing model. I had to wipe out the entire controller, which is very unfortunate.

https://paste.ubuntu.com/p/GHMK99gDzY/

Peter Matulis (petermatulis) wrote :

A K8s namespace gets left behind in the cluster. This means the model name cannot be reused unless the operator knows how to remove the namespace:

$ kubectl delete namespace turbo

Ian Booth (wallyworld) on 2019-05-29
Changed in juju:
milestone: none → 2.6.3
assignee: nobody → Christian Muirhead (2-xtian)
status: New → Fix Committed
importance: Undecided → High
Changed in juju:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers