1

Our GKE cluster can't start any pods with following error:

failed calling webhook "gvisor.common-webhooks.networking.gke.io": Post https://localhost:5443/webhook/gvisor?timeout=30s: dial tcp [::1]:5443: connect: connection refused

luckily we were able to recover by upgrading master version, which triggered restart of a control plane, but what if we were already on a latest master version? As far as I am aware there is no way to cycle master nodes via API.

Maxim Ivanov
  • 11
  • 2

1 Answers1

0

The error message basically just mean that the common webhook went down for a short period of time.

However, version 1.14.3-gke.12 or above should fix this issue in regards to webhook.

But you have said that you already upgraded your master version, so you should be okay! If you haven't upgraded to 1.14.3-gke.12 or above yet, please do so.

If you need to know how to manually upgrading a cluster or node pool, just click on this link.

Hope this helps!

Anthony Leo
  • 181
  • 4
  • Hi Anthony, my question is more about what to do if we hit the bug like that, but there is no new version released yet. – Maxim Ivanov Oct 31 '19 at 10:44
  • Hello again, it seems that this error should not happen again after version 1.14.3-gke.12 is released. However, I cannot provide a date on when that version will be released. – Anthony Leo Nov 04 '19 at 13:18
  • If you do hit a bug similar again in the future, it would be advised to create a support case with GCP so that way, they can fully analyze your project and give you better results. Here is the link: https://cloud.google.com/contact/ – Anthony Leo Nov 04 '19 at 13:19
  • Just as additional information, if you will ever encounter similar situation, you can reach `Google Support` using [Public Issue Tracker](https://issuetracker.google.com/) and get additional information, if this is a Bug or issue with your cluster configuration. – PjoterS Feb 12 '21 at 12:43