Max Gautier
Max Gautier
Upgrading to v1.18.2 does not fix the issue, AFAICT.. Also, I missed the following in the cert-manager-webhook logs : ``` W0904 13:05:30.929520 1 reflector.go:569] k8s.io/[email protected]/tools/cache/reflector.go:251: failed to list *v1.Secret: Get...
I thought that initially as well, but other pods appeart to contact the API server just fine, so I'm not sure what's happening. I'll keep investigating. EDIT: other pods on...
> but with escape chars in inotify the sysctl -p fail and based on that install failed > when i remove \ manualy on each node it started working What...
The rsync stuff is going away soonish (™️ ), see #12299 Would that still be needed when not using synchronize ?
I'd put this in the defaults instead, as done elsewhere, to have one default definition instead of several scattered across usage sites.
IMO the main divide between experimental or not should probably be wheter it's tested in CI. We could add some tests for CRI-O, in particular for upgrade and scale. (Our...
I don't think we will be adding new stuff to kubernetes-apps. Kubespray direction is more to re-focus on the core Kubernetes cluster bring-up. (this migth need an dedicated issue though,...
From what I can tell the reset playbook import internal_facts, which already gather some facts. Would'nt it make more sense to gather it here ? :thinking:
Hum the linked issue appear to invoke the role from remove-node.yml which also include internal facts, it seems.
(and we should probably add a reset node test to one of the rhel9-like CI job to ensure this does not regress)