kops
kops copied to clipboard
Ubuntu 22.04 Jammy Jellyfish tracker
Various things we need to cross off before defaulting to 2204
- [x] Cilium tests are failing
- [x] https://github.com/kubernetes/kops/issues/14139
- [x] Are things ready for cgroupv2?
- [x] https://github.com/aws/amazon-vpc-cni-k8s/issues/2103
https://github.com/kubernetes/kops/issues/14140#tasklist-block-7f00af88-8494-4b65-ab20-1a46222e069b
cgroupv2 is worth of mentioning in release notes at least
With k8s 1.25, support for cgroups v2 graduates to stable.
Was lucky enough to catch one of the weekly cilium u2204 tests passing this night: https://testgrid.k8s.io/sig-cluster-lifecycle-kops#kops-grid-cilium-etcd-u2204-k22-docker
/remove-lifecycle stale
/remove-lifecycle stale
/lifecycle frozen
Last open issue in the checklist here was closed (well, rotted away) - perhaps I'll give 22.04 a try and open issues if I run into any problems with it. Are there any other known issues not already noted here?
FWIW, I haven't had any issues running on 22.04 in the last two weeks, though I haven't tested the specific scenarios that were raised in issues here, like dind.
@danports Ubuntu 22.04 will be the default in kOps 1.27.
FWIW, I haven't had any issues running on 22.04 in the last two weeks, though I haven't tested the specific scenarios that were raised in issues here, like dind.
@danports , could you please let us know what version of KOPS were you testing Ubuntu 22.04 LTS with? I am seeing similar symptoms with KOPS v 1.23 on Ubuntu 22.04 LTS. We see 'timeouts' with e.g. EBS CSI and coredns trying to connect to the kube-apiserver endpoint within the cluster network. We use cilium, but all pods are running fine.
I've been running Ubuntu 22.04 with every version of kOps since 1.26.3. kOps 1.23 is over a year old now, so I'd recommend upgrading to the latest version if possible and seeing if the issue still persists.