Fail to run TASK [vexxhost.kubernetes.cluster_api : Run upgrade if required]

[okozachenko1203]

TASK [vexxhost.kubernetes.cluster_api : Run upgrade if required] ***************
fatal: [ctl1]: FAILED! => {"changed": true, "cmd": ["/usr/local/bin/clusterctl", "upgrade", "apply", "--config", "/etc/clusterctl.yaml", "--core", "capi-system/cluster-api:v1.4.4", "--bootstrap", "capi-kubeadm-bootstrap-system/kubeadm:v1.4.4", "--control-plane", "capi-kubeadm-control-plane-system/kubeadm:v1.4.4", "--infrastructure", "capo-system/openstack:v0.7.1"], "delta": "0:01:03.388960", "end": "2023-07-05 12:42:33.305584", "msg": "non-zero return code", "rc": 1, "start": "2023-07-05 12:41:29.916624", "stderr": "Performing upgrade...\nError: failed to read \"metadata.yaml\" from the repository for provider \"infrastructure-openstack\": failed to get GitHub release v0.8.0-alpha.0: failed to read release \"v0.8.0-alpha.0\": GET https://api.github.com/repos/kubernetes-sigs/cluster-api-provider-openstack/releases/tags/v0.8.0-alpha.0: 404 Not Found []", "stderr_lines": ["Performing upgrade...", "Error: failed to read \"metadata.yaml\" from the repository for provider \"infrastructure-openstack\": failed to get GitHub release v0.8.0-alpha.0: failed to read release \"v0.8.0-alpha.0\": GET https://api.github.com/repos/kubernetes-sigs/cluster-api-provider-openstack/releases/tags/v0.8.0-alpha.0: 404 Not Found []"], "stdout": "", "stdout_lines": []}

[mnaser]

This is due to kubernetes/k8s.io#5510 not being merged.

[mdbooth]

This is due to kubernetes/k8s.io#5510 not being merged.

I don't think so. It's looking for https://api.github.com/repos/kubernetes-sigs/cluster-api-provider-openstack/releases/tags/v0.8.0-alpha.0 which exists but isn't published yet. The primary reason it isn't published is because kubernetes/k8s.io#5510 isn't merged, but I'm pretty sure that's incidental.

I think this may be a bug in either clusterctl (assuming a release exists for a tag) or GH (returning a release which isn't published).

[mdbooth]

This may be fixed in kubernetes-sigs/cluster-api#7889.

[mnaser]

This may be fixed in kubernetes-sigs/cluster-api#7889.

Ah, indeed, you're right. I tested it with 1.4.4 and no good but the 1.5.0-beta.0 (which includes the fix) seems to have it.

I filed kubernetes-sigs/cluster-api#8966 upstream in the meantime.

[mnaser]

For users of this collection, the workaround for now is to login to one of the systems and run this:

GOPROXY=off clusterctl upgrade apply \
  --config /etc/clusterctl.yaml \
  --core capi-system/cluster-api:v1.4.4 \
  --bootstrap capi-kubeadm-bootstrap-system/kubeadm:v1.4.4 \
  --control-plane capi-kubeadm-control-plane-system/kubeadm:v1.4.4 \
  --infrastructure capo-system/openstack:v0.7.1

With that, it should be able to unblock this, and you should be able to run the rest of the roles with no issues.

[mnaser]

kubernetes-sigs/cluster-api#7889 seems to continue the discussion here.