[bug] kube-proxy not upgraded to v1.26.4

[gerhard]

Is there an existing issue for this?

• [X] I have searched the existing issues

Current Behavior

On Omni v0.6.2, I triggered a K8s upgrade from v1.26.3 to v1.26.4. All components except kube-proxy got upgraded correctly:

> omnictl get rcmc -o yaml | rg v1.26
                image: ghcr.io/siderolabs/kubelet:v1.26.4
                image: registry.k8s.io/kube-apiserver:v1.26.4
                image: registry.k8s.io/kube-controller-manager:v1.26.4
                image: registry.k8s.io/kube-proxy:v1.26.2
                image: registry.k8s.io/kube-scheduler:v1.26.4

Expected Behavior

kube-proxy should be upgraded to v1.26.4 too.

Steps To Reproduce

Given a cluster running K8s v1.26.4, click the upgrade K8s to latest button in Omni.

What browsers are you seeing the problem on?

No response

Anything else?

No response

[smira]

The fix will come in the next version of Omni, at the moment it doesn't update the bootstrap manifests (yet), thanks for reporting!

[gerhard]

OK, so after Omni v0.7.0. No worries. Keeping this one open then. We are down to just 1 issue then. Great job!

FWIW, I am very close to rolling this in a production cluster. I am aiming for this week. Looking forward to it!

[gerhard]

I can confirm that this is now fixed in Talos v1.4.0 managed by Omni v0.7.1:

omnictl get rcmc -o yaml | rg image

                image: ghcr.io/siderolabs/kubelet:v1.26.4
                image: ghcr.io/siderolabs/installer:v1.4.0
                image: registry.k8s.io/kube-apiserver:v1.26.4
                image: registry.k8s.io/kube-controller-manager:v1.26.4
                image: registry.k8s.io/kube-proxy:v1.26.4
                image: registry.k8s.io/kube-scheduler:v1.26.4

Thank you @smira 💪