Kenichi Omichi

It would be busy and mess if managing all 3 failed jobs in a single issue. This should be an umbrella issue and the separated issues will be created for...

vagrant_fedra35-kube-router also is failed and that would be the same reason at `TASK [Check the API servers are responding]`

@floryut Thanks for working for those issues. We still have 2 issues, and packet_debian9-macvlan will be solved with https://github.com/kubernetes-sigs/kubespray/pull/8971 We will need to concentrate on packet_oracle7-canal-ha after that.

Could you try Kubespray v2.19 instead of v2.17? We keep updating package versions including docker-ce like the commit 0e22a90579bde7eb864c4a9e5d02cd24b4cbbca4 to follow the upstream changes. Now we don't take effort for...

We can close this pull request because of https://github.com/kubernetes-sigs/kubespray/pull/9463

> > I think it's still worthwhile testing macvlan if we can get it to work but on more recent os like debian 11. WDYT? > > Indeed, maybe move...

> @oomichi you will need to add the handlers suggested by @floryut for the job to work, even with the rebase to debian 11 the race condition is the same...

``` Warning NetworkNotReady 38s (x151 over 5m38s) kubelet network is not ready: container runtime network not ready: NetworkReady=false reason:NetworkPluginNotReady message:Network plugin returns error: cni plugin not initialized ```

Try adding restarting-containerd

Even if restarting containerd, ``` " Type Reason Age From Message", " ---- ------ ---- ---- -------", " Normal Scheduled 6m22s default-scheduler Successfully assigned default/netchecker-agent-dhdt7 to instance-2", " Warning FailedCreatePodSandBox...