CreateContainerError after first `kubeadm join`

General Discussions
1

After successfully running my first kubeadm join, of a worker node to the existing control node, I see three pods on the new node, but all have a status of CreateContainerError or Init:CreateContainerError. I’m not sure how to debug this.

describe pod of the kube-proxy pod (that’s on the new node) gives a curious Events list:

Type    Reason  Age                   From     Message
----    ------  ----                  ----     -------
Normal  Pulled  89s (x775 over 170m)  kubelet  Container image "registry.k8s.io/kube-proxy:v1.31.2" already present on machine

That’s it. I also see in the output

IP:
IPs:          <none>

Not sure whether these should have values yet.

In journalctl, I see a lot of these from kubelet:

E1113 … driver-call.go:262] Failed to unmarshal output for command: init, output: "", error: unexpected end of JSON input
W1113 … driver-call.go:149] FlexVolume: driver call failed: executable: /usr/libexec/kubernetes/kubelet-plugins/volume/exec/nodeagent~uds/uds, args: [init], error: executable file not found in $PATH, output: ""
E1113 … plugins.go:691] "Error dynamically probing plugins" err="error creating Flexvolume plugin from directory nodeagent~uds, skipping. Error: unexpected end of JSON input"

journalctl --grep=CreateContainerError returns messages like:

CreateContainerError: container create failed: cannot open sd-bus: No such file or directory

:sos::exclamation:

Cluster information:

This is an on-prem system, one physical machine, nodes are Xen domUs.

What confounds me about this is that the control node should be set up the same way as the worker node domU. Yet I don’t see the issue over there.

2

I forgot to add the IPv6 address to the worker’s /etc/network/interfaces when I built the domU. That may solve the issue; I’ll check tomorrow.

I’ve been setting this up as a dual-stack cluster. Some of the messages that I’ve been seeing have mentioned “addresses”. I just found one in journalctl -u kubelet that said:

E1113 … kubelet_node_status.go:706] "Failed to set some node status fields" err="failed to validate secondaryNodeIP: node IP: \"…\" not found in the host's network interfaces" …