External worker doesn't work

[vpm-bradleyhession]

Adding an external worker fails on every step. Kubelet doesn't start up because of no certs etc.

Provisioning stalls on kubernetes-master-01 : wireguard configured 10.0% [--------------] kubernetes-worker-01 : wireguard configured 18.2% [>-------------]

FWIW - I am adding an external hetzner Dedicated server.

[eliasp]

Same happens for me when adding a regular worker - seems to be a general issue with adding workers.

[vpm-bradleyhession]

Any updates to this? Really starting to need this now

[JohnnyQQQQ]

I'm investigating right now as we need it at work. It seems like the hcloud api is broken. The request that gets the ssh-key returns nil but no error.

https://github.com/xetys/hetzner-kube/blob/7208c4929a94d9849e71dd855707d6fd73b04f2e/pkg/hetzner/hetzner_provider.go#L47

The command don't check for nil on the ssh-key and panics when referencing nil.

[xetys]

will be closed when the progress bars render correctly, but should basically work now

[sebzimmermann]

I tried to add an external worker, but the progress is stuck for several minutes on all nodes. It doesn't seem to do anything on the external worker.

kube-master-01 : starting 0.0% [--------------] kube-master-02 : starting 0.0% [--------------] kube-master-03 : starting 0.0% [--------------] kube-worker-02 : starting 0.0% [--------------] kube-worker-04 : starting 0.0% [--------------] kube-worker-05 : starting 0.0% [--------------] kube-worker-06 : starting 0.0% [--------------] kube-worker-07 : starting 0.0% [--------------] kube-worker-08 : starting 0.0% [--------------]

How can I debug the situation?

I used the cmd hetzner-kube cluster add-external-worker -i ip --name "cluster" -d

[max-software-net]

It's waiting 10 minutes for cloud-init to finish, which is not used on my external worker...

to skip waiting I did this:

mkdir -p /var/lib/cloud/instance ; touch /var/lib/cloud/instance/boot-finished