Hey all, I have a bit of an odd question. Has anyone tried to run an RKE2 server within a pod on a host cluster / know if that’s possible? I’ve given the pod the minimum resource requirements according to https://docs.rke2.io/install/requirements#linuxwindows, mounted the containerd socket, and in the pod logs I can see RKE2 trying to come up, but etcd is having issues and not coming up even after quite some time:

time="2025-02-24T21:29:18Z" level=error msg="Sending HTTP/1.1 503 response to 127.0.0.1:54792: runtime core not ready"
time="2025-02-24T21:29:18Z" level=info msg="Failed to test data store connection: failed to get etcd status: rpc error: code = Unavailable desc = connection error: desc = \"transport: Error while dialing: dial tcp 127.0.0.1:2379: connect: connection refused\""

If it matters, this is with

debian:stable-slim

plus the rke2 install script since I’m having some trouble with

rancher/rke2-runtime

which should have

/bin/rke2

present based on the Dockerfile, but it’s not found for some reason:

docker run rancher/rke2-runtime:v1.30.9-rke2r1 /bin/rke2
docker: Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: exec: "/bin/rke2": stat /bin/rke2: no such file or directory: unknown.

Also if it helps, I’m running RKE2 with

/usr/local/bin/rke2 server -debug --disable-cloud-controller --disable=rke2-ingress-nginx --disable-kube-proxy --cloud-provider-name=external --cni=none --container-runtime-endpoint=unix:///run/k3s/containerd/containerd.sock

Thanks in advance!