https://rancher.com/ logo
Join Slack
Powered by
This message was deleted.
# general
a
This message was deleted.
p
This is normal, this is because the component responsible for managing these certs is not up yet.
b
Really? It’s been a couple of days now…
p
Oh.
So the bootstrap failed somehow, but this error is a red herring. At least in my experience
wait you said "I’m trying to provision a k3s cluster through rancher" do you really mean a "vanila" k3s cluster managed and provisionned by rancher, or do you actually mean a rke2 cluster (default)?
b
A vanilla k3s cluster. Rancher itself is hosted on a rke2 cluster.
p
Is there a particular reason to use k3s instead of rke2? I'm unfamiliar with k3s sheneningans when piloted by rancher. I know interesting errors happen in the service rke2-server for me, not really the agent
b
I want to deploy a really simple “single node” k8s instance.
p
I understand, but rke2 makes things easier, at least to bootstrap.
Isnt there a k3s service somewhere on your node?
b
yes. It failed:
Copy code
root@icvm0169:~# systemctl status k3s.service
● k3s.service - Lightweight Kubernetes
     Loaded: loaded (/etc/systemd/system/k3s.service; enabled; vendor preset: enabled)
     Active: activating (auto-restart) (Result: exit-code) since Tue 2024-05-14 14:17:44 CEST; 4s ago
       Docs: <https://k3s.io>
    Process: 9220 ExecStartPre=/bin/sh -xc ! /usr/bin/systemctl is-enabled --quiet nm-cloud-setup.service 2>/dev/null (code=exited, status=0/SUCCESS)
    Process: 9222 ExecStartPre=/sbin/modprobe br_netfilter (code=exited, status=0/SUCCESS)
    Process: 9223 ExecStartPre=/sbin/modprobe overlay (code=exited, status=0/SUCCESS)
    Process: 9224 ExecStart=/usr/local/bin/k3s server (code=exited, status=1/FAILURE)
   Main PID: 9224 (code=exited, status=1/FAILURE)
        CPU: 149ms
p
Check its logs (journalctl -eu k3s)
b
it fails to start with the following error message:
Copy code
invalid port range node.example.com: strconv.Atoi: parsing \"node\": invalid syntax
p
No idea.
b
Thanks @powerful-librarian-10572 for your time and help.
p
Just to be sure, is your node a fresh install?
b
yes.
p
I mean, a fresh linux install
b
Yes. A brand new VM
p
So, just in case, delete the cluster, uninstall rancher-system-agent (there should be a script), unistall k3s (i hope there is a script as well) and try again
If it's not working still, i will only be able to help you further if you switch to a rke2 cluster
b
Is rke2 OK to run on a single node?
p
Yessir
But you wont be able to make it go down easily
b
go down?
p
Etcd really like to have a teacher at any time, recovering from a completely shut down cluster is a handful
Go down as in making the rke2 server go down on your node (for testing or following a restart, for example)
b
ok… I wanted to avoid having 5 nodes (3 control-planes and 2 workers) for hosting Harbor but I guess it is the price to pay.
211 Views
Open in Slack
PreviousNext
Powered by