Hi. I've installed the auto-updater into my one no...
# k3s
a
Hi. I've installed the auto-updater into my one node cluster and it immediately started updating it. However, it's been doing it for 20 minutes now without any logs whatsoever. Any insights into how long this usually takes? Ingresses don't work at the moment.
c
what do you mean by “auto-updater”.
Do you mean the system upgrade controller? If so, what Plan did you create for the upgrade? What does the upgrade job pod log show?
a
https://docs.k3s.io/upgrades/automated I did exactly what says here. Installed the CRDs, and then created a char with those two plans, server and agent. I originally installed the k3s with the script, from what I understand that means I can use the controller. The logs say nothing. Both the controller and the job are just empty. Also, I just tried accessing a service with port-forward and with exec and I'm getting 503 service unavailable.
c
why did you create a server and agent plan if you have only a single node?
What version are you currently running, and what version did your Plan say to upgrade to?
a
well for future profing, but that should be a problem since the selectors should prevent the agent node to do anything, right?
c
If you just copy-pasted then that plan would “upgrade” you to 1.24.6 which probably is not an upgrade at all unless you were WAY behind.
a
kubectl currently reports v1.24.6+k3s1, which is the version i set to be upgraded to. i don't know what version was before that...
c
You probably just downgraded yourself and thats why everything is broken
a
i mean i've been running the k3s for a few months
huh, ok
c
these are examples, you’re not supposed to just copy paste them. You need to think critically about what you’re doing.
a
i just figured the docs will have the latest version
well, can i just set the channel to stable
c
no, they have the version that was latest when they were written. We don’t upgrade the examples every time we put out a new release.
a
what do you suggest in terms of recovery?
c
1.24.6 is from like, 3 years ago. when we last updated those docs.
😂 1
you can TRY replacing the k3s binary on disk with whatever version you were running before, then restarting the service. That should hopefully get you back to where you were before.
But we really don’t support downgrading like this, so its possible you’ve hosed things up and will need to either restore from a backup, or rebuild from scratch.
At some point in the last few years we added downgrade protection, so it would have just failed. but I think 1.24.6 is before we added that, so that doesn’t help you any.
a
Yeah, just replacing the binary and restarting it worked, tnx
c