https://rancher.com/ logo
Join Slack
Powered by
This message was deleted.
# rke2
a
This message was deleted.
c
1. did you disable it on all the servers 2. did you enable servicelb?
e
I only have one node.. and yes i use service-lb
Copy code
enable-servicelb: true
however, I do not want to use cloud-controller-manager as we don't use it
c
ah that would do it. servicelb is part of the rke2 cloud controller; if you want to use that you have to use it. If you set disable-cloud-controller it will not enable the node-lifecycle stuff, but it still needs to run to do the service controller stuff.
Are you deploying a different provider to do the node lifecycle stuff?
if you disable it and don’t provide a replacement, new nodes that you deploy will remain tainted as uninitialized.
e
it's just running a single node
never more
with it disabled we get cloud-controller in a continuous CrashLoop where healthz never passes, crashes, restarts, etc.. which is what lead me to my question
Copy code
kube-system          cloud-controller-manager-lima-ubuntu-sensor                 0/1     Running     660 (3m43s ago)   155m
Am I doing something wrong as far as configuration is concerned. This is just a single node system on bare metal, which will never have more nodes added, should I choose a diff provider?
c
No, but if you want to use servicelb you might as well leave it enabled. It costs practically nothing to run it.
e
yeah
I thought I was "optimizing" lol
thanks for clarifying
c
I’ll have to take a look to see why the health-checks are failing. I suspect that it’s because we disable the port if the cloud provider (node-lifecycle stuff) is disabled, to avoid conflicting with other cloud providers. Probably need to change or disable the health check in that case.
e
yeah, thats what I gather
thanks for the help
19 Views
Open in Slack
PreviousNext
Powered by