This message was deleted.
# rke2
a
This message was deleted.
c
That is a really old version of rke2... I'm confused though, you said the servers all started up fine. The error message you're showing here would not happen on an agent (worker) node. Did you go back and change something on the servers after updating the server address on the agents?
s
The master nodes started fine.
The worker nodes didn't
When checking the 50-rancher.yaml file, the server referenced there doesn't exist, so I changed it to one of the working master nodes, then started rke and that's the error that comes out
This cluster was brought up about a year ago, and since it's used for observability, we don't have maintenance windows on it often.
shit - I think my server node IP's may have changed.
that subnet has dhcp on it and someone didn't static reserve the IP's for those machines
c
Yeah that would definitely break things. Server IPs need to be static, or DHCP reservations
The error message is from the server code, youre sure you're not trying to start the server service on a worker node?
s
Now I guess I have to find the old IP addresses of all the nodes. Funny the master nodes came up
c
An agent (worker) will not say "preparing server" in its logs
s
I've gotten about 6 hours of sleep in the last 4 days. What can I do to verify the worker node
c
Just confirm that you're starting rke2-agent on the workers, not rke2-server? If they are not supposed to be servers...
s
rke-agent not found
c
What are you running exactly?
And you said these are agent/worker nodes not servers, are you sure of that?
s
It's a Rancher RKE2 cluster
483042
Something is completely messed up. Set the nodes back to the proper IP, rebooted and Rancher is just not updating the status or anything.
c
Are all the servers running? Can you run
kubectl get node
from one of the servers?
It might be helpful to note which ones are servers and which ones are agents
s
cluster is ready in kubectl
and now it's showing up well in rancher.
Rancher is this mysterious thing that may look broken but eventually fixes itself.
c
yeah the cattle cluster agent needs to be running in the downstream cluster for things to sync up in Rancher
s
Thanks for listening to me complain and live stream troubleshoot. Glad it's up.