https://rancher.com/ logo
#k3s
Title
# k3s
a

adamant-kite-43734

12/07/2022, 4:10 PM
This message was deleted.
c

creamy-pencil-82913

12/07/2022, 4:12 PM
you could do a rolling move, by deleting nodes from the cluster and then re-joining them with their new IP.
as long as the two networks have connectivity with each other
w

wide-author-88664

12/07/2022, 4:15 PM
They kind of do… the new network is going to be behind a NAT gateway, so the old networks’ nodes won’t be able to communicate with new (moved) nodes… New nodes can easily talk to old nodes tho.
And, in this case, it’s only a two-node cluster…
Copy code
root@demo-nb-12110:~# kubectl get nodes
NAME            STATUS   ROLES                  AGE   VERSION
demo-nb-12110   Ready    control-plane,master   31h   v1.25.4+k3s1
demo-nb-12111   Ready    <none>                 11s   v1.25.4+k3s1
c

creamy-pencil-82913

12/07/2022, 4:31 PM
for a single-server cluster you could probably get away with just stopping the server, running the killall script, and then bringing it back up with the new IP. Do the same with the agent, but change the server’s IP in the config before bringing it back up. This is more complicated with embedded etcd, because you have to manage quorum while moving the nodes… but not a problem for you.
w

wide-author-88664

12/07/2022, 4:34 PM
Cool, thanks
3 Views