# rke2


03/30/2023, 12:48 PM
Apologies for the total noob question but I am new to RKE2 and trying to figure out how to upgrade it properly. I inherited a v1.21.3 cluster which I've tried upgrading to v1.21.14 manually per the instructions here: . Details were sparse so I stopped the rke2-server service on a master node and then installed v1.21.14 via the installation scripts. I then restarted the rke2-server service and the node appears functional, with working pods. However, a
kubectl get nodes
says it the master node is still v1.21.3. Checking
rke2 --version
says it is v1.21.14 - what am I missing?


03/30/2023, 9:49 PM
I don't have the full picture either, but if the cluster was created in Rancher Manager then you'll have some tools at your fingertips there. RM is a cluster manager and it can manage the lifecycle of any cluster it creates, and to a certain extent, any cluster that's imported in Rancher Manager. If you don't already have Rancher Manager running, you could start by installing it in Rancher Desktop or by running the container image, e.g.
docker run
. Some people run Rancher Manager inside the cluster they're managing, which is fine, I suppose, if you're very careful not to cut the branch you're standing on.


03/31/2023, 12:48 PM
Also if you do the default manual install, it installs via RPM and it'll install a repo in the process you'll end up updating with your normal yum/dnf upgrades (at least it does with RHEL-clones, so I assume it does apt with Ubuntu as well). Not that it helps with mismatched versions. The only time I had something like that I spent 4-6 weeks debugging, found out what was wrong but had no way to fix it and eventually just went the nuke & pave route (I only waited that long because my task was experimenting and documentation for later troubleshooting).