Has anyone else run into issues with pods not starting when upgrading a cluster's K8 version, specifically

1.24.8-rancher1-1

to

1.24.9-rancher1-1

? I documented the specifics in an issue (https://github.com/rancher/rke/issues/3160), but wondering if this group has seen this issue in their own work.

it seems like using the upgraded packages and

--ignore-docker-version

as an RKE flag could solve your problems

I appreciate the reply, but I fear it is too little, too late. I ended up in a state where my nodes had Docker 23 installed, but the nodes would not start up. What I'm really wondering is why the instability with what should have been a relatively simple upgrade from 1.24.8 to 1.24.9? The agent should have been fully compatible (I was running docker 20.10 on the agents), and yet upgrading from 1.24.8 to 1.24.9 caused all of my pods to simply stop. I am also wondering if RKE2 isn't a better option at this point.

I mean if you can switch very easily, RKE2 is a good option. Did the flag not do anything?

I didn't try: The cluster that went down housed a few critical sites. Thankfully I had a set of blank nodes already provisioned for other testing, so I just moved some things around. I currently have a RKE cluster running some rest workloads, and I am definitely leaning towards that as my future for other clusters.

yup having the issue myself