Hi all. We are running Rancher 2.6.9 and upgraded on of the clusters managed by rancher from 1.23.15 to 1.24.10 We saw a lot of things going wrong. Nodes had trouble communicating to each other, Some storage mounts ( Ceph block storage ) didn't want to unmount. We basically had to restart all our nodes one by one to recover everything. Has anyone else seen such behaviour?
We do know that the cri-docker replaced the docker shim part, could this be a root cause of it?
03/23/2023, 3:15 AM
ive seen the ceph not wanting to unmount on node reboot issue a lot, i thought it was a normal issue tbh since ive seen it on pretty much every version of 1.24 hah
03/23/2023, 9:33 AM
It wasn't a reboot issue or something like that/ we did a kubernetes upgrade via rancher 🙂