https://rancher.com/ logo
Join Slack
Powered by
This message was deleted.
# harvester
a
This message was deleted.
w
b
Ok. Me and my team went through this and tried a few things. • Trying to recreate the upgrade object to send it through again, but the validation hook prevented that as it thinks the cluster is upgraded. • Adding a new node with v1.2.2 - it would not add and was stuck in provisioning. Many errors with kubelete, containers, etc. I've never seen that happen when joining a node to cluster, but clearly even though it thinks its upgraded, it's in a strange state. • Upgrading the manifests by hand. Put the node in not ready and started evicting everything. • We ended up reverting all our changes and now have a stable cluster that is 1/3 upgraded
We're kind of stuck now
r
Do you mind sharing the cluster’s support bundle file? We can better determine the current state of the cluster and plan the next step.
👍 2
b
will do. one sec. I’ll DM both you
w
not me, just @red-king-19196
b
👍
r
Did you ever remove any 
Upgrade
resources?
b
it is not there as it thinks it’s upgraded and at v1.2.2
we tried to re-add it, but the validation hook said no as it’s already at 1.2.2
r
From the support bundle I saw three 
Upgrade
resources. All of them indicate that the upgrade failed at the image preloading state. I do notice the first node was already upgraded to the new RKE2 version, but the OS stayed at v1.2.1
What did it say when recent upgrades were blocked? May I see the screenshot of the messages?
5 Views
Open in Slack
PreviousNext
Powered by