I have a v1.4.3 4-node cluster upgrade to v1.5.1 t...
# harvestera
adventurous-engine-93026
07/02/2025, 6:22 PMI have a v1.4.3 4-node cluster upgrade to v1.5.1 that is hung up on the post-drain of the 1st node (been running ~3 hours). I'm wondering how to unstick it. Looks like it might be waiting for a repo? Any thoughts? I've been looking at this, but was worried about the restart process. since the guide has a huge warning about not stopping and restarting the upgrade without the direction from support. https://docs.harvesterhci.io/v1.5/upgrade/troubleshooting/#restart-the-upgrade
r
rhythmic-article-81903
07/02/2025, 9:35 PMcheck the log of post drain pod in that node... most probably it is waiting for rke2 to upgrade (and most probable fix in my experience is restart of that node). For some reasons rke2 get stuck and continues only after restart of the node. I have upgraded 12 nodes so far and I think I got 2 or 3 of these cases.
a
adventurous-engine-93026
07/02/2025, 10:27 PMThe log looks like it's waiting for the repo to become available (see screenshot). I rebooted the node a couple hours ago.
r
rhythmic-article-81903
07/02/2025, 10:35 PMthat is a new one. Have you tried to restart that pod?
a
adventurous-engine-93026
07/03/2025, 12:51 AMI haven't. How would I do that?
r
rhythmic-article-81903
07/03/2025, 7:55 AMjust delete that pod and it will recreate
a
adventurous-engine-93026
07/03/2025, 1:33 PMI deleted the post drain pod. No change. If I remove the harvester repo VM pod, will it auto recreate?
adventurous-engine-93026
07/03/2025, 2:08 PMAnd I rebooted it again a while after deleting the pod. No change. Ugh, I've never had a cluster upgrade finish without issue.
adventurous-engine-93026
07/03/2025, 3:01 PMTo answer my own qquestion, yes, you can delete teh virt-launcher vm pod, and it will get recreated. And that actually resolved the issue!