Hi, i've been upgrading my 1.8.0 cluster to 1.8.2 The first node went smooth, but the second node has hung after 'waiting for reboot'. There was a kernel panic when booting active.img. When booting with passive.img the node does come up, but I've not found good steps how to proceed from there. The partition is only 8gigs, with 2.2gb and 1.7gb for passive and active images. The node also seems to be missing all trust stores, and quite unhappy. I remember having this issue during the last upgrade but was able to solve that manually copying over the ca-bundle.pem from another node. Now I am wondering if this node has had 'issues' for a bit.

I don't remember what the partition diffs were, but I think the recommended install disk size jumped from like 100Gi->250Gi->500Gi at various points.

Thanks for the comment! I am also not sure. The drive was 2 TB, but the auto partitioning only put 8GB for COS_STATE

sda                       1.8T

├─sda1   COS_GRUB          64M

├─sda2   COS_OEM           64M

├─sda3   COS_RECOVERY       4G

├─sda4   COS_STATE          8G

└─sda5   COS_PERSISTENT   1.8T

My other two nodes have a 1tb drive and have this layout

sdb                     931.5G

├─sdb1                      1M

├─sdb2  COS_OEM            50M

├─sdb3  COS_RECOVERY        8G

├─sdb4  COS_STATE          15G

└─sdb5  COS_PERSISTENT  908.5G

I have no idea why it differs.

Yeah, I think the planned wipe/reinstall is a good idea.

Can you run this cmd and send the output ot this cmd

findmnt /run/initramfs/cos-state

Well, there is this scc.suse.com/rancher-docs/…/COS_STATE-is-too-small.html