Corrupt kernel. How safe to update?
So I've ended up with a corrupt kernel on my PVE host. I didn't even realise until I tried to install Tailscale as an LXC, and it turns out tun functionality is missing. As a workaround, I've got Tailscale running on a VM for now.
I had CoPilot walk me through trying to update it from the shell, but a matching kernel wasn't available. Something about these being custom Proxmox kernels with ZFS support. However, I'm seeing a kernel update to the right version listed in the updates within the UI.
I think the underlying problem was that 'local' had 100% filled. An external drive had failed to automount and a whole bunch of backups had written to the mount folder on the local drive instead and completely filled it. Having got to the bottom of that (dismounted the drive and emptied the underlying folder) I'm back down to <10% full on 'local' rather than the 99.5% I was at before. Possibly an update occurred without my knowledge and failed to complete due to lack of disk space?
My dilemma is:
- Do I leave this well alone? It's working fine as long as I run Tailscale from the VM, so I don't really *need* to fix it.
or
- Do I update it and fix the underlying problem so that I can revert to running Tailscale as an LXC?
Ultimately, I know the worst case scenario is that I'm reinstalling Proxmox from scratch and restoring all my backups. But that's a lot of work for very little reward.
Should I push the button on this anyway?
(Oh, and I'm really struggling to type 'kernel' with an E, having spent many hours writing 65xx code for the Commodore 64 back in the day.)