[SOLVED] Proxmox 6.4-13 node (& VMs) gone in web frontend

[MoxProxxer]

One machine of mine seems to have lost track of everything
The node does not show, also no VMs running on that node are not shown.
However, both node and VMs continue to operate fine.
There's no cluster configured, it's a standalone node. Web frontend looks a little borked too.

Now I tried everything except rebooting the machine.

# service pveproxy restart

to no avail. For the life of me I cannot figure out what's wrong, as I have another machine running the exact same PVE version, also just as standalone node and everything is ok there.

Any hints what I could try?

 

[MoxProxxer]

It seems this is the problem

# qm list
ipcc_send_rec[1] failed: No such file or directory
ipcc_send_rec[2] failed: No such file or directory
ipcc_send_rec[3] failed: No such file or directory

although no idea as of why.

https://forum.proxmox.com/threads/ipcc_send_rec-errors-after-fresh-install.66885/ seems not related.

This seems very similar to what I observe
https://forum.proxmox.com/threads/ipcc_send_rec-failed-on-a-non-cluster-installation.14224/

but /etc/pve is mounted and looks ok to me. "service pve-cluster start" also throws no error

/dev/fuse                  30720         48      30672   1% /etc/pve

 

[fiona]

Hi,
what is the output of the following:

ls -l /etc/pve
ls -l /etc/pve/nodes/<nodename>
pveversion -v
systemctl status pve-cluster.service

Is there anything interesting in /var/log/syslog?

 

[MoxProxxer]

I seem to have solved it. There was a stray "stretch" left in

/etc/apt/sources.list.d/pve-install-repo.list

deb http://download.proxmox.com/debian/pve stretch pve-no-subscription

when system was on buster already. Not sure how that happened, because as I wrote I have a "twin" machine where basically everything is alike (and worked) and I used that for comparison.

I also only stumbled across that apt source problem when I decided to do a dist-upgrade and suddenly saw that the system downgraded proxmox from 6.4 to 5.4 (or 5.3 don't remember).

After changing that source to buster, doing a dist upgrade and restarting pvedaemon and pveproxy, reloading the web frontend, logging out, logging in, reloading again - and probably some rain dance - things started to work.

A few more glitches when clicking on VMs and retrieving infos - again a few reloads, and all seems to be working now.

So "SOLVED", probably the error was sitting in front of the keyboard, but for the life of me I do not fathom where I did take that wrong turn. Everything worked perfectly on both nodes for months and then suddenly one stopped working.