Proxmox comes unresponsive after 1 or 2 days

[Jiggs]

Hi all,
I've been dealing with a Ryzen 9 system with a MSI board and every day or two the system comes unresponsive and needs to be hard rebooted. At the time this happens I get an error from storage: no such logical volume SSDPRO/SSDPRO and also am unable to backup to the PBS server with pbs1: error fetching datastores - 401 Unauthorized in the logs. At the time this happens I can't remote into the windows VM and the backups don't work. I can log into PVE on that machine, but shutting down the VM's don't work and then I have to hold down the power button to shut the host down. After the reboot, all VM's work fine and backups work fine. I really wish I could get this system running properly and am about ready to switch it to an "actual server" and not a Ryzen 9. I've even had dell laptops more stable than this machine! Right now i'm on version Virtual Environment 9.2.21. I've updated the BIOS and maybe thinking it's a memory issue? Have any of you run into this? Any help appreciated! Storage is set up with LVM - no ZFS.

 

[tschmidt]

Seems you are losing assess to storage and then everything else breaks because of that.

 

[UdoB]

First guess: check the system temperature. Some devices produces errors when running too hot. Look at the SMART-data to get an idea. Example, just for demonstration:

~# smartctl    -A  /dev/sda | grep -i temp
194 Temperature_Celsius     0x0022   071   048   000    Old_age   Always       -       29 (Min/Max 0/52)

This one is currently at 29°C and the hottest temperature was 52°C, some time in the past.

Then try to find some helpful information in the journal. man journalctl shows how to use it, but it is the reference doc, not a tutorial and so it is hard to read - but there is endless information "out there"...

If you are running the system now and the last crash was at the end of the previous boot you may look into journalctl -b -1 -e ("-b 1" = the log before the current one; "-e" shows the end, not the beginning.)