Proxmox VM Disks throw Errors (NFS issue?)

[Gandalf1783]

Hey everyone,

 

I have this issue where my Proxmox VM Disks (mounted over NFS from my UnRaid NAS) will throw ata errors after some time.

 

I figured out, that in this case, after the mover run, the issue occured:

 

Unraid Syslog:

Oct  6 03:23:00 NAS  emhttpd: read SMART /dev/sdd
Oct  6 03:40:01 NAS  crond[1062]: exit status 1 from user root /usr/local/sbin/mover &> /dev/null
Oct  6 04:01:55 NAS  emhttpd: spinning down /dev/sdd

Syslog of my VM:

You can see, that just a minute after he mover seemed to run, it started with this errors.

 

 

On Oct 4th, the mover also ran at 03:40, but my VM didnt have any errors until 11:44 that day.

In this case, Unraid mentioned a disk spindown at 10:44, but the issues arrived an hour later.

 

Is there a known problem with running VM Disks on a NFS Share?

I disabled caching (dont have a cache drive though), and i can only guess that my drives dont spin up intime to provide data.

 

I attatched my diagnostics below:

 

Dont know if this is intresting, but my NFS Share is conncted over a 10G Interface (which isnt known to me for problems), but could ofc be a potential issue.

 

Thanks!

nas-diagnostics-20221006-1010.zip

[Squid]

Have you tried reseating the cabling to the drives (power & sata, both ends)

[Gandalf1783]

Not yet, mainly because my parity checks shows 0 errors and the NAS runs fine for other VMs.

 

Keep in mind that the VM that is spitting these errors has the disks on the NAS. 

 

I am currently moving the VM disk onto a different disk in my array, cannot see a different reason. Especially since the VM disk isnt even corrupted or anything, it just seems to loose the connection.

[Gandalf1783]

So, I recently had a power outage. I plugged in a power strip which seems to have caused a short, and well, the breaker turned off.

Therefore, everything was restarted. And it seems, that after this restart, everything works fine.

The VM is now running for multiple days without corruption, in contrast to what happened before.

 

I can only assume that a software update of the NFS Drivers in Proxmox was not completed / needed a restart, or something like DMA didnt work quite out (though I never got to see errors on the PVE-Host).

Therefore, the Issue here is "resolved" while no clear solution was found.

 

For anybody else who has this, try a reboot