Insane load on PVE host

[Nhoague]

Check this out guys, I've got a pretty beefy host, that normally runs under 1 load. I have a Windows 2012 R2 VM doing something, I can't pin point yet (youre help would greatly appreciated with any tips), but when this host gets fired up the PVE host load climbs to 10-12 completely slowing down everything. Ideas?

Currently running PVE 3.4 as I can't ug to PVE 4.x just yet.

 

[spirit]

do you see same cpu usage in windows guest ?

 

[Nhoague]

Yes the windows guest VM is pegged, I see a explorer.exe in suspended status. I'm trying to find the cause of it all, it was working great, and then one night just pegged.

 

[spirit]

So, It's 100% sure that the problem is coming from windows or explorer, not proxmox.

you can still mitigated it to not impact others vm, in vm "cpu options", you can set cpuunits
"
cpuunits: integer (0 - 500000) (default=1000)

CPU weight for a VM. Argument is used in the kernel fair scheduler. The larger the number is, the more CPU time this VM gets. Number is relative to weights of all the other running VMs.

"

 

[Nhoague]

Oh I totally agree, I will never doubt PVE.

Thanks for the great advice on units. I do remember that but was flaky on its usage.

So get this, I migrated to another host, and it picked right up. Perhaps the hardware config of this guest wasnt compatible with the host, and / or other VMs on this host were interferring. I honestly don't have an answer ... but the second the migration was done, the guest came alive and is operating normally. WTF?!

 

[PigLover]

Don't know much about why the Windows VM went wonky on you - but the 1.5+ years uptime on the PVE host is impressive...

 

[Nhoague]

Haha, thanks!

 

[FastLaneJB]

Don't know much about why the Windows VM went wonky on you - but the 1.5+ years uptime on the PVE host is impressive...

Agreed. You don't often see figures like that.

I have to say when ever I've seen servers with such a long uptime I always start to feel concerned if it'll come back up when you reboot it. I'm not sure why but the fear in me grows greater the higher the uptime number. Some update or config change that might have broken the boot but is completely unknown because it's not been rebooted forever. Am I alone in having this fear?

 

[Nhoague]

I'm really only worried about the annual fsck check, but I guess I could disable that.

 

[fabian]

one of the advantages of using virtualization is that you don't need to keep your systems running without reboots for such a long time - it's generally not a good idea anyway, because for kernel upgrades and various maintenance work reboots are required (unless you use some in place patching scheme like kpatch and your systems never require maintenance ).