VM freezes irregularly

[Neobin]

Could it be, that your cam-local-storage is simply full? (I guess, the VM in question uses this additional storage?) And therefore you get the io-error?
At least, by looking at the little indicator beside the storage, it looks quite(/too?) full...

 

[biosek]

You can try to change CPU type. I my case some VM need to use Host CPU

It is actually already on the Host CPU type

Could it be, that your cam-local-storage is simply full? (I guess, the VM in question uses this additional storage?) And therefore you get the io-error?
At least, by looking at the little indicator beside the storage, it looks quite(/too?) full...

It seems that way. It got full one time, then I deleted all the data on it and then it stayed like this in the proxmox UI

but when I look at it from the VM in that <1min, it is actually quite empty.


@fabian Will try to get the logs soon. Life just gets in the way lately

 

[Impact]

You should enable trim/discard support for your VM's disk and run fstrim -av in it.

 

[fabian]

your screenshot shows cam-local-storage being an LVM-thin storage - you also seem to have a local mountpoint that uses the same name, but that is not the storage you see on the UI

what does "lvs" output? please also post your storage.cfg!

 

[biosek]

You should enable trim/discard support for your VM's disk and run fstrim -av in it.

Ou my god...
After trimming and proxmox showing 12% used. The VM didn't crash in 1minute.
If this was it, I feel so stupid...
So far 5hours uptime...

Update: uptime 28hours, seems that was really the problem

 

[exagui]

Hey everyone,
I have a N5105 Topton mini PC and I still experience VM crashes even with updated microcode. In my case, it's the pfSense VM that goes up to 100% CPU and stops responding :


I'm on PVE 7.4-3 with kernel 5.19.17-2

cat /proc/cpuinfo | grep microcode gives me this :
microcode : 0x24000024
microcode : 0x24000024
microcode : 0x24000024
microcode : 0x24000024

This occurs every 60 days or so as you can see on my monitoring graph for uptime :


I've tried messing with CPU flags on the VM as well but it did not change anything :


Any idea what I could do ? Would upgrading the kernel to 6.2 change anything ?

I've checked the logs but I do not see any relevant info and there is no crash log in pfsense.

Thank you for your help !

 

[yurus]

Hey everyone,
I have a N5105 Topton mini PC and I still experience VM crashes even with updated microcode. In my case, it's the pfSense VM that goes up to 100% CPU and stops responding :
View attachment 67200

I'm on PVE 7.4-3 with kernel 5.19.17-2

cat /proc/cpuinfo | grep microcode gives me this :
microcode : 0x24000024
microcode : 0x24000024
microcode : 0x24000024
microcode : 0x24000024

This occurs every 60 days or so as you can see on my monitoring graph for uptime :
View attachment 67204

I've tried messing with CPU flags on the VM as well but it did not change anything :
View attachment 67201

Any idea what I could do ? Would upgrading the kernel to 6.2 change anything ?

I've checked the logs but I do not see any relevant info and there is no crash log in pfsense.

Thank you for your help !

Upgrading to PVE 8.2.2 and installing the proprietary intel-microcode package fixed the issue for me. Also try setting the BIOS back to the optimized defaults.