VM freeze at 100% CPU (no windows)

[pos-sudo]

Dear,

I really hope you guys could help me out. We have an Proxmox 8.0.3 installation with kernel 6.2.16-3. Lately we encouter issues the VM becoming unresponsive after it reached 100% CPU for a time, I have to stop and start again in order to get the VM back online. We have no Windows VMS only Ubuntu 20.04 LTS or Ubuntu 22.04 LTS. I never had this before on previous installations of PVE, is this something that is fixed in PVE 8.1 or something else I can do in order to prevent it from happening again.

I would like to hear from you.

Thanks regards,

 

[gzader]

I believe you need 6.2.16-4 or greater.
Upgrading your installation (and rebooting) should get you what you need.

 

[pos-sudo]

I believe you need 6.2.16-4 or greater.
Upgrading your installation (and rebooting) should get you what you need.

Thanks for the quick reply. Is it true that this is completey random or is it an behaviour which I simple can reproduce?

 

[gzader]

This thread:
https://forum.proxmox.com/threads/vms-freeze-with-100-cpu.127459/page-12

Has a huge amount of details. Solutions start to show up around page 8. People can cause the freeze / lockup.
Having said that. I've just posted because I'm running all the latest and I still have a freezing issue.

 

[pos-sudo]

This thread:
https://forum.proxmox.com/threads/vms-freeze-with-100-cpu.127459/page-12

Has a huge amount of details. Solutions start to show up around page 8. People can cause the freeze / lockup.
Having said that. I've just posted because I'm running all the latest and I still have a freezing issue.

Just me thinking, could It be that it is the new cpu unit default which is 100 but this should be 1024 compared to other PVE installations?

 

[gzader]

Just me thinking, could It be that it is the new cpu unit default which is 100 but this should be 1024 compared to other PVE installations?

That's outside of my knowledge level. I've got enough knowledge to set up a cluster, ceph, ha, replication and import a drive image.

 

[pos-sudo]

That's outside of my knowledge level. I've got enough knowledge to set up a cluster, ceph, ha, replication and import a drive image.

Okay thank you! But you havent managed to mitigate the VM freeze problem?

 

[gzader]

Okay thank you! But you havent managed to mitigate the VM freeze problem?

Not mine. But most of the people in that other thread seem to have fixed their issues. It's marked as solved.
My post is awaiting moderation before it shows up in the list. I may have a different issue entirely from the ones in that post.
In your case, I'd say run the upgrade. You're running a version with a known issue.

Also, let me know if it fixes it. I could use some good news today.

 

[pos-sudo]

Not mine. But most of the people in that other thread seem to have fixed their issues. It's marked as solved.
My post is awaiting moderation before it shows up in the list. I may have a different issue entirely from the ones in that post.
In your case, I'd say run the upgrade. You're running a version with a known issue.

Also, let me know if it fixes it. I could use some good news today.

I would many many thanks!!

 

[pos-sudo]

Not mine. But most of the people in that other thread seem to have fixed their issues. It's marked as solved.
My post is awaiting moderation before it shows up in the list. I may have a different issue entirely from the ones in that post.
In your case, I'd say run the upgrade. You're running a version with a known issue.

Also, let me know if it fixes it. I could use some good news today.

Hello,

I just updated and rebooted one of our "affected" servers to kernel 6.5 I will let you know if we encouter any further issues.

 

[pos-sudo]

Hello,

We just encountered another freeze but this time the freeze was 4000 above the integer previously noticed on "2147483648". The VM froze at 2147488074. I think we now could consider that the next update of the kernel (which already launched) fixed this random freeze issue. What do you think?