[SOLVED] High CPU Ussage till complete UI Freeze after update to 3.2-8

[robinrm]

Hello guys.

after the update that was released on Sunday last week i've major issuse with an abnormal CPU Ussage on my two backupservers.
There is no task aktive and the IO time ist at 0.00 till 0.1%.
The CPU Ussage will go in little steps up till 100% and than the UI is complete unresponsible. SSH is working fine.
After a restart with

systemctl restart proxmox-backup-proxy

the CPU Ussage is ~0 again and the UI works again.

First Server runs on an 6 Core Intel(R) Xeon(R) CPU E5-1650 v4 with 128gb of RAM and second on an 6 Core Intel(R) Core(TM) i7-8700 CPU with 64gb of RAM.
Both use the latest Kernel 6.8.12-4 (2024-11-06T15:04Z) with no Subscription Repository.

An downgrade to Kernel 6.8.12-3-pve (2024-10-23T11:41Z) doesn't change anything with my problem.

Does anyone have an idea how to fix that? Sadly where is no update jet that fixed this issues.

 

[wisepds]

I'm in the same problem like you...
I have post this problem on Twiiter... when i update from original iso version to Linux 6.8.12-4-pve my PBS goes crazy... cpu at 100% and alwais Backup task even is in idle.

The problem comes from PROXMOX repositories... if you use only debian, no problem at all...

https://x.com/wisepds/status/1856398182948700605

(Only in spanish, sorry)...

 

[TwiX]

Hi,

Can confirm the issue for 50% of my PBSs

 

[wisepds]

Here a working workaround...

https://forum.proxmox.com/threads/pbs-goes-crazy-when-upgrading-to-version-6-8-12-4.157349/

 

[robinrm]

This workaround is great!

I've found the reason for this issue in the german support categorie. It seems like any kind of monitoring system which pings Port 8007 ist the problem. If you stop monitoring everything is fine.
Maybe this will be fixed in the next update, we'll see...

 

[wisepds]

Can you share how to stop this monitor?
Thanks!

 

[robinrm]

This depends on the service you're using for monitoring.

In my case i've stopped Uptimerobot because they ping ever 60seconds Port 8007.

Thats the Problem. Each ping will open a session on the server which normally closes automaticly but with the latest update thats not the case anymore.

So if you stop whatever service you're using to ping Port 8007 you've solved it temporaly.

 

[wisepds]

I use Uptime Kuma too.... WTF? I'll try!!!

 

[t.lamprecht]

This is being tracked under https://bugzilla.proxmox.com/show_bug.cgi?id=5868 and should now be fixed again with the proxmox-backup-server package in version 3.2.9-1 which has been uploaded to the pbstest repository a bit ago.

 

[AKEB]

I had the same problem with uptime kuma and checking tcp 8007
I removed the monitoring and everything started working fine

 

[wisepds]

This is being tracked under https://bugzilla.proxmox.com/show_bug.cgi?id=5868 and should now be fixed again with the proxmox-backup-server package in version 3.2.9-1 which has been uploaded to the pbstest repository a bit ago.

Do you know date of release for normal repository? For now it's 3.2.8-1 available.
Thanks!

 

[Pcom]

3.2.9 released for everybody. I'm testing it

 

[t.lamprecht]

Yeah, we move it to no-subscription today after positive feedback and no sight of regressions from other changes in that release.

The problematic version never hit the enterprise repo, so no need to move anything fast there.

 

[Pcom]

3.2.9 released for everybody. I'm testing it

I confirm my problem is solved with 3.2.9

 

[wisepds]

I confirm my problem is solved with 3.2.9

I confirm too... problem is solved! Thanks!