• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion
  • Please beaware of a breaking change in the REST API on the current Plesk release (18.0.62).
    Starting from Plesk Obsidian 18.0.62, requests to REST API containing the Content-Type header with a media-type directive other than “application/json” will result in the HTTP “415 Unsupported Media Type” client error response code. Read more here

Resolved Update 18.0.60 on AlmaLinux = high cpu steal value

rkbonatto

New Pleskian
Server operating system version
Almalinux 8
Plesk version and microupdate number
18.0.60
The recent update 18.0.60/61 caused the cpu steal value to increase sharply, causing slowdowns on websites and servers. The same behavior is repeated on all servers that I have AlmaLinux 8.10 (Cerulean Leopard).

The other servers with Centos7 that also received the 18.0.60/61 updates are not having any problems.

Take a look on 3 different servers with Plesk on AlmaLinux8, the problem starts with the update at 05/29

Uzfc3l5YiK.png
rbLp0AQ7Bs.png
sJt1ZI38Yj.png

Plesk on Centos7 with the same update but with no problem.
9VHMIhuqtA.png

Anyway to solve this?
 

Attachments

  • Uzfc3l5YiK.png
    Uzfc3l5YiK.png
    124.2 KB · Views: 1
It's hard to tell why there is such a high CPU load. The date it started is the date Almalinux 8.10 went live.
- Did you have a look at /var/log/messages ? Are there any related issues in there?
- Did you reboot the server after the upgrade so it runs on the new kernel?
- Any related issues shown in htop or glances?
 
Kernel version - 4.18.0-477.21.1.el8_8.x86_64
It is the same kernel in all servers with alma linux
This is strange because I restarted the server 3 or 4 times.

In logs there are no related issues.
 
Did you see anything in htop or glances? Are there processes that use a lot of CPU?

For the next time, I advise you to wait a couple of weeks before upgrading to a major update. This won't happen after the 8.10 release (it's the last one in the 8.x release) but for other servers, try not to upgrade immediately as soon as there is a new release.
 
Did you see anything in htop or glances? Are there processes that use a lot of CPU?

For the next time, I advise you to wait a couple of weeks before upgrading to a major update. This won't happen after the 8.10 release (it's the last one in the 8.x release) but for other servers, try not to upgrade immediately as soon as there is a new release.
I just updated to the latest kernel (4.18.0-553), let's see if the problem persists.
All servers are with automatic updates On
 
You should turn the automatic updates off. Keep an eye on the Plesk changelog. As soon as there is a new version, install the version before that using the command line. That way, you're always a release behind but that gives you the assurance there are no more major issues:
Code:
plesk installer --all-versions

This is my configuration on all servers:
Screenshot 2024-06-03 at 18.59.04.png
 
Take a look on glances, don't seem to be any processes that justify stealing the cpu, I updated the kernel to the recent version it is running, I also disabled docker but the problem persists.
glances.jpg
 
After a detailed analysis, it was possible to verify that the problem faced was occurring due to the AlmaLinux 8.10 update that is causing some type of bug with the Proxmox virtualizer, on other occasions I have experienced other bugs related to AlmaLinux and Proxmox, it turns out that Proxmox it does not update as frequently as Alma and ends up not receiving recent Alma kernels and packages properly.FOPKyuoFA5.png
As we can see now, the steal value has decreased by almost 100% compared to before.
Unfortunately, to solve the problem it was necessary to migrate the entire VM to another server.
 
Back
Top