Resolved High memory-cached usage and CPU stuck during file backup

[jescayola]

Hello everyone!
First time here (not with Plesk), please don't be to hard and let me know if I'm missing something essential to help with this case so you don't waste time looking into it.
We had a VPS upgrade (number of CPUs and RAM) and since that upgrade the Plesk Backup can't finish because the CPUs stuck and there seems to be a high memory-cached value from Plesk. Here are some screenshots attached.
Before the upgrade we had no problem with the backups.
BBDD and email Backup seem to work just fine.
We are running Plesk Obsidian 18.0.58 update 2 on Debian 10.13

Thank you for your time!

 

[Peter Debik]

The soft lockup is likely caused by a general hardware issue of your system OR by some software, but it is not necessarily linked to a Plesk component. Last time I saw it the cause was a Linux Kernel that did not perfectly match to the other apps that were long updated (old Kernel, new apps), and simply updating the Kernel helped.

The RAM usage is o.k. Even if it reaches 100% - as long as that is cached RAM.

Are you really on Plesk Onyx? What operating system are you using?

 

[jescayola]

Hola Peter!
Thank you very much for taking the time and providing valuable info about this issue!
About the SO and Plesk version, here is from the dashboard and CLI:



Let me know if I can provide more info or perform any test!

 

[jescayola]

Update!
I performed a dist upgrade following your lead and the backup worked!!! It reached again high levels of memory-cached and CPU (a few warnings) but the server survived:



If you need more details to help other cases please let me know!
Thank you!

 

[Peter Debik]

High memory usage is fine. That's what memory is for. High cpu usage is typically caused by the backup compression. On your "Backup Settings" page you can make some changes to reduce the cpu load, e.g. use a compression level "normal", "fast" or "fastest" (normal compression, less compression, only a tiny little bit of compression etc.). You can also reduce the priority of the backup job. For example change the priority to a number 12, 13 or 14 (which is a quite low priority on Linux, the higher the number, the lower the priority), also change the IO priority to 7 (which is the lowest priority possible with 7 = lowest, 0 = highest). This will give other processes on the server more time slices so that even in a high cpu load situation, other process will continue to run swiftly.