Comment # 14 on bug 1032027 from Ales Seifert

We just had the same issue two days in the row on our DB server which freezed
in the middle of business hours for 10-15 minutes. System was not absolutely
frozen but even ssh login took a 30 sec before succeeded. Executing anything
from shell was taking a long delay before able to run. The "top" command
reveled 100% CPU utilization on btrfs-cleaner kernel thread. I was able
eventually run "btrfs quota disable /" and system started to be responsive
again after a while.

It was in the middle of the day and as a database server running MariaDB and
PostgreSQL it is not acceptable for the server to stop responding and do some
btfrs cleanup even for few seconds nor 15 minutes IMHO.

Server: Leap15.2 up to date, 32GB RAM, SSD 512GB, 2x8core XEON, FS 20% full

We are using timeline and number snapshots on /backups subvolume and number
snapshost (created by zypper) on /

Is there any way I can help to resolve this issue?