MariaDB crashes after update on a Plesk server: mariadb.service: main process exited, code=exited, status=218/CAPABILITIES

kb: technical

Symptoms

After an update on a Plesk server with CentOS 7.9 and MariaDB 10.5, MariaDB is crashing and does not start.
Output of systemctl status mariadb shows the following:

CONFIG_TEXT: May 15 09:34:50 server.example.com systemd[1]: Starting MariaDB 10.5.25 database server...
May 15 09:34:51 server.example.com systemd[1]: mariadb.service: main process exited, code=exited, status=218/CAPABILITIES
May 15 09:34:51 server.example.com systemd[1]: Failed to start MariaDB 10.5.25 database server.
May 15 09:34:51 server.example.com systemd[1]: Unit mariadb.service entered failed state.
May 15 09:34:51 server.example.com systemd[1]: mariadb.service failed.
Yum log /var/log/yum.log shows a recent update:

CONFIG_TEXT: May 15 03:30:58 Updated: MariaDB-common-10.5.25-1.el7.centos.x86_64
May 15 03:30:58 Updated: MariaDB-compat-10.5.25-1.el7.centos.x86_64
May 15 03:31:02 Updated: MariaDB-client-10.5.25-1.el7.centos.x86_64
May 15 03:31:04 Updated: galera-4-26.4.18-1.el7.centos.x86_64
May 15 03:31:22 Updated: MariaDB-server-10.5.25-1.el7.centos.x86_64
May 15 03:31:22 Updated: MariaDB-shared-10.5.25-1.el7.centos.x86_64

Outdated kernel incompatibility with systemd unit functionality used by MariaDB starting from recent 10.5 update and above.

# [root@plesk my.cnf.d]# uname -a
Linux hostname 3.10.0-327.28.3.el7.x86_64 #1 SMP Thu Aug 18 19:05:49 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux

Alternatively, if the kernel cannot be updated at the present moment, a workaround can be used:

Comment out the affecting line:

Connect to the server via SSH
Create the folder /etc/systemd/system/mariadb.service.d/

# mkdir -p /etc/systemd/system/mariadb.service.d/ && cd /etc/systemd/system/mariadb.service.d/
Within that folder, create the file override.conf with your preferred text editor.
Add the following lines to the file:

MYSQL_LIN: [Service]
CapabilityBoundingSet=
AmbientCapabilities=
Save the file.
Reload the daemon:

# systemctl daemon-reload
Restart MariaDB:

# systemctl restart mariadb

Downgrade the MariaDB packages:

Check the last update list:

# yum history
Confirm the latest update with Action corresponding to I,U or Update that altered MariaDB packages:

# yum history info <ID>
Undo that transaction ID:

# yum history undo <ID>
Restart MariaDB

# systemctl restart mariadb