View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0015120||CentOS-7||mariadb||public||2018-07-30 19:44||2018-07-30 20:20|
|Priority||normal||Severity||major||Reproducibility||have not tried|
|Platform||x84_64||OS||CentOS 7||OS Version||7.5.1804|
|Target Version||Fixed in Version|
|Summary||0015120: systemd-tmpfiles --create mariadb.conf is not run by MariaDB 5.5.56-2 package install process|
|Description||Installed MariaDB 5.5.56-2 as part of a LAMP server setup. Initially, MariaDB was running fine, but after a restart I noticed that MariaDB was not running and was failing to start.|
Eventually found out this was due to /var/run/mariadb not existing, even though other /var/run/ folders were created.
Also found that /user/lib/tmpfiles.d/mariadb.conf existed and looked fine.
Verified that everything other entry in /usr/lib/tmpfiles.d/ was being executed correctly...
Since I have no way of checking existing entries created with the "systemd-tmpfiles --create" command, I had to assume that the package installer had a bug similiar to https://bbs.archlinux.org/viewtopic.php?id=160302 where it either didn't run the command, or had a typo.
Sure enough, after running "systemd-tmpfiles --create mariadb.conf", everything is working fine -- MariaDB now starts correctly after a reboot.
|Steps To Reproduce||1) Normal installation of MariaDB 5.5.56-2 on CentOS 7.5.1804|
2) Restart system
3) MariaDB will not be running and will fail to start.
Creating /var/run/mariadb with the correct permissions temporarily and will allow it to start until next system reboot.
|Tags||No tags attached.|
Looks like I spoke too soon. I performed another reboot and the /var/run/mariadb folder was missing again.
Running "systemd-tmpfiles --create mariadb.conf" does fix it, and it does appear to be a problem with MariaDB and systemd-tmpfiles but I'm unclear on the permanent fix.