Well okay I didn't know the config files were being split up. There are two options of the top of my head:
1) As you said. Just assume the pid file location 2) Check the parent process id being 1 In this case parent of the parent because of mysqld_safe # test $(ps -o ppid= -p $(ps -o ppid= -p $PID)) -eq 1 This would work in most cases I can think of. mysqld run by a user or a container would not be started by the init. But seems like a rather complex solution to a fairly simple problem. January 26 2016 9:18 AM, "Otto Kekäläinen" <[email protected]> wrote: > Thanks Lennart, the patch is much nicer to read now. > > It seems to rely on the fact that it should find the line 'pid-file = > /var/run/mysqld/mysqld.pid' in the file /etc/mysql/my.cnf > > However, since the new mysql/mariadb config decoupling effort (driven > by Ubuntu developers) the file /etc/mysql/my.cnf is no longer the main > config file itself, and does not contain the pid line. > > The actual line is now found in: > grep pid /etc/mysql/mariadb.conf.d/* > /etc/mysql/mariadb.conf.d/50-server.cnf:pid-file = > /var/run/mysqld/mysqld.pid > > Ironically simply assuming the exact location > /var/run/mysqld/mysqld.pid would be more reliable than grepping the > configs :) > > I am glad the fix works in your situation in Debian Jessie. > Unfortunately we need to think a bit more to come up with a universal > solution.. I am happy to review any alternative solutions/patches > anybody posts to this issue.
