the file /etc/opt/rh/rh-mysql56/my.cnf.d/mysql-server.cnf specifies log-error=/var/opt/rh/rh-mysql56/log/mysql/mysqld.log this is not a problem in default configuration, as the file gets created on service startup however, if you change log path to something else, e.g. like [mysqld] general_log=1 general_log_file=/var/opt/rh/rh-mysql56/log/mysql/mysqld.fifo and delete the file 'mysqld.log', then service startup then fails as the general_log_file path is checked/gets created if needed and not the error-log path this is regression from previous version, as both files were touched before running mysqld_safe, and this was dropped from scripts, leaving it upon mysqld_safe
Created attachment 1314225 [details] proposed patch There is one line in mysqld_safe that was not converted into SCL-ized path, which seems like causing these troubles.
ahem, this is in RHSA-2017:28374 which is 2.4.z but has target 3.0 ... trying to correct note that also 3.0 branches commits are referred but it got fixed in the new 2.4 build
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2017:2787