Red Hat Bugzilla – Bug 91943
Postgresql init script system boot failure
Last modified: 2007-04-18 12:54:09 EDT
The postgresql init script is trying to figure out how it is called as and
therefore at the startup does somethin like:
Based on that $NAME then it tries to read the /etc/sysconfig/pgsql/$NAME
configuration file, if it exists.
There are 2 problems that keep this from working right:
- first, during the system bootup time, the script is called as S85postgresql,
which means that $NAME is assigned 'S85postgresql' instead of the desired
'postgresql'. This works correctly when doing a "system postgresql restart", but
fails to read the postgresql config file during the first boot.
- secoindly, in the config file one could define the PGOPTS variable with
additional options to be passed to the postgres engine by the postmaster daemon.
While that variable is exported from the init script, it is used nowhere in the
start() function, therefore rendering it useless.
The init script is currently in the process of being overhauled. I've never
really seen the point of the basename ilne anyways, but that is an easy fix.
Usage of PGOPTS can be integrated with the coming enhancements. Re-assigning to
fnasser who is working on the init script.
These two issues are fixed in RAWHIDE.
*** Bug 97605 has been marked as a duplicate of this bug. ***