Make setting pcp_rc_dir in configure.in more robust
configure.in defines pcp_rc_dir by testing for the existence of the
directories /etc/rc.d/init.d /etc/init.d /sbin/init.d, and setting
pcp_rc_dir to the first one it finds.
We run SLES on our systems, and what I'll call a "rogue" RPM happened
to create /etc/rc.d/init.d and install an init script there on the
host I did my pcp build on. This makes chkconfig unhappy, as it
expects to find init scripts in /etc/init.d.
Arguably /etc/rc.d/init.d shouldn't be present on a SuSE machine, but
nonetheless, I think it would be a good idea to explicitly set
pcp_rc_dir to /etc/init.d for SuSE. Might be a good idea to do the
same for other distros as well. The following patch should do the
trick, though I've only been able to test on SLES 10 SP2.
Make setting pcp_rc_dir in configure.in more robust
configure.in defines pcp_rc_dir by testing for the existence of the
directories /etc/rc.d/init.d /etc/init.d /sbin/init.d, and setting
pcp_rc_dir to the first one it finds.
We run SLES on our systems, and what I'll call a "rogue" RPM happened
to create /etc/rc.d/init.d and install an init script there on the
host I did my pcp build on. This makes chkconfig unhappy, as it
expects to find init scripts in /etc/init.d.
Arguably /etc/rc.d/init.d shouldn't be present on a SuSE machine, but
nonetheless, I think it would be a good idea to explicitly set
pcp_rc_dir to /etc/init.d for SuSE. Might be a good idea to do the
same for other distros as well. The following patch should do the
trick, though I've only been able to test on SLES 10 SP2.