if httpd-perl starts by itself, it binds to :80; subsequent attempt to start vanilla httpd will tell us "OK" but not start it. well httpd-perl *will* serve the requests, but the message is confusing: "service httpd status" will tell it's down, and /etc/httpd/conf/httpd.conf changes won't do any difference. should we check httpd-perl running in /etc/initd.d/httpd::start() ? should we then do something to httpd-perl and restart it after httpd start? or notify the administrator and fail?
done like this: start() { if [ "$1" != "again" ]; then if status --pidfile "$PERLPIDFILE" --expect-user root \ --expect-user root -- httpd-perl >&/dev/null; then echo "Warning: httpd-perl is already running, check 'service httpd-perl extendedstatus'" passed "httpd-perl already running" fi fi start_daemon --pidfile "$PIDFILE" --lockfile "$LOCKFILE" \ --expect-user root --name libhttpd.ep -- httpd RETVAL=$? return $RETVAL } stop() { stop_daemon --pidfile "$PIDFILE" --lockfile "$LOCKFILE" \ --expect-user root --name libhttpd.ep -- httpd if [ "$1" != "nocheck" -a -f "$PERLPIDFILE" ]; then echo "Warning: httpd-perl is running, check 'service httpd-perl extendedstatus'" passed "httpd-perl already running" fi RETVAL=$? return $RETVAL } restart() { stop nocheck conftest || exit $? start again } First implementation (involving e.g. httpd-perl stop when httpd is stopped and so on) is now considered as too non-modifiable by local administrator who can e.g. port-separate the instances and thus not need the chain restarts.
* Wed Sep 03 2003 Michael Shigorin <mike@altlinux.ru> 1.3.28rusPL30.18-alt2 - httpd.init.Sisyphus now checks for some situations involving httpd-perl when silent breakage could occur and issues warnings (#2928)
rolled back in 1.3.33-alt3 (along with something like a fix, see #4994)
closing
