For some reason apparantly something in /sbin/runscript.sh doesn't account for a process that doesn't start up properly. For example, when I setup bind9 it seemed to work fine - but when I imported my old configuration and shut it down and restarted it, named encountered a problem because it didn't have permission to read my named.conf file - so it exited with a fatal error (and wrote to /var/log/daemon.log). However, /etc/init.d/named status still reported it running, and attempting to do /etc/init.d/named stop would always fail: init.d # ./named stop * Stopping named... [ !! ] And then curiously: init.d # ./named start * WARNING: "named" has already been started. This happens even when no named.pid exists, which I'd think should be the mechanism for detecting whether or not it's running. The only way I got around this was by manually removing the "named" symlink in /mnt/init.d/started and then restarting (wth?).
The same thing happened to me with apache 1.3.27... Que
`/etc/init.d/named zap`
It doesn't help that /etc/init.d/named doesn't say anything about zap. Where is "zap" documented that you'd know it does that? Maybe this should be moved to a documentation bug. :)
run any init.d script w/out parameters
Ah, geez, there it is. I guess I never thought of doing that as its only printed start/stop/restart/reload on my other systems. thx.
