Service stop/start/restart/reload failed when you kill a process manually. You have to remove the daemon pid by hand. Solution: add a sort of "rm $PIDFILE" in the routine. This affects also all runscript in /etc/init.d Regards ph
If you nuke a process/service manually, you have to "zap" it to get runscript to restart it: /etc/init.d/<service> zap /etc/init.d/<service> start There are rare cases where the pid file does need to be deleted before a service can start again, even after zapping. The only service I've seen do this repeatedly is openssh, so for that case, a separate bug would need to be filed. if you come across any other packages with this behavior after zapping, please file separate bugs for them. Resolving as INVALID.