apache2 startup script fails to stop apache2 at times.

[Xeno Campanoli]

I am having a hard top with /etc/init.d/apache2 stops and restarts (due to the stop phase) from my Hardy Server. Typically after some usage involving mod_ruby (not clear that it's from ruby, or just from traffic), a restart is unable to stop any of the apache processes and I have to kill them by hand to proceed. Either apache needs to be fixed, or an alternative harder kill done as a matter of course in the script when this happens. Otherwise you spend all day killing apache to get things done. I have no experience with this outside of my Hardy Server.

[Chuck Short (zulcss)]

Please describe your apache setup. Unfortunately I dont have enough information from what you have provided.

Thanks
chuck

[Xeno Campanoli (xeno)]

Part of the problem here is you delete the PID file before you are sure you killed the thing. Why delete the pid file if you haven't checked? Keep it around until you're sure it's gone.

[Xeno Campanoli (xeno)]

And if you're trying to short circuit or obscure workings to make them more secure, please see Schneier on topics like "Obscurity is not Security" (_Secrets_And_Lies_, for instance). If you make the script run cleanly, it is more secure. If you make it so somebody can modify it, it is more of a real user product.

[Stefan Fritsch (sf-sfritsch)]

It is normal behavior that the apache processes stay around until they have finished processing the currently active requests (since we are using graceful-stop). Normally you should just be able to start apache again even before all processes are dead. Is this not the case for you? What error message do you get?

[Chuck Short (zulcss)]

This has been fixed for intrepid.

Thanks
chuck