The following reply was made to PR os-linux/303; it has been noted by GNATS.
From: Marc Slemko <[EMAIL PROTECTED]> To: Randy Carpenter <[EMAIL PROTECTED]> Subject: Re: os-linux/303: apache server will not die when 'kill'ed Date: Sun, 6 Apr 1997 00:10:55 -0700 (MST) On Sun, 6 Apr 1997, Randy Carpenter wrote: > On Sat, 5 Apr 1997, Marc Slemko wrote: > > > Are you sure it is an actual process? Can you see it in a ps? Is it the > > parent or a child process? > > Yes it is a real process. > > > How are you killing the server? > > well, at first by using the rc. script that starts and stops, but it > leaves the one process. It cannot even be 'kill -9'ed I have heard that > a newer kernel with some patches helps this problem... I would suggest you try that; it looks like it may be a Linux kernel issue since I have heard no other reports of Apache doing this and the behavior you are reporting sounds like kernel trouble. Let us know if you find that a kernel change works when you get time for it. [...] > > Does it cause problems when you send a HUP to it? (I assume stopping and > > restarting referrs to actually shutting it down completely) > > Havent tried HUPing the actual server... we need to be abel to shut it > down, and bring it back up automatically for the purpose of safely > calculating webstats, and rotating logs You can do log rotations by HUPing the parent; move the log file to another file on the same partition, HUP the server, wait a little, then it will start writing to the new file. However, HUPing it may result in the same problem.
