The following reply was made to PR mod_log-any/1396; it has been noted by GNATS.
From: Marc Slemko <[EMAIL PROTECTED]>
To: Steve Resnick <[EMAIL PROTECTED]>
Cc: Apache bugs database <[EMAIL PROTECTED]>
Subject: Re: mod_log-any/1396: Logging to a pipe causes server to stop serving
documents
Date: Sat, 15 Nov 1997 23:32:52 -0700 (MST)
Does this problem always happen right away after starting the
server?
How many copies of your process does ps show running?
If you try the simplest possible setup (eg. a perl script that
does something like:
#!/usr/local/bin/perl
open (FILE, ">>/tmp/log.out") || die "blarg";
while (<>) {
print FILE;
}
) does it still have problems? We need to eliminate your code as
being the problem.
If the above is still a problem, how about when you don't have any
virtual hosts?
What state does ps show your logging process to be in at that time?
If you compile it with debugging and try running gdb on it after
it starts having problems, does it show it blocking on anything?
You may want to try 1.3b3 when it comes out, because it has better piped
logging support. I suspect, however, that there is something else
going on here.
Do you have any other logfiles for each vhost? Grab lsof from somewhere
and see how many file descriptors each httpd process has open.
