httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Oliver Jones <>
Subject Weird problem with my logs.
Date Fri, 12 Jul 2002 11:54:47 GMT
Today I noticed that two of my Vhosts had stopped logging into their
access_log files.  I tried restarting the server but this didn't fix the
problem.  So I moved the files out of the way and restarted the server
so it created new log files.  Once I did this the server started to log
requests into the new files.

I also noticed that the modified time on the files wasn't changing like
it normally does as requests are logged.

Hmm.  Something else I just noticed.  My main /var/log/httpd/error_log
was full of these errors:

[Fri Jul 12 23:08:01 2002] [notice] child pid 18169 exit signal File
size limit exceeded (25)

It appears this is the problem.  I am running with Mod_perl 1.27 and
Apache::Resource to keep processes from expanding beyond my boxes
limitations.  Ie, to stop dinial of service attacks etc.

It appears that these errors started cropping up at the same time that
one of my most active vhosts stopped logging so it certainly looks like
the source of the problem.

I've now commented out my PerlSetEnv PERL_RLIMIT_FSIZE config var.  The
weird thing is that up until this incident I've been running just fine
and my log files have gotten to be 135mb in size.

One thing that I did do recenly was upgrade from mod_perl 1.26 to 1.27. 
I did this when I rebuilt my Apache RPM to patch the chunked encoding
security issue. From my error log it appears that the file size exceeded
errors popped up within minutes of my first restarting apache with
mod_perl 1.27.

Sigh.  I guess this is just another lesson in why not to upgrade server
daemons on a production box.

If anyone can shed anymore light on this I'd love to hear from them.



Oliver Jones > Software Engineer >
Deeper Design Limited. 

View raw message