httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dean Gaudet <dgau...@arctic.org>
Subject Re: [SPLAT] Invalid Content-Length 49, 49, 49
Date Wed, 03 Dec 1997 02:56:09 GMT
Wouldn't it be nice if you could customize the error log to log the
uniqueid for the request causing the error?  Then you could join(1) this
with your hit log or your user agent log, which you could keep separate
from your hit log, and even do three way joins if you were so inclined.

Too bad mod_cgi needs a complete overhaul to stop it from blindly
attaching scripts' stderr to error_log.  It should really intervene and
prepend a uniqueid to each line of stderr.  But to do that requires it to
run a select() loop.  Something I keep wishing I had the time to implement
:)

Dean

On Wed, 3 Dec 1997, Rob Hartill wrote:

> 
> I recently added a 5 minute cron job to poll for unusual happenings,
> log them (via output that cron then mails me) before a kill -15 and
> complete restart of the server.
> 
> Here's one that looks like it's not mod_perl related (my usual achilles
> heel).
> 
> 
> ---------- Forwarded message ----------
> Date: Tue, 2 Dec 1997 17:30:35 -0600
> From: Cron Daemon <root@imdb.com>
> To: robh@imdb.com
> Subject: Cron <robh@us6> /home/robh/bin/emergency_httpd_restart
> 
> Error log was '38125221' bytes long. Emptying it and restarting httpd
> 
> [Tue Dec  2 17:25:26 1997] [error] Invalid Content-Length 49, 49, 49
> [Tue Dec  2 17:25:26 1997] [error] Invalid Content-Length 49, 49, 49
> [Tue Dec  2 17:25:26 1997] [error] Invalid Content-Length 49, 49, 49
> [Tue Dec  2 17:25:26 1997] [error] Invalid Content-Length 49, 49, 49
> [Tue Dec  2 17:25:26 1997] [error] Invalid Content-Length 49, 49, 49
> [Tue Dec  2 17:25:26 1997] [error] Invalid Content-Length 49, 49, 49
> [Tue Dec  2 17:25:26 1997] [error] Invalid Content-Length 49, 49, 49
> etc..
> 
> this came from a tail -100 error_log.
> 
> Now, I don't know how much of the 38mb was filled with this message
> but it was caught kinda red handed. In the past I've had problems with
> error_logs growing too big for the filesystem so my cron job is there
> to catch such problems.
> 
> The server is a 1.3b of a few weeks back, ~Nov 9th I think.
> 
> 
> 


Mime
View raw message