Joshua Slive <[EMAIL PROTECTED]> wrote:
On Tue, 27 Jul 2004, Dan Wilga wrote: > [Fri Jul 02 11:14:30 2004] [info] Initial (No.1) HTTPS request received for > child 0 (server [server name]:443) > [Fri Jul 02 11:14:30 2004] [info] (104)Connection reset by peer: > core_output_filter: writing data to the network > [Fri Jul 02 11:14:30 2004] [error] 5894: ModPerl::Registry: Software caused > connection abort at [path to script and line number].\n > [Fri Jul 02 11:14:30 2004] [error] 5894: ModPerl::Registry: Software caused > connection abort at [path to script and line number].\n
> If you have ideas about the possible cause of this problem, I can privately > email you an error log of the SSL session produced with "LogLevel debug". I > can also send you the environment variables at the time of the error.
I don't really understand what the problem is. The "Connection reset by peer" is a perfectly normal condition resulting from -- as you've already figured out -- the client going away in the middle of the response. The only thing that I see wrong here is that your mod_perl stuff is generating a message at loglevel "error" which is really not an error condition. You'd probably need to address that on the mod_perl list.
I have two concerns:
1. The superficial one, being the fact that my ErrorDocument 500 script is being called every time this happens. It normally sends me email every time a 500 occurs. Sure, I can (and already have) put in some code to ignore this particular condition, but it worries me that my script might now be failing to notify me of legitimate errors because of this.
2. The change in behavior between versions of Apache suggests that something fundamental changed in error reporting, and I am concerned that this might have other effects. I know that it's due to a change in Apache, because I did not change versions of mod_perl when upgrading Apache.
I will take your suggestion to the mod_perl list, to see if they agree that this condition should be re-labeled as "info"-level, instead of "error". However, I do think someone should reexamine the changes that were made between Apache 2.0.47 and 2.0.49, to see if there might be other, more negative fallout. I worry that changing the logging behavior might just be masking the symptom of a bigger problem.
--
Dan Wilga [EMAIL PROTECTED]
Web Technology Specialist http://www.mtholyoke.edu
Mount Holyoke College Tel: 413-538-3027
South Hadley, MA 01075 "Who left the cake out in the rain?"
