On Mon, Apr 13, 2009 at 10:12:27PM -0400, Perrin Harkins wrote: > > In your Image.pm module, do you declare a package name? Do you export > get_image_name()? How do you load this module in your new Mason > setup? > The entire Example/Image.pm:
----------8<---------->8---------- package Example::Image; use Image::ExifTool; use vars qw( %dirs %urls ); sub get_image_data ($;$$) { my $image_name = $_[0]; my $image_src = $_[1] ? $_[1] : $urls{'img'}; my $image_dir = $_[2] ? $_[2] : $dirs{'img'}; my %image_data; $image_data{'src'} = $image_src . $image_name; my $info = Image::ExifTool::ImageInfo($image_dir . $image_name); $image_data{'comment'} = $$info{'Comment'}; $image_data{'width'} = $$info{'ImageWidth'}; $image_data{'height'} = $$info{'ImageHeight'}; return %image_data; } 1; ----------8<---------->8---------- The module is loaded with a 'PerlModule Example::Image' directive. I have also tried without the PerlModule directive and with 'use' in the global.asa (for Apache::ASP) and the main component (for HTML::Mason). > > This has really been causing me to wrack my brain since > > it only occurs on the production server and not on the development > > server (both are identically configured). > > Are they really identically configured? Are you sure there isn't a > difference in startup.pl or in MaxRequestsPerChild or something like > that? > I have no startup.pl and the MaxRequestsPerChild setting is identical for both servers. The only differences that I found were mod_include and mod_disk_cache were enabled on the development server, but not on the production server. I have enabled them in production to see if that makes a difference. For the moment, the 500 errors have seemingly disappeared. However, they have on occasion disappeared only to return after a day or so. But for now it appears to be OK for a few tens of thousands of requests. Regards, -Roberto -- Roberto C. Sánchez http://people.connexer.com/~roberto http://www.connexer.com
signature.asc
Description: Digital signature