We've been seeing, with an alarming frequency, the following in our error log:
Apache2 IO write: (103) Software caused connection abort
This is typically connected to print statements in ContentGenerator/Problem.pm (please see below for the full pattern in the error log). I haven't been able to reproduce this--or, even to debrief anyone who has. But, given that it sometimes happens repeatedly with the same timestamp, it is not impossible that apache keeps trying until it succeeds--and that the end-user remains blissfully unaware (or so I hope). It is connected to the new perl/mod_perl we're using--we had never seen anything like it before. And, as much as it looks like a problem on the client side, I'm certain it is something in our mod_perl.
Mainly, I'm wondering if anyone else has seen this in their logs; if anyone knows what this looks like from the client's perspective; and, of course, if there's a fix.
Very many thanks,
(I couldn't compile a higher mod_perl under a higher perl or apache without test failures)
The full pattern in the error log:
[Thu Aug 31 14:02:17 2017] [error] [client 10.16.113.195] [/webwork2/course_name/set_id/problem_id/] :Apache2 IO write: (103) Software caused connection abort at /opt/webwork/webwork2/lib/WeBWorK/ContentGenerator/Problem.pm line 1209\n * in WeBWorK::ContentGenerator::Problem::output_form_start called at line 155 of /opt/webwork/webwork2/lib/WeBWorK/Template.pm\n * in WeBWorK::Template::template called at line 560 of /opt/webwork/webwork2/lib/WeBWorK/ContentGenerator.pm\n * in WeBWorK::ContentGenerator::content called at line 363 of /opt/webwork/webwork2/lib/WeBWorK/ContentGenerator/Problem.pm\n * in WeBWorK::ContentGenerator::Problem::content called at line 232 of /opt/webwork/webwork2/lib/WeBWorK/ContentGenerator.pm\n * in WeBWorK::ContentGenerator::go called at line 384 of /opt/webwork/webwork2/lib/WeBWorK.pm, referer: http://euclid.vaniercollege.qc.ca/webwork2/course_name/set_id/problem_id/