Alan Thanks for taking an interest - I have to admit, although I have been in the business for some years, I am outside my comfort zone when working with PERL - I was born & bred a windows boy :-)
Forgive me then if I sometimes ask the bloomin obvious trying to make things fit my frame of reference. Firstly, yes I get <gl: <STARTTLS in the debug logs when I have SSL enabled - the problem pretty much goes away when I have it turned off, and I need to allow business mails through - since I can't replicate the fault I am having to observe it 'in the wild' which means I have to turn SSL off for enough of the time that the retries go through. However - with SSL _OFF_ I have one connection that still fails and times out - it is pretty spammy, but is targeted. However I have allTestMode on so it should get through but it exhibits the <stateReset> (is that some kind of watchdog triggering?) I have captured a section of that log, I tried uploading it, but it is stuck in moderation - if you think it will be eye strain for no use then ignore it - I will try & capture a transaction while SSL is enabled which will be a more relavant test Kevin -----Original Message----- From: Alan McNeil [mailto:a9...@mac.com] Sent: 08 April 2009 03:44 To: ASSP development mailing list Subject: Re: [Assp-test] 1.5.1.2 - Many Connection Idle - timeout Kevin, I'm running on Linux (Ubuntu Hardy) with perl v5.8.8 built for i486- linux-gnu-thread-multi, IO::SOCKET::SSL v1.24, Net::SSLeay 1.35 so our environments are quite different. Do you still have a log with a <stateReset> in it? I'd love to see a log and try to puzzle out what was going on in the code. Do you ever see <gl: <STARTTLS... in the log? That <gl: is a getline subroutine and should be hit. The way this bug is behaving smells like an uninitialized variable or pointer reads random data that is different on different machines. It's a lot harder to make that mistake in perl but maybe one of the modules has a problem. I hope your eyes are OK. Mine are bloodshot and asking for shut-eye time. Log reading isn't fun. Alan > However - I have been running more tests, and I have tracked one > failed > connection, it proceeded in what looked to my unpracticed eye good > order > until it did a state reset - I am loath to clutter the list with log > dumps, but the last couple of lines might show what I mean by that > > 7 Apr 09 13:12:07 <IO::Socket::SSL=GLOB(0x7054984) HASH(0x70c3d14) > l=269> > 7 Apr 09 13:12:07 <wrote: (269)<long text>> > 7 Apr 09 13:12:07 <6> > 7 Apr 09 13:12:07 <50> > 7 Apr 09 13:12:07 <8> > 7 Apr 09 13:12:07 <2> > 7 Apr 09 13:12:07 <doing <MAIL FROM:<ag...@ukex12.pur3.net> > SIZE=37972>> > 7 Apr 09 13:12:07 <15> > 7 Apr 09 13:12:07 <gl: <MAIL FROM:<ag...@ukex12.pur3.net> SIZE=37972>> > 7 Apr 09 13:12:07 <stateReset> > > I am running ASSP Version: 1.5.1.2(0.0.03) on Windows with active > state > PERL 5.10.0 IO::SOCKET::SSL is at version 1.22 which is the latest I > could find in a repository shown as working with Windows. I am > starting > to loose the ability see for staring at log files so any help would be > gratefully received - I'll even let you tell me how daft I am if it > points me closer to the fix > > Thanks > > Kevin ------------------------------------------------------------------------ ------ This SF.net email is sponsored by: High Quality Requirements in a Collaborative Environment. Download a free trial of Rational Requirements Composer Now! http://p.sf.net/sfu/www-ibm-com _______________________________________________ Assp-test mailing list Assp-test@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/assp-test ------------------------------------------------------------------------------ This SF.net email is sponsored by: High Quality Requirements in a Collaborative Environment. Download a free trial of Rational Requirements Composer Now! http://p.sf.net/sfu/www-ibm-com _______________________________________________ Assp-test mailing list Assp-test@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/assp-test
