On Jan 23, 10:39 am, [EMAIL PROTECTED] (David Nicol) wrote: > On 1/23/07, Hanno Hecker <[EMAIL PROTECTED]> wrote: > > > Yes, had that too.. not that many connections, but I've seen running > > connections that long. It turned out that some clients where sitting > > there and just issuing NOOP every 30s without disconnecting. If you > > can, log with LOGDEBUG and see if it's the case for you, too... > > alternate solution: pick the connection details and use tcpdump or > > similar tools to see what's happening in this connection. > > > Hanno > > a max_noop would not be a bad thing -- after five or six consecutive noops, > we trigger the trapdoor
I have created a bug entry for SMESERVER which uses qpsmtpd. I think the bug is related to this issue and I have attached a logfile extract and a wireshark trace. The trace shows no evidence of NOOPs, just a bad connection. It seems to me 1200s is an excessive timeout for this situation. I would appreciate suggestions as to how to adjust this timeout. The bug entry is: http://contribs.org/bugzilla/show_bug.cgi?id=2578 Anthony
