Yes, that time out setting is still definitely the default. Snoop is basically Solaris' native version of tcp dump, and shouldn't be putting the packets out of order when capturing. Maybe the client is making more than one connection attempt? The part that made me suspect my server is that the same message, from the very same client, is processed just fine when CC'd to a different account (gmail, et al). Two test messages from the client side have yet to make it over server-side. Then, a third test message from the client side was deferred for six hours before it was finally delivered.
On 03/10/2011 01:56 AM, Victor Duchovni wrote: > On Thu, Mar 10, 2011 at 01:14:48AM +0100, Jeroen Geilman wrote: > > >>> Mar 9 18:38:03 pmx4 postfix/smtpd[13358]: [ID 197553 mail.info] >>> connect >>> from unknown[134.53.6.74] >>> >>> >> okay >> >> >>> Mar 9 18:41:03 pmx4 postfix/smtpd[13243]: [ID 197553 mail.info]> >>> unknown[134.53.6.74]: 421 4.4.2 smtp-in.montclair.edu Error: timeout >>> exceeded >>> >> That's 3 minutes (180 seconds); any particular reason you changed it >> from >> the default of 300 seconds ? >> > No, these are different smtpd(8) processes, and unrelated connections. > > Ouch, indeed. So this client is making connections in rapid succession - and failing ? -- J.
