Steve Kemp wrote:
This sequence can repeat multiple times though I cannot seem to find why the RSET is being issued by the client side. At each step my plugins are returning the correct responses (e.g. "250 QUEUED" after the DATA phase). Of course the MS-Exchange admin believes their server is working correctly and nobody else is reporting problems.
Historically, I have seen this behavior if the time between steps 4 and 5 was significant (for some value of significant). What was happening was that the sending MTA would timeout before the receiving MTA would send the 250. However, I've never seen an instance like you describe where the sending MTA would reuse the connection in that case. Do you have log entries with timestamps?
Of course the MS-Exchange admins won't acknowledge their server is acting erroneously, despite evidence to the contrary. It would make their tiny little minds explode to realize that the very expensive server they are supporting is crap... ;-)
John
