I have a problem with a "protocol synchronization error":
We use an exchange server internally and an external exim server as relay
and spam and virus filter. Usually this works fine. We have pipelining
generally disabled and enforce protocol synchronization to fight spam. This
would not be necessary for our exchange but it also doesn't hurt. Until
today!
Our exchange wanted to send a bounce. For some reason it seems to have used
pipelining - what never occurred before and I am not even sure, it really
did. Exim was expected to reject that and close the connection - which would
have ment that the outgoing bounce would stay in the queue of our exchange
server.
Yet there was a big problem: Exim DID detect the pipelining AND closed the
connection BUT sent out that bounce ANYWAY.
Unfortunally five minutes later exchange tried again. So every five minutes
a bounce was sent. Over night starting at 6pm yesterday. Our project partner
was a litte "unpleased".
No to the question:
WHY is exim sending out a message if it dropped that connection?
How can I prevent this in the future? (Note that this never happened in the
last 3 years - with basically the same setup.)
Our anonymized log is:
2010-06-02 08:33:29 1OJhW4-000368-JI <= <> H=xxx (xxx) [134.147.xxx.xxx]
P=esmtp S=661369 id=xxx
2010-06-02 08:33:29 SMTP protocol synchronization error (next input sent too
soon: pipelining was not advertised): rejected "" H=xxx (xxx)
[134.147.xxx.xxx] next input="\nJM\025\374\204\277\260\241<\257,\017xxxx"
2010-06-02 08:33:30 1OJhW4-000368-JI => xxx@xxx R=remote T=remote H=yyy
[192.134.yyy.yy]
2010-06-02 08:33:30 1OJhW4-000368-JI Completed
