msmilter connect timeouts.

[Mark Sapiro]

I have installed MailScanner 5.1.2-2 and am running with msmilter.
Sometimes the connect from postfix to the milter times out after 30
seconds. If thin occurs on a message reinjected by the milter, other
than the delay, it is not a problem, but if it occurs on an incoming
message, MailScanner is bypassed for that message.

Here are some relevant log entries for one such message.

Oct 31 09:05:53 sbh16 postfix/smtpd[10899]: connect from
fg90.mta.exacttarget.com[13.111.54.90]

Oct 31 09:06:23 sbh16 postfix/smtpd[10899]: warning: milter
inet:127.0.0.1:33333: can't read SMFIC_OPTNEG reply packet header:
Connection timed out

Oct 31 09:06:23 sbh16 postfix/smtpd[10899]: warning: milter
inet:127.0.0.1:33333: read error in initial handshake

Followed by the normal postfix messages, but the milter is bypassed so
MailScanner is never invoked for this message.

This does not always occur. It occured 13 times so far on Oct 31. In the
same period, there are 1159 'status=sent' log entries, so it seems to
affect roughly 1% of messages.

I haven't looked at the code so I have no idea why it might occasionaly
time out, but perhaps it is RBL lookups or similar. In any case it might
be good to put a shorter timeout on such things within the milter itself
so MailScanner can be invoked.

-- 
Mark Sapiro <mark at msapiro.net>        The highway is for gamblers,
San Francisco Bay Area, California    better use your sense - B. Dylan