Broken pipe and timeout waiting for input from [hostname] during client greeting

Hi. I had a fairly old sendmail running on a cobalt qube (running
Linux 2.2.) and about a month
ago, I upgraded to the latest 8.13.8. I wanted to run milter-greylist
with it and did all the setting
up and recomping of both the milter and sendmail.

I think I have been rather thorough with the complying with the
documentation and FAQs etc, but
what I have noticed is that I now am getting lots of "Broken pipes" and
timeouts.

I have uncommented all the sendmail.cf timeout values so that they are
defined (as suggested)
but it does not seem to be sufficient.

I am sure I am missing something, but because I have been staring at
the docs, FAQs etc for
so long, I am not seeing the solution.

I tried 8.14.0-beta4 and the errors are lesser, but still happening.
So, I think I am on the right
track. Please suggest what else I should look for.

Thanks.

Harish

Re: Broken pipe

On Sat, 29 Mar 2008 07:09:44 +0100, Peter Boosten wrote:

> Once in a while I receive an email that stays in the queue of my relay
> for days, and cannot be delivered to the internal Exchange server. Mailq
> reports 'broken pipe', and every morning I receive these kind of
> messages:
....
> Mar 29 03:24:38 validmst04 sendmail[89198]: m2PFSVuS078681: 10: fl=0x6,
> mode=140444: SOCK (Connection reset by valaps04.valid.loc.)->(Socket is
> not connected)
....
> Is this fixable, and if so: how?

Fix the "Exchange server".

Re: Broken pipe

Greg Russell wrote:
> On Sat, 29 Mar 2008 07:09:44 +0100, Peter Boosten wrote:
>
>> Once in a while I receive an email that stays in the queue of my relay
>> for days, and cannot be delivered to the internal Exchange server. Mailq
>> reports 'broken pipe', and every morning I receive these kind of
>> messages:
> ...
>> Mar 29 03:24:38 validmst04 sendmail[89198]: m2PFSVuS078681: 10: fl=0x6,
>> mode=140444: SOCK (Connection reset by valaps04.valid.loc.)->(Socket is
>> not connected)
> ...
>> Is this fixable, and if so: how?
>
> Fix the "Exchange server".

Greg, thanks for your answer, although I tast some cynism in your answer.
Why would that Exchange server be doing this, and especially for some, by far
not all emails?

TIA

Peter

--
http://www.boosten.org

Mail: peter at boosten dot org