Hi-

I'm running sendmail 8.13 and using a set of dnsbl entries to block
spam. I've enabled delay_checks so that check_mail and check_relay are
called on RCPT TO.

I've recently noticed that hotmail.com servers disconnect abruptly
after the RCPT TO, even though I didn't issue a 550/reject. I suspect
that they are timing out while I'm doing the dnsbl lookups. When I add
an entry to whhitelist hotmail.com in /etc/access (so that the dnsbl
tests are bypassed) I receive mail properly.

Does anyone have ideas as to a better fix than whitelisting
hotmail.com? Maybe a way to set the maximum time I will wait for a
dnsbl server to respond (so I time out slow servers?)? Any idea what
their timeout is set to?


The error I get is:

>Jan 18 06:48:16 memnoch sendmail[16223]: l0IBkPm5016223: lost input
>channel from bay0-omc1-s4.bay0.hotmail.com [65.54.246.76] to MTA
>after rcpt
>
>Jan 18 06:48:16 memnoch sendmail[16223]: l0IBkPm5016223:
>from=<xxx@msn.com>, size=0, class=0, nrcpts=1, proto=ESMTP,
>daemon=MTA, relay=bay0-omc1-s4.bay0.hotmail.com [65.54.246.76]

Thanks,

Lou

climberlou@gmail.com writes:
> I'm running sendmail 8.13 and using a set of dnsbl entries to block
> spam. I've enabled delay_checks so that check_mail and check_relay are
> called on RCPT TO.
>
> I've recently noticed that hotmail.com servers disconnect abruptly
> after the RCPT TO, even though I didn't issue a 550/reject. I suspect
> that they are timing out while I'm doing the dnsbl lookups. When I add
> an entry to whhitelist hotmail.com in /etc/access (so that the dnsbl
> tests are bypassed) I receive mail properly.
>
> Does anyone have ideas as to a better fix than whitelisting
> hotmail.com? Maybe a way to set the maximum time I will wait for a
> dnsbl server to respond (so I time out slow servers?)? Any idea what
> their timeout is set to?
>
>
> The error I get is:
>
>>Jan 18 06:48:16 memnoch sendmail[16223]: l0IBkPm5016223: lost input
>>channel from bay0-omc1-s4.bay0.hotmail.com [65.54.246.76] to MTA
>>after rcpt
>>
>>Jan 18 06:48:16 memnoch sendmail[16223]: l0IBkPm5016223:
>>from=<xxx@msn.com>, size=0, class=0, nrcpts=1, proto=ESMTP,
>>daemon=MTA, relay=bay0-omc1-s4.bay0.hotmail.com [65.54.246.76]

Simple startup question for would be optimization:
Which DNSBL do you use?
Do you use local copies of any of the above?

--
[pl>en: Andrew] Andrzej Adam Filip : anfi@priv.onet.pl : anfi@xl.wp.pl
Before You Ask: http://anfi.homeunix.net/sendmail/B4UAsk-Sendmail.html
http://anfi.homeunix.net/sendmail/

Andrzej Adam Filip wrote:
> climberlou@gmail.com writes:
> > I'm running sendmail 8.13 and using a set of dnsbl entries to block
> > spam. I've enabled delay_checks so that check_mail and check_relay are
> > called on RCPT TO.
> >
> > I've recently noticed that hotmail.com servers disconnect abruptly
> > after the RCPT TO, even though I didn't issue a 550/reject. I suspect
> > that they are timing out while I'm doing the dnsbl lookups. When I add
> > an entry to whhitelist hotmail.com in /etc/access (so that the dnsbl
> > tests are bypassed) I receive mail properly.
> >
> > Does anyone have ideas as to a better fix than whitelisting
> > hotmail.com? Maybe a way to set the maximum time I will wait for a
> > dnsbl server to respond (so I time out slow servers?)? Any idea what
> > their timeout is set to?
> >
> >
> > The error I get is:
> >
> >>Jan 18 06:48:16 memnoch sendmail[16223]: l0IBkPm5016223: lost input
> >>channel from bay0-omc1-s4.bay0.hotmail.com [65.54.246.76] to MTA
> >>after rcpt
> >>
> >>Jan 18 06:48:16 memnoch sendmail[16223]: l0IBkPm5016223:
> >>from=<xxx@msn.com>, size=0, class=0, nrcpts=1, proto=ESMTP,
> >>daemon=MTA, relay=bay0-omc1-s4.bay0.hotmail.com [65.54.246.76]
>
> Simple startup question for would be optimization:
> Which DNSBL do you use?
> Do you use local copies of any of the above?
>

Thanks for the quick reply. I use a list, which I reorder and track
performance of from time to time. I just went back and noticed that I
was querying ordb (off the air). Removing it seems to have fixed the
problem.

I'm surprised that msn/hotmail is that sensitive to delays...haven't
seen this problem elsewhere.

In any case, thanks again. And for completeness, my dnsbl list is
currently below, with stats on spam blocked for the last 5.5 days:

log start: Jan 14 04:03:31
log end: Jan 19 14:25:57

spamhaus: 4472
dsbl: 205
spamcop: 650
njabl: 147
rfc-ignorant: 0
--------------------------
total spam rejected = 5474
total dial rejected = 2931
----------
total rejected= 8405

climberlou@gmail.com writes:

> Andrzej Adam Filip wrote:
>> climberlou@gmail.com writes:
>> > I'm running sendmail 8.13 and using a set of dnsbl entries to block
>> > spam. I've enabled delay_checks so that check_mail and check_relay are
>> > called on RCPT TO.
>> >
>> > I've recently noticed that hotmail.com servers disconnect abruptly
>> > after the RCPT TO, even though I didn't issue a 550/reject. I suspect
>> > that they are timing out while I'm doing the dnsbl lookups. When I add
>> > an entry to whhitelist hotmail.com in /etc/access (so that the dnsbl
>> > tests are bypassed) I receive mail properly.
>> >
>> > Does anyone have ideas as to a better fix than whitelisting
>> > hotmail.com? Maybe a way to set the maximum time I will wait for a
>> > dnsbl server to respond (so I time out slow servers?)? Any idea what
>> > their timeout is set to?
>> >
>> >
>> > The error I get is:
>> >
>> >>Jan 18 06:48:16 memnoch sendmail[16223]: l0IBkPm5016223: lost input
>> >>channel from bay0-omc1-s4.bay0.hotmail.com [65.54.246.76] to MTA
>> >>after rcpt
>> >>
>> >>Jan 18 06:48:16 memnoch sendmail[16223]: l0IBkPm5016223:
>> >>from=<xxx@msn.com>, size=0, class=0, nrcpts=1, proto=ESMTP,
>> >>daemon=MTA, relay=bay0-omc1-s4.bay0.hotmail.com [65.54.246.76]
>>
>> Simple startup question for would be optimization:
>> Which DNSBL do you use?
>> Do you use local copies of any of the above?
>>
>
> Thanks for the quick reply. I use a list, which I reorder and track
> performance of from time to time. I just went back and noticed that I
> was querying ordb (off the air). Removing it seems to have fixed the
> problem.

Implementtion of DNSBL checks in sendmail via *sequential DNS queries*
"could be improved" by source code changes (understatement).

> I'm surprised that msn/hotmail is that sensitive to delays...haven't
> seen this problem elsewhere.

If hotmail is so sensitive during initial/first delivery then in may
be treated as (too?) aggressive optimization. I have used pretty "brave"
timeouts for *initial* (at once) delivery attempts myself.
If hotmail is so sensitive during consecutive delivery attempts then it
would constitute (for me) a lack of professionalizm.

> In any case, thanks again. And for completeness, my dnsbl list is
> currently below, with stats on spam blocked for the last 5.5 days:
>
> log start: Jan 14 04:03:31
> log end: Jan 19 14:25:57
>
> spamhaus: 4472
> dsbl: 205
> spamcop: 650
> njabl: 147
> rfc-ignorant: 0
> --------------------------
> total spam rejected = 5474
> total dial rejected = 2931
> ----------
> total rejected= 8405

Have you presented results of "parallel checks" or results of
"sequential checks" up to first listing?

I bet the later.

BTW XBL of spamhaus org contains big part of najbl.org listings.

--
[pl>en: Andrew] Andrzej Adam Filip : anfi@priv.onet.pl : anfi@xl.wp.pl
Before You Ask: http://anfi.homeunix.net/sendmail/B4UAsk-Sendmail.html
http://anfi.homeunix.net/sendmail/