SEO

I have set up an email server in one of my boxes, running
Slackware 12. The box is called my_server.my_domain.net (well, not
really; these are just pseudonyms) and has an IP address xx.yy.zz.aa
(again, not really) reachable from the external world.

When an email is received from the net, the following diagnostics
get printed out to my /var/log/maillog file:

Oct 8 10:39:32 folkvang sm-mta[10995]: l98HdWwt010995:
from=<xxx@yyy.com>, size=1252, class=0, nrcpts=1,
msgid=<346742.21731.qm@web35615.mail.mud.yyy.com>, pro
to=SMTP, daemon=MTA, relay=web35615.mail.mud.yyy.com [66.163.179.154]

Oct 8 10:42:41 folkvang sm-mta[10997]: l98HdWwt010995:
to=<xxx@my_server.my_domain.net>, delay=00:03:09, xdelay=00:03:09,
mailer=esmtp, pri=121252, relay=my_server.my_domain.net. [xx.yy.zz.aa],
dsn=4.0.0, stat=Deferred: Connection timed out with
my_server.my_domain.net

The message has been received by my_server from a yahoo.com MTA.
However, it never gets delivered the local user xxx, which does exist all
right in the my_server box. The traces are obtained in my_server. Why the
"Connection timed out" message? Everything is running in my_server; how
come it can't connect to itself? The messages received are sitting in
/var/spool/mqueue, waiting to be delivered to the local users they are
being sent to. What else do I have to do in my email server configuration
to fix this? What am I missing?

On Mon, 8 Oct 2007, H.K. Kingston-Smith wrote:

>
> I have set up an email server in one of my boxes, running
>
> to=SMTP, daemon=MTA, relay=web35615.mail.mud.yyy.com [66.163.179.154]

Ok, received well so far..



> Oct 8 10:42:41 folkvang sm-mta[10997]: l98HdWwt010995:
> to=<xxx@my_server.my_domain.net>, delay=00:03:09, xdelay=00:03:09,
> mailer=esmtp, pri=121252, relay=my_server.my_domain.net. [xx.yy.zz.aa],
> dsn=4.0.0, stat=Deferred: Connection timed out with
> my_server.my_domain.net

What sendmail instances are running? Sounds like the queue runner is not
running

> come it can't connect to itself? The messages received are sitting in
> /var/spool/mqueue, waiting to be delivered to the local users they are

Are you starting like this?, dont worry if -L options differ, it's the
arguments after that, that matter.

/usr/sbin/sendmail -L sendmail -bd -q15m
/usr/sbin/sendmail -L sm-msp-queue -Ac -q15m


--

Cheers
Res

On Tue, 09 Oct 2007 08:50:11 +1000, Res wrote:

> On Mon, 8 Oct 2007, H.K. Kingston-Smith wrote:
>
>
>> I have set up an email server in one of my boxes, running
>>
>> to=SMTP, daemon=MTA, relay=web35615.mail.mud.yyy.com [66.163.179.154]
>
> Ok, received well so far..
>
>
>
>> Oct 8 10:42:41 folkvang sm-mta[10997]: l98HdWwt010995:
>> to=<xxx@my_server.my_domain.net>, delay=00:03:09, xdelay=00:03:09,
>> mailer=esmtp, pri=121252, relay=my_server.my_domain.net. [xx.yy.zz.aa],
>> dsn=4.0.0, stat=Deferred: Connection timed out with
>> my_server.my_domain.net
>
> What sendmail instances are running? Sounds like the queue runner is not
> running
>
>> come it can't connect to itself? The messages received are sitting in
>> /var/spool/mqueue, waiting to be delivered to the local users they are
>
> Are you starting like this?, dont worry if -L options differ, it's the
> arguments after that, that matter.
>
> /usr/sbin/sendmail -L sendmail -bd -q15m /usr/sbin/sendmail -L
> sm-msp-queue -Ac -q15m

Sendmail is started from the rc.sendmail file shipped with
Slackware 12. According to that file (which I have not modified) sendmail
is started as follows:

/usr/sbin/sendmail -L sm-mta -bd -q25m
/usr/sbin/sendmail -L sm-msp-queue -Ac -q25m

I seem to have sorted it out but, unfortunately, I am not sure
how :-( I monkeyed a bit with sendmail.cnf and after restarting sendmail
the diagnostic line above becomes:

Oct 8 11:15:28 folkvang sm-mta[11228]: l98IAUY0011177:
to=<xxx@my_server.my_domain.net>, delay=00:04:58, xdelay=00:00:01,
mailer=local, pri=210896, dsn=2.0.0, stat=Sent

and everything works fine. The problem is, after reinstating the original
sendmail.cnf and restarting, everything still works!

On Tue, 9 Oct 2007, H.K. Kingston-Smith wrote:

> the diagnostic line above becomes:
>
> Oct 8 11:15:28 folkvang sm-mta[11228]: l98IAUY0011177:
> to=<xxx@my_server.my_domain.net>, delay=00:04:58, xdelay=00:00:01,
> mailer=local, pri=210896, dsn=2.0.0, stat=Sent
>
> and everything works fine. The problem is, after reinstating the original
> sendmail.cnf and restarting, everything still works!

LOL, well at least its all good now


--

Cheers
Res

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 2007-10-09, H.K. Kingston-Smith <HKK-S@yahoo.com> wrote:
> I seem to have sorted it out but, unfortunately, I am not sure
> how :-( I monkeyed a bit with sendmail.cnf and after restarting sendmail
> the diagnostic line above becomes:

That's the absolute worst feeling in the world: "It works now, but I
have no idea why." It's almost as if this was an Exchange server!

Seriously, take a look at postfix for a solid, easily configured mail
server. The documentation is great, and you can get SlackBuild scripts
for 11.0 and 12.0 from http://www.slackbuilds.org.

- --
It is better to hear the rebuke of the wise,
Than for a man to hear the song of fools.
Ecclesiastes 7:5
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (GNU/Linux)

iD8DBQFHC7AlrZS6hX/gvjoRAumKAKCdL04GSZlXlohx/5mZP9w6bP3hpACfbUzK
r2xJPM31+EFJm78xfYhPHn0=
=A8q7
-----END PGP SIGNATURE-----

+Alan Hicks+ wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 2007-10-09, H.K. Kingston-Smith <HKK-S@yahoo.com> wrote:
>
>> I seem to have sorted it out but, unfortunately, I am not sure
>>how :-( I monkeyed a bit with sendmail.cnf and after restarting sendmail
>>the diagnostic line above becomes:
>
>
> That's the absolute worst feeling in the world: "It works now, but I
> have no idea why." It's almost as if this was an Exchange server!

"Monkeying" with the configs instead of systematic working _can_
contiribute to such a feeling. :-)

The second log line in the original posting showed that sendmail
tried to use the "relay" mailer te deliver the message. For local
delivery the "local" mailer should have been used. This indicates
that sendmail didn't recongnize it's own name as a local distination.
A second issue was that sendmail could not reach its ounw IP
address.

Mode details were not supplied and even these few details
(host name an IP number) were obfuscated.
A little more details about configuration and network environment
would have been helpfull to diagnos the problem.
Of course there should be a logical explaination. Sendmail is not
Exchange after all!

I'll make a wild guess. That might make the "absolute worst feeling
in the world" a little better.
Combineing the "not recognised own name" and the "unreachable
own IP number" I start directly thinking about a local
network using a private IP range and located behind a
masquerading/NAT-ing firewall. The host name and IP number are
the official/external name and number. At startup sendmail
tries to determine its name, and the name(s) of its network
interface(s) via the resolver library (reverse dns lookups,
/etc/hosts file). Of course a reverse lookup of a local
(masqueraded) IP number will not resolve to the external
host name. The unreachable external IP number could have
been a firewall issue.
Specifying the host name as a local detination would solve
both problems. The easy way to do this is putting the
name in /etc/mail/local-host-names. The hard way is putting
it in sendmail.mc and generating a new sendmail.cf.
Of course this won't change the firewall rules but sendmail
doesn't need to contact the external IP number anymore
once the name is recognized as a local destination.

As I said it's a wild guess, but it's the best I can do with
as little information as the OP supplied. I'm curious if it
makes any sense.


Regards,

Kees.

--
Kees Theunissen.

Well, I was wondering when the spammers would get around to it.

On Tue, 9 Oct 2007, +Alan Hicks+ wrote:

> Seriously, take a look at postfix for a solid, easily configured mail
> server. The documentation is great, and you can get SlackBuild scripts

For that mater, look at Exim and Qmail as well.

If you are going to spam that an MTA is better, how about you prove why
its better, somthing nobody has been able to do yet, since postfix
basically copies everything sendmail does anyway, also postfixes
implimentation of the sendmail (look-ahead|milter-ahead|smd-sav) milter,
fails dismally under heavy use.


--

Cheers
Res

On 2007-10-09, Res <res@ausics.net> wrote:
>
> Well, I was wondering when the spammers would get around to it.
>
> On Tue, 9 Oct 2007, +Alan Hicks+ wrote:
>
>> Seriously, take a look at postfix for a solid, easily configured mail
>> server. The documentation is great, and you can get SlackBuild scripts
>
> For that mater, look at Exim and Qmail as well.
>
> If you are going to spam that an MTA is better,

I see no mention in Alan's post about postfix being better than any
other MTA.

Qmail is an utter disaster from a configuration and maintenance
standpoint. I'd suggest against it for anyone except my enemies.

--keith

--
kkeller-usenet@wombat.san-francisco.ca.us
(try just my userid to email me)
AOLSFAQ=http://www.therockgarden.ca/aolsfaq.txt
see X- headers for PGP signature information

On Tue, 9 Oct 2007, Keith Keller wrote:

>> If you are going to spam that an MTA is better,
>
> I see no mention in Alan's post about postfix being better than any
> other MTA.
>

IMHO thats what he was doing, the OP says he is using MTA a, why
should anyone say 'check out MTA b', its like posting on this list and
saying we should check out SuSE or fuckdora.

> Qmail is an utter disaster from a configuration and maintenance
> standpoint. I'd suggest against it for anyone except my enemies.

on its own, yes, but it is powerfull with a few patches and has its place,
especially for those in large hosting arenas or ISP's with millions and
millions of customers, Qmail/Vpopmail combo is far superior performer
then Sendmail/Cyrus(2) and postfix's implimentation, disastor recovery is
so simple my 12yo nephew could do it.

Its only set back in djb is an arrogant prick , and has not maintained
it, its reliance upon others to supply modern day patching, many large
ISP/ASP/OSP's around the world still use it, and probably always will.
I don't use B.Shupps or JMS's patches, I took netqmail and got the
other patches I want that do much the same sort of thing, and work from
there, and remember, when djb wrote qmail, spam was something you went to
a supermarket to buy and eat.

--

Cheers
Res

On 2007-10-10, Res <res@ausics.net> wrote:
> On Tue, 9 Oct 2007, Keith Keller wrote:
>
>>> If you are going to spam that an MTA is better,
>>
>> I see no mention in Alan's post about postfix being better than any
>> other MTA.
>>
>
> IMHO thats what he was doing, the OP says he is using MTA a, why
> should anyone say 'check out MTA b', its like posting on this list and
> saying we should check out SuSE or fuckdora.

Not really, but you're welcome to your own opinion. Some software is
more suited to certain environments. Some people just grok some
software better than others. I certainly wouldn't claim that postfix is
better than sendmail (or vice versa), but I understand postfix better
than sendmail. Other admins understand sendmail better. The person who
said he's using MTA a might try MTA b and decide for himself which he
prefers.

--keith



--
kkeller-usenet@wombat.san-francisco.ca.us
(try just my userid to email me)
AOLSFAQ=http://www.therockgarden.ca/aolsfaq.txt
see X- headers for PGP signature information