Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.user > #172104 > unrolled thread

exim4 some config error causing error how to pinpoint

Back to article view | Back to linux.debian.user

Contents

  exim4 some config error causing error how to pinpoint Harry Putnam <reader@newsguy.com> - 2016-09-13 04:00 +0200
    Re: exim4 some config error causing error how to pinpoint Liam O'Toole <liam.p.otoole@gmail.com> - 2016-09-13 12:30 +0200
      Re: exim4 some config error causing error how to pinpoint Harry Putnam <reader@newsguy.com> - 2016-09-14 23:30 +0200
        Re: exim4 some config error causing error how to pinpoint Liam O'Toole <liam.p.otoole@gmail.com> - 2016-09-15 11:10 +0200
          Re: exim4 some config error causing error how to pinpoint Harry Putnam <reader@newsguy.com> - 2016-09-15 19:40 +0200
            Re: exim4 some config error causing error how to pinpoint Liam O'Toole <liam.p.otoole@gmail.com> - 2016-09-16 10:50 +0200
              Re: exim4 some config error causing error how to pinpoint Harry Putnam <reader@newsguy.com> - 2016-10-01 15:40 +0200
                Re: exim4 some config error causing error how to pinpoint Mark Fletcher <mark27q1@gmail.com> - 2016-10-01 16:10 +0200

#172104 — exim4 some config error causing error how to pinpoint

I've attempted to setup exim4 on a second debian OS, using the
working configuration from the older one.

Before getting too detailed I think I see something in output of my
debug technique that indicated somewhere my host is telling exim the
wrong host name.

(dv is host with working config )

I use this technique; sending a mail like this:
mailx -v -s "TEST 160912_211756 dv" reader@newsguy.com < ~/txtmsg.txt
                               
I run that command then watch the smtp output from the little
test script named `tmail'.

On the working host the first bit of output:

 $ tmail reader@newsguy.com
sending like this:
mailx -v -s "TEST 160912_211756 dv" reader@newsguy.com < ~/txtmsg.txt
LOG: MAIN
  <= harry@dv U=harry P=local S=546

[...]

Now the second, non-working host:

 tmail reader@newsguy.com
sending like this: (Non-working host is `d2')
mailx -v -s "TEST 160912_211617 d2" reader@newsguy.com < ~/txtmsg.txt
                           
LOG: MAIN      
  <= harry@d2local.lan U=harry P=local S=440
 ^^^^^^^^^^^^^^^^^^^^^
Notice the above difference .. in the working host it shows
  harry@dv (hostname is dv.local.lan)
  In the non-working host:
  harry@d2local.lan (hostname is d2.local.lan)

Somewhere in my setup exim4 is finding the hostname without the
separating dot in d2.local.lan   and is using d2local.lan

Which may be causing the failure:  Following the smtp otuput:

-------       -------       ---=---       -------       -------

 delivering 1bjcKi-0001Ey-9B
R: nonlocal for reader@newsguy.com
LOG: MAIN
  ** reader@newsguy.com R=nonlocal: Mailing to remote domains not supported
LOG: MAIN
  <= <> R=1bjcKi-0001Ey-9B U=Debian-exim P=local S=1650
LOG: MAIN
  Completed
delivering 1bjcKi-0001F0-Fy
R: nonlocal for harry@d2local.lan
LOG: MAIN
  ** harry@d2local.lan R=nonlocal: Mailing to remote domains not supported
LOG: MAIN
  Frozen (delivery error message)

------- ------- ---=--- ------- -------

Or am I making too much of the missing dot and it really is that I've
missed an important setting that lets exim4 deliver to outside of lan?

[toc] | [next] | [standalone]

#172115

On 2016-09-13, Harry Putnam <reader@newsguy.com> wrote:
> I've attempted to setup exim4 on a second debian OS, using the
> working configuration from the older one.
>
> Before getting too detailed I think I see something in output of my
> debug technique that indicated somewhere my host is telling exim the
> wrong host name.
>
> (dv is host with working config )
>
> I use this technique; sending a mail like this:
> mailx -v -s "TEST 160912_211756 dv" reader@newsguy.com < ~/txtmsg.txt
>                                
> I run that command then watch the smtp output from the little
> test script named `tmail'.
>
> On the working host the first bit of output:
>
>  $ tmail reader@newsguy.com
> sending like this:
> mailx -v -s "TEST 160912_211756 dv" reader@newsguy.com < ~/txtmsg.txt
> LOG: MAIN
>  <= harry@dv U=harry P=local S=546
>
> [...]
>
> Now the second, non-working host:
>
>  tmail reader@newsguy.com
> sending like this: (Non-working host is `d2')
> mailx -v -s "TEST 160912_211617 d2" reader@newsguy.com < ~/txtmsg.txt
>                            
> LOG: MAIN      
>  <= harry@d2local.lan U=harry P=local S=440
>  ^^^^^^^^^^^^^^^^^^^^^
> Notice the above difference .. in the working host it shows
>   harry@dv (hostname is dv.local.lan)
>   In the non-working host:
>   harry@d2local.lan (hostname is d2.local.lan)
>
> Somewhere in my setup exim4 is finding the hostname without the
> separating dot in d2.local.lan   and is using d2local.lan
>
> Which may be causing the failure:  Following the smtp otuput:
>
> -------       -------       ---=---       -------       -------
>
>  delivering 1bjcKi-0001Ey-9B
> R: nonlocal for reader@newsguy.com
> LOG: MAIN
>   ** reader@newsguy.com R=nonlocal: Mailing to remote domains not supported
> LOG: MAIN
>  <= <> R=1bjcKi-0001Ey-9B U=Debian-exim P=local S=1650
> LOG: MAIN
>   Completed
> delivering 1bjcKi-0001F0-Fy
> R: nonlocal for harry@d2local.lan
> LOG: MAIN
>   ** harry@d2local.lan R=nonlocal: Mailing to remote domains not supported
> LOG: MAIN
>   Frozen (delivery error message)
>
> ------- ------- ---=--- ------- -------
>
> Or am I making too much of the missing dot and it really is that I've
> missed an important setting that lets exim4 deliver to outside of lan?
>

Run 'dpkg-reconfigure exim4-config' as root on the second host to repair
the hostname. See §8.5.3. of the documentation[1] for details.

You can also inspect the working settings on the first machine using
'debconf-show exim4-config'.

1: https://www.debian.org/releases/stable/i386/ch08s05.html.en


-- 

Liam

[toc] | [prev] | [next] | [standalone]

#172181

Liam O'Toole <liam.p.otoole@gmail.com> writes:

> dpkg-reconfigure exim4-config

Thanks, that was pretty painless.  And thanks for the url to the
documentation section.

I've now progressed on to where I was aiming for.

I wanted to have one debian box as main mail client.  Yet be able to
send mail from a couple of others.

I've managed to get that working by aiming the secondary hosts at the
main mail host as smarthost.  That main host is inturn aimed at an
internet smarthost.

So far all working except one glitch.

Here is the cast of characters:
HOST1 Main mail host
HOST2 secondary host
User1 HOST1
User2 HOST1
user HOST2

I can send a mail from HOST2 to a user on HOST1 using the correct
address.  user1@HOST1

That main host has only two users.  I can send mail from HOST2 to
only one of those users.  That is. if, from HOST2, I send mail
to a HOST1 user like:

  user1@HOST1   It arrives in user1 slot at /var/spool/mail/user1 on
HOST1. 

However If I try sending to user2@HOST1 it never gets there. Even
though the smtp output on sending HOST2 appears to be saying it was
sent and received successfully.

   ------- ------- ---=--- ------- -------

orignating on HOST2 

tmail user2@HOST1
sending like this:
mailx -v -s "TEST 160914_163910" user1@HOST1 < ~/txtmsg.txt
LOG: MAIN
  <= user@HOST2.local.lan U=user1 P=local S=422

smtp output:

 delivering 1bkGxi-0002L3-U4
R: smarthost for user2@HOST1
T: remote_smtp_smarthost for user1@HOST1
Connecting to HOST1.local.lan [192.168.1.5]:25 ... connected
  SMTP<< 220 dv ESMTP Exim 4.86 Wed, 14 Sep 2016 16:47:36 -0400
  SMTP>> EHLO HOST2.local.lan
  SMTP<< 250-dv Hello HOST2.local.lan [192.168.1.17]
         250-SIZE 52428800
         250-8BITMIME
         250-PIPELINING
         250 HELP
  SMTP>> MAIL FROM:<user@HOST2.local.lan> SIZE=1464
  SMTP>> RCPT TO:<user2@HOST1>
  SMTP>> DATA
  SMTP<< 250 OK
  SMTP<< 250 Accepted
  SMTP<< 354 Enter message, ending with "." on a line by itself
  SMTP>> writing message and terminating "."
  SMTP<< 250 OK id=1bkH5k-0008FT-UQ
  SMTP>> QUIT
LOG: MAIN
  => user2@HOST1 R=smarthost T=remote_smtp_smarthost H=HOST1.local.lan [192.168.1.5] C="250 OK id=1bkH5k-0008FT-UQ"
LOG: MAIN
  Completed

Mail never appears at HOST1 /var/spool/mail/user2

[toc] | [prev] | [next] | [standalone]

#172204

On 2016-09-14, Harry Putnam <reader@newsguy.com> wrote:
> Liam O'Toole <liam.p.otoole@gmail.com> writes:
>
>> dpkg-reconfigure exim4-config
>
> Thanks, that was pretty painless.  And thanks for the url to the
> documentation section.
>
> I've now progressed on to where I was aiming for.
>
> I wanted to have one debian box as main mail client.  Yet be able to
> send mail from a couple of others.
>
> I've managed to get that working by aiming the secondary hosts at the
> main mail host as smarthost.  That main host is inturn aimed at an
> internet smarthost.
>
> So far all working except one glitch.
>
> Here is the cast of characters:
> HOST1 Main mail host
> HOST2 secondary host
> User1 HOST1
> User2 HOST1
> user HOST2
>
> I can send a mail from HOST2 to a user on HOST1 using the correct
> address.  user1@HOST1
>
> That main host has only two users.  I can send mail from HOST2 to
> only one of those users.  That is. if, from HOST2, I send mail
> to a HOST1 user like:
>
>   user1@HOST1   It arrives in user1 slot at /var/spool/mail/user1 on
> HOST1. 
>
> However If I try sending to user2@HOST1 it never gets there. Even
> though the smtp output on sending HOST2 appears to be saying it was
> sent and received successfully.
>
>    ------- ------- ---=--- ------- -------
>
> orignating on HOST2 
>
> tmail user2@HOST1
> sending like this:
> mailx -v -s "TEST 160914_163910" user1@HOST1 < ~/txtmsg.txt
> LOG: MAIN
>  <= user@HOST2.local.lan U=user1 P=local S=422
>
> smtp output:
>
>  delivering 1bkGxi-0002L3-U4
> R: smarthost for user2@HOST1
> T: remote_smtp_smarthost for user1@HOST1
> Connecting to HOST1.local.lan [192.168.1.5]:25 ... connected
>   SMTP<< 220 dv ESMTP Exim 4.86 Wed, 14 Sep 2016 16:47:36 -0400
>   SMTP>> EHLO HOST2.local.lan
>   SMTP<< 250-dv Hello HOST2.local.lan [192.168.1.17]
>          250-SIZE 52428800
>          250-8BITMIME
>          250-PIPELINING
>          250 HELP
>   SMTP>> MAIL FROM:<user@HOST2.local.lan> SIZE=1464
>   SMTP>> RCPT TO:<user2@HOST1>
>   SMTP>> DATA
>   SMTP<< 250 OK
>   SMTP<< 250 Accepted
>   SMTP<< 354 Enter message, ending with "." on a line by itself
>   SMTP>> writing message and terminating "."
>   SMTP<< 250 OK id=1bkH5k-0008FT-UQ
>   SMTP>> QUIT
> LOG: MAIN
>  => user2@HOST1 R=smarthost T=remote_smtp_smarthost H=HOST1.local.lan [192.168.1.5] C="250 OK id=1bkH5k-0008FT-UQ"
> LOG: MAIN
>   Completed
>
> Mail never appears at HOST1 /var/spool/mail/user2
>
>

Does user2 appear in the file /etc/aliases on HOST1? Is there a
/home/user2/.forward file on that host? Either of those would cause mail
addressed to user2@HOST1 to be relayed to another user instead.

Is there anything suspicious in the file /var/log/exim4/mainlog on
HOST1?

-- 

Liam

[toc] | [prev] | [next] | [standalone]

#172249

Liam O'Toole <liam.p.otoole@gmail.com> writes:

>> Mail never appears at HOST1 /var/spool/mail/user2
>>
>>
>
> Does user2 appear in the file /etc/aliases on HOST1? Is there a
No

> /home/user2/.forward file on that host? Either of those would cause mail
No

> Is there anything suspicious in the file /var/log/exim4/mainlog on
> HOST1?
Yes, there is this bit that seems to indicate that the message is
being processed by procmail.

dv.local.lan is HOST1  smarthost set to smtp.gmail.com
d2.local.lan is HOST2  smarthost set to HOST1

There is a user named `harry' on both hosts.

-------       -------       ---=---       -------       ------- 
2016-09-15 12:40:29 1bkZi9-0000Ne-MU <= harry@d2.local.lan H=(d2.local.lan) [192.168.1.17] P=esmtp S=608 id=E1bkZa7-0006Mb-CG@d2.local.lan
2016-09-15 12:40:31 1bkZi9-0000Ne-MU => harry <harry@dv> R=procmail T=procmail_pipe
2016-09-15 12:40:31 1bkZi9-0000Ne-MU Completed
2016-09-15 12:52:31 Start queue run: pid=1574
2016-09-15 12:52:31 End queue run: pid=1574
-------       -------       ---=---       -------       ------- 

There is a ~/.procmailrc for user2 but no specifc recipe that would
grab mail addressed to USER2@HOST1 (harry@dv).

However my .procmailrc has a final rule that catches anything not
already handled:

  :0
  post_ex.in

However the mail addressed user2@host1 is not going to that inbox (or
any other) either.

But dropping that mystery for the time being there seems to be
something bigger happening.

Since my previous post I have found that all those test messages to
user2 and user1 on HOST1 from HOST2 are being sent out to the
HOST1 smarthost and piling up as undeleverable on a gmail account.  My
smarthost is smtp.gmail.com

I guess that does make sense.  However I guess I just supposed that
somehow HOST1 would magically know not to try to send mail from other
lan hosts addressed to users on HOST1, out to the internet smarthost.

Maybe I need some other mechanism for that to happen?

To summarize the setup again very briefly... since it seems to be
getting more confusing ... at least to me.

HOST1 is setup to relay mail from its users to an internet smarthost
(smtp.gmail.com).

HOST2 is setup to relay mail from its users to HOST1 as its smarthost

When I try to send a message from USER@HOST2 to USER@HOST1, instead of
just delivering to USER@HOST1 it appears to be sending that message on
to the gmail smtp server.

The local lan is not a real FQDN so not suprising that the gmail
server cannot deliver it.... but how to prevent that local mail from
being sent to gmail server to start with?

[toc] | [prev] | [next] | [standalone]

#172289

On 2016-09-15, Harry Putnam <reader@newsguy.com> wrote:
> Liam O'Toole <liam.p.otoole@gmail.com> writes:
>
>>> Mail never appears at HOST1 /var/spool/mail/user2
>>>
>>>
>>
>> Does user2 appear in the file /etc/aliases on HOST1? Is there a
> No
>
>> /home/user2/.forward file on that host? Either of those would cause mail
> No
>
>> Is there anything suspicious in the file /var/log/exim4/mainlog on
>> HOST1?
> Yes, there is this bit that seems to indicate that the message is
> being processed by procmail.
>
> dv.local.lan is HOST1  smarthost set to smtp.gmail.com
> d2.local.lan is HOST2  smarthost set to HOST1
>
> There is a user named `harry' on both hosts.
>
> -------       -------       ---=---       -------       ------- 
> 2016-09-15 12:40:29 1bkZi9-0000Ne-MU <= harry@d2.local.lan H=(d2.local.lan) [192.168.1.17] P=esmtp S=608 id=E1bkZa7-0006Mb-CG@d2.local.lan
> 2016-09-15 12:40:31 1bkZi9-0000Ne-MU => harry <harry@dv> R=procmail T=procmail_pipe
> 2016-09-15 12:40:31 1bkZi9-0000Ne-MU Completed
> 2016-09-15 12:52:31 Start queue run: pid=1574
> 2016-09-15 12:52:31 End queue run: pid=1574
> -------       -------       ---=---       -------       ------- 
>
> There is a ~/.procmailrc for user2 but no specifc recipe that would
> grab mail addressed to USER2@HOST1 (harry@dv).
>
> However my .procmailrc has a final rule that catches anything not
> already handled:
>
>  :0
>   post_ex.in
>
> However the mail addressed user2@host1 is not going to that inbox (or
> any other) either.
>
> But dropping that mystery for the time being there seems to be
> something bigger happening.
>
> Since my previous post I have found that all those test messages to
> user2 and user1 on HOST1 from HOST2 are being sent out to the
> HOST1 smarthost and piling up as undeleverable on a gmail account.  My
> smarthost is smtp.gmail.com
>
> I guess that does make sense.  However I guess I just supposed that
> somehow HOST1 would magically know not to try to send mail from other
> lan hosts addressed to users on HOST1, out to the internet smarthost.
>
> Maybe I need some other mechanism for that to happen?
>
> To summarize the setup again very briefly... since it seems to be
> getting more confusing ... at least to me.
>
> HOST1 is setup to relay mail from its users to an internet smarthost
> (smtp.gmail.com).
>
> HOST2 is setup to relay mail from its users to HOST1 as its smarthost
>
> When I try to send a message from USER@HOST2 to USER@HOST1, instead of
> just delivering to USER@HOST1 it appears to be sending that message on
> to the gmail smtp server.
>
> The local lan is not a real FQDN so not suprising that the gmail
> server cannot deliver it.... but how to prevent that local mail from
> being sent to gmail server to start with?
>
>

You need to tell exim4 on HOST1 that it is the 'final destination' for
emails within your local network. If you go back through the debconf
questions for exim4-config, you'll be asked about 'Other destinations
for which mail is accepted'. In your case the answer should be
'*.local.lan'. Then local emails will no longer be forwarded upstream.
Obviously, you will need to ensure that the various local fully
qualified domain names are resolvable, either by running your own DNS or
maintaining /etc/host files correctly.

-- 

Liam

[toc] | [prev] | [next] | [standalone]

#173047

Liam O'Toole <liam.p.otoole@gmail.com> writes:

[...]

>> When I try to send a message from USER@HOST2 to USER@HOST1, instead of
>> just delivering to USER@HOST1 it appears to be sending that message on
>> to the gmail smtp server.
>>
>> The local lan is not a real FQDN so not suprising that the gmail
>> server cannot deliver it.... but how to prevent that local mail from
>> being sent to gmail server to start with?
>>
>>
>
> You need to tell exim4 on HOST1 that it is the 'final destination' for
> emails within your local network. If you go back through the debconf
> questions for exim4-config, you'll be asked about 'Other destinations
> for which mail is accepted'. In your case the answer should be
> '*.local.lan'. Then local emails will no longer be forwarded upstream.
> Obviously, you will need to ensure that the various local fully
> qualified domain names are resolvable, either by running your own DNS or
> maintaining /etc/host files correctly.

Thank you.  That fixed HOST1.  Still haven't worked out how to be able
to send mail from a user@HOST2 to a user@HOST1 (both inside home lan ).

With the setting you suggest above... mail from user@HOST2 to
user@HOST1 fails with smtp output:

  sending like this:
  mailx -v -s "TEST 161001_092048 d2" harry@dv.local.lan < ~/txtmsg.txt
  LOG: MAIN
    <= harry@d2.local.lan U=harry P=local S=442
  delivering 1bqKDj-00009o-AK
  R: smarthost for harry@dv.local.lan
  T: remote_smtp_smarthost for harry@dv.local.lan
  LOG: MAIN
    ** harry@dv.local.lan R=smarthost T=remote_smtp_smarthost: retry time not reached for any host after a long failure period
  LOG: MAIN
    <= <> R=1bqKDj-00009o-AK U=Debian-exim P=local S=1680
  LOG: MAIN
    Completed
  delivering 1bqKDj-00009s-Hi
  R: system_aliases for harry@d2.local.lan
  R: userforward for harry@d2.local.lan
  R: procmail for harry@d2.local.lan
  R: maildrop for harry@d2.local.lan
  R: lowuid_aliases for harry@d2.local.lan (UID 1000)
  R: local_user for harry@d2.local.lan
  T: appendfile for harry@d2.local.lan
  LOG: MAIN
    => harry <harry@d2.local.lan> R=local_user T=mail_spool
  LOG: MAIN
    Completed

And from /var/log/exim4/mainlog:

  2016-10-01 09:20:51 1bqKDj-00009o-AK <= harry@d2.local.lan U=harry P=local S=442
  2016-10-01 09:20:51 1bqKDj-00009o-AK ** harry@dv.local.lan R=smarthost T=remote_smtp_smarthost: retry time not reached for any host after a long failure period
  2016-10-01 09:20:51 1bqKDj-00009s-Hi <= <> R=1bqKDj-00009o-AK U=Debian-exim P=local S=1680
  2016-10-01 09:20:51 1bqKDj-00009o-AK Completed
  2016-10-01 09:20:51 1bqKDj-00009s-Hi => harry <harry@d2.local.lan> R=local_user T=mail_spool
  2016-10-01 09:20:51 1bqKDj-00009s-Hi Completed

I think those ouputs mean that the local mail being sent from HOST2 to
HOST1 is still attempting to be sent to smtp.gmail.com (The smarthost
for HOST1)

I have HOST2 set like this in /etc/exim4/update-exim4.conf.conf:

  dc_eximconfig_configtype='smarthost'
  dc_other_hostnames=''
  dc_local_interfaces='127.0.0.1 ; ::1'
  dc_readhost='d2.local.lan'  # (HOST2)
  dc_relay_domains=''
  dc_minimaldns='false'
  dc_relay_nets=''
  dc_smarthost='dv.local.lan' # (HOST1)
  CFILEMODE='644'
  dc_use_split_config='false'
  dc_hide_mailname='false'
  dc_mailname_in_oh='true'
  dc_localdelivery='mail_spool'

Is this wrong?  Is setting HOST1 as smarthost for HOST2 a mistake?

Contents of /etc/exim4/update-exim4.conf.conf on HOST1

  dc_eximconfig_configtype='smarthost'
  dc_other_hostnames='*.local.lan' # NOTE: This was supposed to tell
                                   # HOST1 that mail from the local
                                   # lan is to be delivered to local
                                   # users... not sent to smarthost 
        
  dc_local_interfaces='127.0.0.1 ; ::1'
  dc_readhost=''
  dc_relay_domains=''
  dc_minimaldns='false'
  dc_relay_nets='192.168.1.0/24'
  dc_smarthost='smtp.gmail.com::587'
  CFILEMODE='644'
  dc_use_split_config='false'
  dc_hide_mailname='false'
  dc_mailname_in_oh='true'
  dc_localdelivery='mail_spool'
  smtp_accept_queue_per_connection=300 

The attempted send from HOST2 to HOST1 does not appear in the
/var/log/exim4/mainlog on HOST1 at all.

I'm sorry to be so confused about setting this up... but am pretty
confused at this point.

[toc] | [prev] | [next] | [standalone]

#173050

On Sat, Oct 01, 2016 at 09:37:34AM -0400, Harry Putnam wrote:
> Thank you.  That fixed HOST1.  Still haven't worked out how to be able
> With the setting you suggest above... mail from user@HOST2 to
> user@HOST1 fails with smtp output:
> 
> 
> I'm sorry to be so confused about setting this up... but am pretty
> confused at this point.
> 

I suggest you read through the thread that is live right now in this 
list, started by mo and replied to by Liam, myself and Brian (plus a 
couple of others) about getting exim set up to deliver mail around a 
local network and to external addresses. The big clue to getting it 
working is "hubbed hosts".

Mark

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.user

csiph-web