Unable to get mail forwarded from qmail to Scalix

[enantiodromia]

I am trying to get my existing qmail server to forward inbound mail to my Scalix server. As soon as I can get this happening, I get to ditch qmail, but for now, I am unable to demonstrate a working Scalix account to management.

Im sure I am just missing a simple concept here, but if I cant produce a working Scalix account soon, I'm afraid I'll be stuck with qmail.

From what I can tell, the problem exists when qmail tries to forward the mail to Scalix. My qmail skills arent great, so I'm not really sure how to track down any logs showing errors.

I am able to send mail FROM my Scalix test accounts just fine, to other Scalix accounts, to qmail, and to external web mail sites.

Mail going TO my Scalix test accounts never work, unless they are sent from another Scalix account.

From reading other posts, I have determined that the below information is most likely useful to share.

Any help or insight would be great.

-e

Account Real Name: Test Account
LDAP and Scalix Login: test
Addresses: testaccount@mycompany.com, test.account@mycompany.com, test@mycompany.com

[root@mailserver test]# more .qmail
| /var/qmail/bin/forward test@my.scalix.server.com
/home/test/Maildir

[root@mailserver root]# ls -la /var/spool/mail/test
-rw------- 1 test mail 130108 May 4 16:48 /var/spool/mail/test

[root@mailserver root]# more /etc/hosts
# Do not remove the following line, or various programs
# that require network functionality will fail.
127.0.0.1 localhost.localdomain localhost

[root@mailserver root]# omsearch -e IA=Test.Account@my.scalix.server.com
[OM 16954] Entry not in the Directory

[root@mailserver root]# omshowu -n "Test Account"
Authentication ID: test
User Name : test account /CN=test account
MailNode : lantern,in
Internet Address : "test account" <test.account@mycompany.com>="test account" <test@mycompany.com>="test account" <testaccount@mycompany.com>
System Login : 60541
Password : set
External authentication only: NO
Admin Capabilities : NO
Mailbox Admin Capabilities : NO
Language : C
Virtual Vault : Enabled (default)
Mail Account: Unlocked
Last Signon : 05.04.05 15:02:46
Receipt of mail : ENABLED
Service level : 0
Excluded from Tidying : NO

[pete]

You need to add testaccount@my.scalix.server.com to the list of internet addreses aliases on the Scalix server. Scalix must match the incoming address to one that it has in it's database.

/Pete

[enantiodromia]

I found that all my test email have actually been delivered to the Scalix server, but they are sitting in /var/spool/mail/(username) .

Any ideas why Scalix isnt taking them, or, how to force the spool to try and deliver its contents?




"PostPosted: Thu May 05, 2005 2:41 pm Post subject:
You need to add testaccount@my.scalix.server.com to the list of internet addreses aliases on the Scalix server. Scalix must match the incoming address to one that it has in it's database.

/Pete"


Thanks for the insight Pete. I'll take a look at this. Perhaps I am confusing my mailserver.mycompany.com with mycompany.com when it comes to the Internet Address?

regards,

-e

[pete]

Your qmail server believes that it is the final resting place for this domain. Check your DNS settings - I don't know qmail, but there may be a way to force it to deliver to another server. You really need to clear up any domain issues.

[enantiodromia]

I am using the qmail equiv of a .forward file to move the mail over to the scalix server. as the mail is in the spool on the scalix machine, i think everything up to the point scalix actually accepting the mail is working as intended.

I have also verified the daemons and services are running, and have restarted them all just in case.

[pete]

Hmmm, the mail should definately NOT be in the sendmail spool. The Scalix service is suposed to be listening on port 25, sendmail listens on the loopback interface 127.0.0.1.

check the status of all Scalix services:

omstat -a
omstat -s

and post the output

[enantiodromia]

I did that right before I left the office, and all daemons and services were running. I also restarted them all for "practice".

so we are thinking that mail was properly routed to the scalix server itself, but for some reason scalix didnt see it. do i have the concept right?

thanks again for the help.

-e

[pete]

Try this:

On the Scalix server:

telnet localhost 25
quit

telnet <ip address of scalix server> 25
quit

Post output. I want to see which server is listening on port 25 - sendmail or Scalix.
Scalix should be on the IP address port 25
sendmail should be on the loopback port 25

[enantiodromia]

logical. thank you. I will try this in the morning.

-e

[enantiodromia]

[root@scalixserverroot]# telnet localhost 25
Trying 127.0.0.1...
Connected to localhost.localdomain (127.0.0.1).
Escape character is '^]'.
220 scalixserver.mycompany.com ESMTP Sendmail 8.12.11/8.12.11; Fri, 6 May 2005 07:44:30 -0700
quit

[root@scalixserver root]# telnet 10.250.100.27 25
Trying 10.250.100.27...
Connected to scalixserver.mycompany.com (10.250.100.27).
Escape character is '^]'.
220 scalixserver.mycomapny.com ESMTP Scalix SMTP Relay 9.2.0.115; Fri, 06 May 2005 07:44:53 -0700 (PDT)

So sendmail listens on 127.0.0.1, and Scalix on 10.250.100.27.

[enantiodromia]

User accounts must be added only after the Scalix server has the correct DNS configuration. When user accounts are created Scalix refers to the global DNS config for the user's info. If you change your DNS configuration (say from test.server.domain.com to domain.com) it seems the user accounts' info is not updated.

Scalix, maybe there needs to be a warning in the Admin console when DNS info is changed saying that any users previously added will no longer work without re-adding them, or, not have the users' DNS info static?

-e

[ScalixSupport]

Page 316 of the Scalix Administration Guide gives details on what you need to do after changing the server name.

This also applies if you are changing from a host name to a domain name.

Cheers

Dave

[enantiodromia]

The manuals are of course very nice, but I doubt most people print out or read 500+ pages re their email systems.

Still need some interactive warnings when breaking stuff :)

-e

[florian]

the thing here is - and that is often misinterpreted in scalix - is that scalix has NO concept whatsoever of a domain that it owns. it only checks FULL internet addresses, so for an address "florian@my.scalix.server" to arrive, this exact full address MUST be setup as an alias for the Florian user. it is NOT enough to have a user with "florian@company.com" as the match will be on name and domain, i.e. the FULL address.

we often use such setups in coexistence scenarios like the one described, i.e. add a user with his first address set to florian@company.com and a secondary address set as florian@newscalix.company.com. The first one will be used for all outbound traffic, the second one will be used for matching inbound traffic.

When the SMTP relay receives a message for which it cannot match the full address, it will determine that it is relaying. It will then check against smtpd.cfg if relaying is allowed. in the default install and the qmail scenario, this will probably be the case; smtpd.cfg will contain a line saying RELAY accept .company.com, so traffic from qmail.company.com will be accepted for releaying. In relaying, the message will just be handed over to sendmail for further outbound processing. sendmail in it's default configuration will just do a DNS lookup for an MX record to my.scalix.server - if it's not found it will sit in the queue, if it points back to the scalix server, it will go into a loop and bounce at some point.

proper configuration would be to put my.scalix.server in /etc/mail/local-host-names; then sendmail would bounce/reject addresses that are unknown because of incomplete directory entries.

Hope this helps - as a quick answer, one responder was right - add florian@scalix.server.my to the aliases of this user and you'll be fine.

[tchen]

florian,

Thanks for this clear desciption, which is very useful.

However, I have a problem in sendmail. I want my local sendmail to be able to send messages, but I found when I try to send mail from user1@mail.dom.ain to user2@dom.ain, (the former user1 is a local user, mail is our hostname, the latter is a scalix user2) the mail always get lost. (at the same time both address can send/receive mail to/from elsewhere)

It seems as if sendmail is trying to look for user2 in the local host, but since it doesn't exist, mail is lost.

Is there any hint how can I solve it?

I had another try by modifying submit.mc, changing from
FEATURE(`msp', `[127.0.0.1]')dnl
to
FEATURE(`msp', `[our mail server's ip]')dnl

then apply it by regenerate submit.cf and then omsendin, it kinda works, mails from user1@mail.dom.ain to user2@dom.ain get received, but this brings up another problem: user2@dom.ain can't send out because it goes into a loop, and rejected after a while for "too many hops".

Is there any clean way to do it? I'm sure it shouldn't be hard, but I just get lost here.
Thanks much!

tchen