[SOLVED] Zimbra 6 -> 7 - Zimbra does not know its own virtual hosts

[TheInfinity]

Hi everyone,

I made an upgrade this evening - and got a really strange bug. I can send mails to external hosts like before - but any mail to zimbra fails because of "Relay Access denied". If I try to send something via webinterface I get an "mail for vereinlandwege.de loops back to myself" error.

My setup:
- Ubuntu 10.04
- Zimbra OpenSource Edition 7, 64bit

Server:
- internal hostname: sun.vereinlandwege.de
- external hostname: mail.vereinlandwege.de
- mail adresses: name@vereinlandwege.de

If I send to outside it looks like this in target hostname:

Feb  8 00:46:12 lvps83-169-1-219 postfix/smtpd[16090]: connect from mail.vereinlandwege.de[88.79.167.149]
Feb  8 00:46:27 lvps83-169-1-219 postfix/smtpd[16090]: warning: 149.167.79.88.list.dsbl.org: RBL lookup error: Host or domain name not found. Name service error for name=149.167.79.88.list.dsbl.org type=A: Host not found, try againFeb  8 00:46:28 lvps83-169-1-219 postfix/smtpd[16090]: 19085166E19FD: client=mail.vereinlandwege.de[88.79.167.149]Feb  8 00:46:28 lvps83-169-1-219 postfix/cleanup[22216]: 19085166E19FD: message-id=<1c4da0a0-605d-408e-85cf-319ac434d616@sun>Feb  8 00:46:28 lvps83-169-1-219 postfix/qmgr[2007]: 19085166E19FD: from=, size=811, nrcpt=1 (queue active)Feb  8 00:46:28 lvps83-169-1-219 postfix/smtpd[16090]: disconnect from mail.vereinlandwege.de[88.79.167.149]

mail.vereinlandwege.de looks perfect - and it is perfect, it works. But internally it gets strange. A mail from webinterface to the same account:

Feb  8 01:26:46 sun postfix/smtpd[25565]: connect from localhost[127.0.0.1]
Feb  8 01:26:46 sun postfix/smtpd[25565]: E21AA4C14F4: client=localhost[127.0.0.1]Feb  8 01:26:46 sun postfix/cleanup[25568]: E21AA4C14F4: message-id=<70be5408-81a1-45f9-a96a-2fc115108577@sun>
Feb  8 01:26:46 sun postfix/qmgr[21006]: E21AA4C14F4: from=, size=593, nrcpt=1 (queue active)Feb  8 01:26:46 sun postfix/smtpd[25565]: disconnect from localhost[127.0.0.1]
Feb  8 01:26:47 sun postfix/smtp[25569]: E21AA4C14F4: to=, relay=none, delay=0.07, delays=0.05/0.02/0/0, dsn=5.4.6, status=bounced (mail for vereinlandweg
e.de loops back to myself)
Feb  8 01:26:47 sun postfix/cleanup[25568]: 043354C14F6: message-id=<20110208002647.043354C14F6@sun.localdomain>
Feb  8 01:26:47 sun postfix/bounce[25571]: E21AA4C14F4: sender non-delivery notification: 043354C14F6
Feb  8 01:26:47 sun postfix/qmgr[21006]: E21AA4C14F4: removed
Feb  8 01:26:47 sun postfix/smtp[25569]: 043354C14F6: to=, relay=none, delay=0.03, delays=0.03/0/0/0, dsn=5.4.6, status=bounced (mail for vereinlandwege.de loops back to myself)
Feb  8 01:26:47 sun postfix/qmgr[21006]: 043354C14F6: removed

This looks strange. It looks even more strange if I try to send a mail from outside:

Feb  8 01:22:51 sun postfix/smtpd[24666]: connect from sectio-aurea.org[83.169.1.219]
Feb  8 01:22:51 sun postfix/smtpd[24666]: NOQUEUE: reject: RCPT from sectio-aurea.org[83.169.1.219]: 554 5.7.1 : Relay access denied; from= to= proto=ESMTP helo=

Why does it not know its own virtual domains?! Because - vereinlandwege.de is active in zimbra admin interface and I can login with vereinlandwege.de users.

/etc/hosts:

192.168.1.1     sun.vereinlandwege.de sun.vereinlandwege.de mail.vereinlandwege.de vereinlandwege.de sun

MX-Records always point to mail.vereinlandwege.de, intern and extern.

So - any ideas where the problem might be? Thanks for any idea!

[bku09]

I have exactly the same problems since I updated from 6.0.10GA to 7.0.0GA, everything was working fine with 6.0.10GA.

still trying to fix the issue, will keep you updated ...

Bernd

[waltholman]

Another me too.... I noticed in the /opt/zimbra/postfix/conf directory that the main.cf file was a zero (0) byte file. Re-running zmmtactl stop/start did not create the correct file (yet).

Update: After restoring the old main.cf from my 6.0.10 backup, the system is receiving and delivering mail once again. Glad this was my home server & not work

Now, to try and figure out why the main.cf doesn't get created properly....

[bku09]

jep right, there is definitely something wrong with the main.cf config file generation.

in the meantime you can simply replace the main.cf with the previous version even if you don't have any "real" backups :-)

as root enter:
cd /opt/zimbra/postfix/conf
cp /opt/zimbra/postfix-2.6.7.2z/conf/main.cf . (change 2.6.7.2z to your last version)
su - zimbra -c "zmmtactl restart"

everything should be back to normal now, but still trying to figure out the root cause ...

Bernd

[quanah]

The configuration is generated by the command /opt/zimbra/libexec/configrewrite mta

Can someone who is seeing this issue paste the output of running this manually?

[bku09]

Originally Posted by quanah

The configuration is generated by the command /opt/zimbra/libexec/configrewrite mta

Can someone who is seeing this issue paste the output of running this manually?

[zimbra@zimbra ~]$ /opt/zimbra/libexec/configrewrite mta
[zimbra@zimbra ~]$

no errors on stdout, no changes to main.cf after this command (tested with user zimbra and root)

[Mark L Sung]

Guys, I though ZCS7 is still in beta, people just can wait to moving forward. lol

[bku09]

Originally Posted by Mark L Sung

Guys, I though ZCS7 is still in beta, people just can wait to moving forward. lol

just go to Zimbra offers Open Source email server software and shared calendar for Linux and the Mac. and read the headline surprise!

[quanah]

Does the problem reappear if you run (as zimbra):
postfix stop
postfix start

If it does, then we'll actually have to do the following, since configrewrite squelches its output:

1. postfix stop (as zimbra user)
2. echo REWRITE mta | /bin/nc -w 120 localhost 7171 (as zimbra user)

--Quanah

[KevinH]

Take a look at this bug. Permissions and sudoers file is important.

https://bugzilla.zimbra.com/show_bug.cgi?id=55541