postfix crash. Where is the log?

[partypooper]

Hi!

I had a working Zimbra 4.03 installation running on Debian. It was serving a main mail domain and several virtual mail domains. It was working fine, but I had to reinstall the machine, so I did a big tgz of the /opt/zimbra and reinstalled: I installed a fresh copy of zimbra 4.03 and removed the /opt/zimbra directory and copied my backup instead. The rest of the machine configuration has stayed the same.

Everything is working ok except for one important thing: postfix!
If I send mail within the main mail domain, everything is fine, but if I want to receive mail sent to a virtual domain, I see in the zimbra log: com.sun.mail.smtp.SMTPAddressFailedException: 550 relay not permitted
However, the virtual mail domain is correctly put in the zimbra administration console, so I am lost here.
This error msg makes postfix crash.

I do not see any additional info in the various logs; i checked the permissions of redo.log and did a chown -R zimbra.zimbra of /opt/zimbra/index and /opt/zimbra/store but that does not fix the problem.
Is it a permission wrongly set somewhere? Where can I have a detailed log of what is happening to postfix?

Thanks for your help!

[partypooper]

Hi, i am still struggling with my postfix problem and I tried this:

zimbra@web:~/bin$ postfix upgrade-configuration
SAFETY: editing main.cf, setting unknown_local_recipient_reject_code=450.
See the LOCAL_RECIPIENT_README file for details.


Is it normal? If not, how do I correct that?

Thanks,

Patrick

[partypooper]

I found in /var/log/zimbra.log the following when I launch zimbra:
Nov 30 21:12:13 web postfix/postqueue[2441]: fatal: Queue report unavailable - mail system is down

and later on:
Nov 30 21:12:34 web amavis[2853]: Creating db in /opt/zimbra/amavisd/db/; BerkeleyDB 0.29, libdb 4.2
Nov 30 21:12:36 web postfix/postfix-script: warning: not owned by root: /opt/zimbra/postfix-2.2.9/conf/main.cf
Nov 30 21:12:36 web postfix/postfix-script: starting the Postfix mail system
Nov 30 21:12:36 web postfix/master[2883]: fatal: bind 0.0.0.0 port 25: Address already in use
Nov 30 21:12:38 web clamd[2785]: Protecting against 79289 viruses.

and finally:
Nov 30 21:14:05 web postfix/postqueue[3241]: fatal: Queue report unavailable - mail system is down
Nov 30 21:14:07 web zimbramon[3138]: 3138:info: 2006-11-30 21:14:01, STATUS: web.clicodrome.eu: antispam: Running
Nov 30 21:14:07 web zimbramon[3138]: 3138:info: 2006-11-30 21:14:01, STATUS: web.clicodrome.eu: antivirus: Running
Nov 30 21:14:07 web zimbramon[3138]: 3138:info: 2006-11-30 21:14:01, STATUS: web.clicodrome.eu: ldap: Running
Nov 30 21:14:07 web zimbramon[3138]: 3138:info: 2006-11-30 21:14:01, STATUS: web.clicodrome.eu: logger: Running
Nov 30 21:14:07 web zimbramon[3138]: 3138:info: 2006-11-30 21:14:01, STATUS: web.clicodrome.eu: mailbox: Running
Nov 30 21:14:07 web zimbramon[3138]: 3138:info: 2006-11-30 21:14:01, STATUS: web.clicodrome.eu: mta: Stopped
Nov 30 21:14:07 web zimbramon[3138]: 3138:info: 2006-11-30 21:14:01, STATUS: web.clicodrome.eu: snmp: Running
Nov 30 21:14:07 web zimbramon[3138]: 3138:info: 2006-11-30 21:14:01, STATUS: web.clicodrome.eu: spell: Running


so I checked, and yes, indeed, I had something running on port 25 (the debian was shipped with exim4, that I had aparently incompletely removed). I feel really stupid (I also had a DNS problem on top of that).

Anyway, this experience has raised the following questions that remain unanswered:

1) why is there many log files, split accross the file system. The doc count 3 (/opt/zimbra/log/zimbra.log, /var/log/zimbra.log, and a 3rd one for tomcat).
Is it impossible to merge them? At least, put an information in one log (I am thinking /opt/zimbra/log/zimbra.log) reminding that additional info could be found at <and list here the other log files>!
2) I am sure I am not the only one that gets ports conflict with the smtp server. Port conflicts are correctly indicated for the web servers but not for the smtp server. This should be nothing to do (I guess). When I do a zmcontrol status, I would like to have something that would have told me: mta not running: port conflict.

3) looking at the logs now, I found this:
Nov 30 21:43:38 web postfix/smtpd[9538]: initializing the server-side TLS engine
Nov 30 21:43:38 web postfix/smtpd[9538]: warning: cannot get private key from file /opt/zimbra/conf/smtpd.key
Nov 30 21:43:38 web postfix/smtpd[9538]: warning: TLS library problem: 9538:error:0B080074:x509 certificate routines:X509_check_private_key:key values mismatch:x509_cmp.c:389:
Nov 30 21:43:38 web postfix/smtpd[9538]: cannot load RSA certificate and key data

should I get worried? How do you regenerate this?

Anyway, keep on the good job guys, zimbra needs just a little bit of polishing for being perfect.

Time for a good night sleep (I actually spent all night long yesterday trying to fix those issues)...

Pat

[jholder]

Hope you had a good nights sleep.
I could use a good nap now.

Please see this thread for SSL Cert problems:
http://wiki.zimbra.com/index.php?tit...icate_Problems

Good luck,
john