[SOLVED] ZCS 6.0: Postfix does not start

[rty]

After upgrading to ZCS 6.0, my server is in a rather useless state since the MTA does not want to start.
Installation seemed to go smoothly, no error messages. (Debian 4.0, 32-bit)

The zimbra.log file is not very verbose:

zimbra postfix/postqueue[28984]: fatal: Queue report unavailable - mail system is down

How can I debug this?

[rty]

To be more specific:

The last thing I hear from postfix is:

Sep 3 14:45:13 zimbra postfix/postfix-script[15617]: starting the Postfix mail system

But it seems to die pretty much immediately after that.

Please help!

[phoenix]

Originally Posted by rty

zimbra postfix/postqueue[28984]: fatal: Queue report unavailable - mail system is down

How can I debug this?

What's in the log files? Is this NE or OSS? Which operating system? Which version/release of Zimbra did you upgrade from? Did you read the Release Notes for any problems you may encounter?

[rty]

Originally Posted by phoenix

What's in the log files?

I did not see much more than what I posted.
I'll quote it with the context:

Sep  3 14:39:02 zimbra postfix/postqueue[8756]: fatal: Queue report unavailable - mail system is down
Sep  3 14:39:09 zimbra postfix/postfix-script[9010]: warning: not owned by root: /opt/zimbra/postfix-2.6.2.2z/conf/main.cf
Sep  3 14:39:09 zimbra postfix/postfix-script[9011]: warning: not owned by root: /opt/zimbra/postfix-2.6.2.2z/conf/master.cf
Sep  3 14:39:10 zimbra postfix/postfix-script[9015]: warning: not owned by postfix: /opt/zimbra/data/postfix/./spool
Sep  3 14:39:10 zimbra postfix/postfix-script[9018]: warning: not owned by postfix: /opt/zimbra/data/postfix/./spool/maildrop/6388427D
E8
Sep  3 14:39:10 zimbra postfix/postfix-script[9020]: warning: not owned by postfix: /opt/zimbra/data/postfix/./spool/maildrop/1794327E
A7
Sep  3 14:39:10 zimbra postfix/postfix-script[9021]: warning: not owned by postfix: /opt/zimbra/data/postfix/./spool/maildrop/2C6C927E
BB
Sep  3 14:39:10 zimbra postfix/postfix-script[9022]: warning: not owned by postfix: /opt/zimbra/data/postfix/./spool/maildrop/1A0FF27E
2B
Sep  3 14:39:10 zimbra postfix/postfix-script[9038]: starting the Postfix mail system
Sep  3 14:39:32 zimbra postfix/postqueue[9664]: fatal: Queue report unavailable - mail system is down
Sep  3 14:39:36 zimbra zmmailboxdmgr[9739]: status requested
Sep  3 14:39:36 zimbra zmmailboxdmgr[9739]: status OK
Sep  3 14:39:37 zimbra zmmailboxdmgr[9797]: status requested
Sep  3 14:39:37 zimbra zmmailboxdmgr[9797]: status OK
Sep  3 14:40:01 zimbra CRON[9831]: (pam_unix) session opened for user zimbra by (uid=0)
Sep  3 14:40:01 zimbra CRON[9833]: (pam_unix) session opened for user zimbra by (uid=0)
Sep  3 14:40:01 zimbra CRON[9834]: (pam_unix) session opened for user zimbra by (uid=0)
Sep  3 14:40:01 zimbra CRON[9836]: (pam_unix) session opened for user zimbra by (uid=0)
Sep  3 14:40:01 zimbra CRON[9834]: (pam_unix) session closed for user zimbra
Sep  3 14:40:01 zimbra postfix/postqueue[9849]: fatal: Queue report unavailable - mail system is down
Sep  3 14:40:01 zimbra zimbramon[9832]: 9832:info: 2009-09-03 14:40:01, QUEUE: 0 0

(and yes, of course I fixed the ownership issues, to no avail - and Zimbra was very fast with recreating them).

What other log files should I look at? You pointed at 4.5 log files, but I am not sure they are all still the same with 6.0

Originally Posted by phoenix

Is this NE or OSS?

OSS.

Originally Posted by phoenix

Which operating system?

as I said, Debian 4.0 32bit

Originally Posted by phoenix

Which version/release of Zimbra did you upgrade from?

5.0.18 GA Release

Originally Posted by phoenix

Did you read the Release Notes for any problems you may encounter?

I looked at the release notes, yes. What are you referring to?

Cheers,

Rainer

[phoenix]

Originally Posted by rty

I looked at the release notes, yes. What are you referring to?

Specifically P22 about checking your certificates expiry and the openLDAP changes, did you review both of those items?

[rty]

Originally Posted by phoenix

Specifically P22 about checking your certificates expiry and the openLDAP changes, did you review both of those items?

thanks for looking into this.

Yes, everything seems fine - the relevant passwords are the same:
zmlocalconfig -s | grep password
ldap_amavis_password = XYZ
ldap_nginx_password = XYZ
ldap_postfix_password = XYZ
ldap_replication_password = XYZ
ldap_root_password = XYZ
mailboxd_keystore_base_password = zimbra
mailboxd_keystore_password = ***(unique)
mailboxd_truststore_password = changeit
mysql_root_password = ***(unique)
zimbra_ldap_password = XYZ
zimbra_logger_mysql_password = (empty)
zimbra_mysql_password = ***(unique)

I just also set the zimbra user's password to something non-empty, just to be sure.

However, the problem persists.

(side remark: you have a typo in the release notes: "chance the passwords" should read "change the passwords")

So - where else can I look to see why postfix refuses to start?
It is not the most memory-hungry application in the world, is it?

[Enake]

I have the same environment...

Debian 4.0 32 bit and I upgraded today from 5.0.18 to 6.0...

Everything went smooth...

Can you do this and see what happens?

As user zimbra:

zmcontrol stop

As root:

#/opt/zimbra/libexec/zmfixperms -extended

As zimbra:

zmcontrol start

As root:

#/opt/zimbra/libexec/zmfixperms -extended

Hope this helps.

Regards.

[rty]

thanks, but no, this does not help...

anything else?

does postfix really die that quietly? is there no other log where it would leave any traces?

[rty]

BTW, my certificate is still valid (just a few months old, a commercial one).

[brian]

can you post the output of zmcontrol status. are you sure you don't have any OS installed versions of postfix running?