Results 1 to 10 of 14

Thread: mta, logger and logswatch won't start - also opendkim issue

Hybrid View

  1. #1
    Join Date
    Oct 2012
    Posts
    4
    Rep Power
    3

    Default mta, logger and logswatch won't start - also opendkim issue

    Hi,

    After a clean install, I was able to start all services on Zimbra 8.0. After shutting down and restarting services, I get:

    zimbra@neutron:/opt/zimbra$ zmcontrol start
    Host neutron.thesoniccloud.com
    Starting ldap...Done.
    Starting zmconfigd...Done.
    Starting logger...Failed.
    Starting logswatch...failed.


    Starting convertd...Done.
    Starting mailbox...Done.
    Starting antispam...Done.
    Starting antivirus...Done.
    Starting snmp...Done.
    Starting spell...Done.
    Starting mta...Failed.
    Starting saslauthd...done.
    opendkim: /opt/zimbra/conf/opendkim.conf: ldap://neutron.thesoniccloud.com:389/?DKIMSelector?sub?(DKIMIdentity=$d): dkimf_db_open(): Connect error
    Failed to start opendkim: 0
    zmopendkimctl failed to start


    Starting stats...Done.
    zimbra@neutron:/opt/zimbra$


    I believe my DNS settings are OK and as I mentioned I was able to start all services previously. Is the OpenDKIM error the problem? I didn't see anything too noteworthy in zimbra.log.

    I hope someone can help me solve this.

    Thanks.

    -Jon

  2. #2
    Join Date
    Oct 2012
    Posts
    6
    Rep Power
    3

    Default

    I experienced the exact same problem last week and desperately tried the last two days to resolve it.
    ZCS 7.2 upgrade to 8.0.0 went flawless and 8.0.0 was running and was completely usable. A little more checking on the logs made me execute the fixes for mysql stats stuff and some time later the zimbra restart was resulting in the same error message as above. The logs arent telling much and raising the loglevels did not point me to problems either.
    LDAP server is running and acessible, DNS is ok, no iptables. What could this be?

    It appears the LDAP query that the opendkim startup is printing is incomplete but I am not into LDAP enough to figure it. As I am not planning on actually using dkim - how could I completely disable it and make the mta service start at least again?

    This is a true show stopper


    Maybe someone who has it running ok could post the output of "zmlocalconfig | grep -i ldap" for some checking?
    Last edited by Ash74; 10-09-2012 at 02:06 AM. Reason: addition

  3. #3
    Join Date
    Oct 2012
    Posts
    4
    Rep Power
    3

    Cool MTA Logger and Logswatch Won't Start

    Does anyone have a solution to this?

    Is it a problem with DNS or OpenDKIM or something else. This is costing major downtime for my company's email server. It started with the update to Zimbra 8.0 and is a massive showstopper.

  4. #4
    Join Date
    Nov 2011
    Location
    Hungary
    Posts
    30
    Rep Power
    4

    Default

    after the upgrade many files and directory permission will be wrong, use the zimbra permission fixer script for reset all permission:
    with root user:
    /opt/zimbra/libexec/zmfixperms -verbose -extended

    then restart zimbra
    it may help

    run this after every manual config edit.
    also check your language collation settings.
    zimbra services may fail to start if its wrong!

    login to zimbra user
    echo $LANG
    echo $LC_ALL

    you need to see 'C' in both output.
    if you see en_US.UTF-8 for example, that will be problem.

    the most easy way to check the LDAP server:
    telnet <YOUR-SERVER-IP-ADDRESS> 389

    example: telnet 192.168.1.20 389

    if you see this:
    Trying 192.168.1.20...
    Connected to 192.168.1.20.
    Escape character is '^]'.

    then your LDAP server is up and running
    Last edited by vadonka; 10-09-2012 at 06:15 PM.

  5. #5
    Join Date
    Oct 2012
    Posts
    6
    Rep Power
    3

    Default

    I only have problems in getting opendkim to properly work which in turn blocks a proper zmmta start.

    LDAP connectivity is there and perfect. Adding and removing keys using "/opt/zimbra/libexec/zmdkimkeyutil" works just fine, so the structure on LDAP server is there as well.

    From the error above:
    "ldap://neutron.thesoniccloud.com:389/?DKIMSelector?sub?(DKIMIdentity=$d)" seems to be missing something between "389/" and "?DKIMSelector" but as I am not into LDAP much and found no docs on how to get e.g. phpldapadmin to access a local Zimbra tree - I am stuck.

    Fixing permissions was done before and didn't help as well.

    Something that puzzles me and may be related - the 8.0 Release Notes note to enter

    "zmldapupgrade -b 66387"

    after upgrade. The output is:


    --------------
    com.zimbra.cs.account.ldap.upgrade.LdapUpgrade -b 66387
    --------------
    unrecognized bug number


    Could that be related? How could I disable opendkim within Zimbra or how can the db_connect be debugged?

  6. #6
    Join Date
    Oct 2012
    Posts
    4
    Rep Power
    3

    Lightbulb Any Progress

    What distro of linux are you using? I've tried everything including a clean install. One thing I notice is that after the install, all services start up properly. However, if I shut down and then restart, that's when mta won't start, nor logger or logswatch. Should I try Ubuntu 10.04 or perhaps use the 32 bit version?


    Quote Originally Posted by Ash74 View Post
    I experienced the exact same problem last week and desperately tried the last two days to resolve it.
    ZCS 7.2 upgrade to 8.0.0 went flawless and 8.0.0 was running and was completely usable. A little more checking on the logs made me execute the fixes for mysql stats stuff and some time later the zimbra restart was resulting in the same error message as above. The logs arent telling much and raising the loglevels did not point me to problems either.
    LDAP server is running and acessible, DNS is ok, no iptables. What could this be?

    It appears the LDAP query that the opendkim startup is printing is incomplete but I am not into LDAP enough to figure it. As I am not planning on actually using dkim - how could I completely disable it and make the mta service start at least again?

    This is a true show stopper


    Maybe someone who has it running ok could post the output of "zmlocalconfig | grep -i ldap" for some checking?

  7. #7
    Join Date
    Oct 2012
    Posts
    6
    Rep Power
    3

    Default

    We are running CentOS 6.3 64-bit on VMWare ESXi 5.1 (would have to check for the patchlevel).

    I recognized some oddities on the lags when searching for things:

    Code:
    ./messages-20121007:Oct  4 15:56:22 draven kernel: opendkim[22564]: segfault at 1f0 ip 000000369b009220 sp 00007fff1ecfd7c8 error 4 in libpthread-2.12.so[369b000000+17000]
    ./messages-20121007:Oct  5 12:16:13 draven kernel: opendkim[23686]: segfault at 1f0 ip 000000369b009220 sp 00007fffcd4afb78 error 4 in libpthread-2.12.so[369b000000+17000]
    These messages came up after initial installation of 8.0 when it was working fine, but before a reboot (I think) indeed. I have no idea wether this is related nor wether these error messages would reoccur whenever opendkim starts again.

  8. #8
    Join Date
    Oct 2012
    Posts
    4
    Rep Power
    3

    Default Zimbra 8.0 MTA Failure

    I was told to disable skim in the configs. I did this. But even with the disablement, it still seems to look for it. So perhaps there are other places to disable. Interesting that you are running in a virtualized environment. Could that be part of the problem? Have you tried not running virtual? Also, did you see similar problems running earlier versions of zimbra such as 7.2 in the same environment?

    Quote Originally Posted by Ash74 View Post
    We are running CentOS 6.3 64-bit on VMWare ESXi 5.1 (would have to check for the patchlevel).

    I recognized some oddities on the lags when searching for things:

    Code:
    ./messages-20121007:Oct  4 15:56:22 draven kernel: opendkim[22564]: segfault at 1f0 ip 000000369b009220 sp 00007fff1ecfd7c8 error 4 in libpthread-2.12.so[369b000000+17000]
    ./messages-20121007:Oct  5 12:16:13 draven kernel: opendkim[23686]: segfault at 1f0 ip 000000369b009220 sp 00007fffcd4afb78 error 4 in libpthread-2.12.so[369b000000+17000]
    These messages came up after initial installation of 8.0 when it was working fine, but before a reboot (I think) indeed. I have no idea wether this is related nor wether these error messages would reoccur whenever opendkim starts again.

  9. #9
    Join Date
    Oct 2012
    Posts
    6
    Rep Power
    3

    Default

    I was told to disable skim in the configs.
    How and where did you do that?

    I don't have the possibility to test on a real physical machine currently and did not have problems on ESXi 5.1 with ZCS 7.2 at all, but we didn't use it excessively either.

    One thing that still puzzles me and seems unanswered in all Zimbra context:
    Where from does that LDAP connect string get its "path"? opendkim LDAP notes has examples for SigningTable and KeyTable containing "proper" ou/cn strings. How come Zimbra doesn't need them? Wherefrom are they inherited?

    Why is disabling this "feature" nowhere documented?

Similar Threads

  1. logswatch is not starting
    By dhull in forum Installation
    Replies: 3
    Last Post: 02-05-2010, 06:16 PM
  2. logswatch.out and zmlogger.out BIG!
    By LMStone in forum Administrators
    Replies: 12
    Last Post: 05-09-2008, 07:45 AM
  3. zimbra logger not start
    By utdom in forum Administrators
    Replies: 1
    Last Post: 12-17-2007, 10:24 PM
  4. Logger won't start
    By InternetGuy in forum Administrators
    Replies: 0
    Last Post: 05-10-2007, 11:25 PM
  5. Cannot start logger or snmp
    By hikenboots in forum Administrators
    Replies: 5
    Last Post: 01-16-2006, 02:39 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •