DNS not working in ISPconfig3 cluster

Discussion in 'Installation/Configuration' started by archange, Sep 6, 2010.

  1. archange

    archange New Member


    Thank you for having me on the Forums.
    I am a Linux beginner, however, I have followed the "HowtoForge" guides with great interest and finally I decided to use ISPconfig3 for my hosting needs.

    I installed ISPconfig3 on Centos 5.5 x64 using this guide: http://www.howtoforge.com/perfect-server-centos-5.5-x86_64-ispconfig-3-p3

    Additionally, I needed to start running my own DNS and I proceeded adding to ISPconfig3 two more servers, configured as slaves.

    I have registered ns1, ns2 and ns3 with my registrar (not GoDaddy though), according to this guide:


    After they have propagated, I added the DNS zones in the control panel. The next day, when issuing the following command I get:

    # dig @ns1.arenahosting.org any arenahosting.org
    ; <<>> DiG 9.3.6-P1-RedHat-9.3.6-4.P1.el5_4.2 <<>> @ns1.arenahosting.org any arenahosting.org
    ; (1 server found)
    ;; global options:  printcmd
    ;; connection timed out; no servers could be reached

    I can see no errors in ISPconfig3 panel, nor at systems' innitialisation.

    Can anyone please help me out?

    Thank you for your patience, time and understanding.
  2. damir

    damir New Member

    What DNS software are you using and is it running? Do you have firewall, maybe port 53 is closed?
  3. archange

    archange New Member

    Servers are running BIND9, chrooted. Firewall is disabled atm. I just restarted the named service and it appears to be running fine.
  4. till

    till Super Moderator Staff Member ISPConfig Developer

    Please make sure that you configured BIND to start at boot time. The command is:

    chkconfig --levels 235 named on
  5. archange

    archange New Member

    I remember doing that during the setup and the service is indeed running; I just rebooted the server once again - this is the bind status output:

    Is it just me, or does Bind not read the DNS configuration from ISPconfig?
  6. till

    till Super Moderator Staff Member ISPConfig Developer

    You should check that the configuration directory and named.conf file path of your bind installation match the bind config settings in ispconfig.

    and if you want to test your local dns server, use this command:

    dig @localhost arenahosting.org

    Dont use @ns1.arenahosting.org as the nameserver domain then is a subdomain of the domain that you query, so you have to add glue records and a failure does not indicate if it is a problem with your dns server or if you have not setup the gluue records at the domain registry correctly.
    Last edited: Sep 6, 2010
  7. archange

    archange New Member

    Thanks Till, I think you have pointed me into the right direction with your reply:
    These were the instructions I have followed during configuration of Bind:

    I changed:

    To these to get it right:
    Last edited: Sep 7, 2010
  8. archange

    archange New Member

    OK, after the changes, I added the DNS zones again and this time I can see them:
    Also, on localhost, the dig command shows correct output:
    Question is, how and when do the zones transfer to the additional name servers? I have also modified the Bind9 file paths in their respective entries under ISPconfig3 and did a service restart.

    Also, the example that I have followed showed the addition of Slave zones - as far as I understand, this is no longer the case in ISPconfig3 v.
    Is it OK to set the slave servers' IPs under "DNS Zone":
    Last edited: Sep 7, 2010
  9. falko

    falko Super Moderator ISPConfig Developer

    Zone transfers start immediately after a zone has changed on the master.

  10. archange

    archange New Member


    Now, on server3 i get the following error:
    And another question: in ISPconfig3 panel, under DNS Zones, is it normal for zones to be displayed only under "Server1"? When I pick "Server2" and "Server3" I get no records. Do I have to add entries manually for these aswell and if so, could you tell me what need I to do?

    Thank you!
  11. till

    till Super Moderator Staff Member ISPConfig Developer

    1) Make sure that you selected server 2 and 3 to be a mirror of server 1 in the ispconfig server settings if you want that records get mirrored to the other servers in the cluster automatically.

    2) ou added incomplete zones for arenahosting.org and dtxigaming.com. Both zones have no NS records. To fix this, add NS records for these zones. I recommend that you use the dns wizard the next time when you create zones as the wizard creates all nescessary records for a zone automatically.
  12. archange

    archange New Member

    Till, what I mean is that I was expecting to see the DNS records added to the other servers; I do not need any of the additional servers to mirror the master.

    I did use the wizard - I know you have repeatedly suggested this in the past to other forum members. And I have the NS records in place.


    I found the problem: somehow, while playing around the DNS control panel I managed to defer partial Zone entries to the 3rd server, prior to the correct creation of NS entries. Named was unhappy about the results and refused to start.
    I solved the problem by removing the truncated /var/named/chroot/etc/named.conf.local and recreating a new empty file, according to the original installation instructions. After that named was happy to start and now I see DNS working properly across the cluster.
    Last edited: Sep 7, 2010
  13. till

    till Super Moderator Staff Member ISPConfig Developer

    How should that happen if you did not configure the servers to mirror the records? Iif a record shall show up on another server, you will have to set the server to be a mirror of the master.

    That ok, but according to the error message, there wre at least ns records missing or a short periiod of time or you missed to enable the mirroring and added a zone on one of the slaves manually without ns records.
  14. archange

    archange New Member

    Thanks Till, you are right; I was just editing the above post while you added your reply. I took the liberty of removing the screenshot.
    And regarding the DNS records, I think there's a difference in the way I perceived the ISPconfig3 interface and the way that it actually works. Anyway, I'm getting used to how it works and I'm starting to get quite fond of it :)
    Last edited: Sep 7, 2010
  15. archange

    archange New Member

    In the server monitor section, how can I get past this error?

    Data from: ????-??-?? ??:??
    No data about the services available at the moment. Please check again later.
  16. till

    till Super Moderator Staff Member ISPConfig Developer

    Thats not an error. Only data for services that are installed on a specific node gets gathered and some of the data gets collected only once a day, so it may take until data shows up or no data may show up at all if a specific service is not installed on a node.
  17. archange

    archange New Member

    I see. Thing is that no monitoring stats show up on that specific machine (server2), although services are all installed. And this has been going on since I first configured ISPconfig as slave on that particular machine - should date at least a couple of weeks back. At least Raid space would be nice to monitor, not to mention logs...

    I figured that probing must somehow be blocked and that's why I was asking for help.

    And on a different note, now that DNS and hosting is running, I need to also get emailing to work. Shall we continue the discussion here, or on another topic?
    My problem is this:
    Seems Postfix has no msql support?

    Reinstalled postfix from the Centos Plus repositories, although I also did so the first time. Sending mail now, but not receiving; not even inside the domain.

    Now I get this error:

    My ISP has port 25 open for the IP; firewall and SELinux are turned off.

    Telnet to localhost returns:
    Telnet to mail.dtxigaming.com results in:
    OK, so I figured iptables was blocking port 25, so I chose to add a standard firewall rule in ISPconfig 3, which did the job. Now port 25 is open and I got flooded with replies to my external mailboxes saying:
    I guess I have to change the relayhost settings in /etc/postfix/main.cf
    Can anyone help me to change it so as to reflect the multiple domains created in ISPconfig 3?

    Falko, Till, please help.
    Last edited: Sep 9, 2010
  18. archange

    archange New Member

    I found a similar problem in a post and both of you (Till & Falko) have suggested to run an ISPconfig update. So i did. And it got me back to:

    postfix/qmgr[26418]: fatal: unsupported dictionary type: mysql
    I'm stuck in a loop. :eek:
    Last edited: Sep 9, 2010
  19. till

    till Super Moderator Staff Member ISPConfig Developer

    You installed a postfix package (binary) which does not support mysql. Please install your server exactly as described in the perfect server guide for ISPConfi´g 3 and the linux distribution that you use to ensure that all installed software components match the requirements for this setup.
  20. archange

    archange New Member

    No, Till, that was not the case. Please read carefully.

    I installed using this tutorial, step 11: http://www.howtoforge.com/perfect-server-centos-5.5-x86_64-ispconfig-3-p4

    11 Install Postfix With MySQL Support
    The "normal" Postfix package from the CentOS repository doesn't have MySQL, but the Postfix package from the CentOS Plus repository does. Therefore we remove Postfix...
    yum remove postfix
    ... and install it again, this time from the CentOS Plus repository:
    yum install postfix
    Then turn off Sendmail and start Postfix and MySQL:
    chkconfig --levels 235 mysqld on
    /etc/init.d/mysqld start
    chkconfig --levels 235 sendmail off
    chkconfig --levels 235 postfix on
    /etc/init.d/sendmail stop
    /etc/init.d/postfix start
    After I finished installing ISPconfig3 I got this, as posted above:
    Then I redid Step 11 in the tutorial, unistalled Postfix, reinstalled from Centos Plus repos, and I got the error, also posted above:
    So I started searching for a solution on the Howto Forums and I found this one:

    In which u advised user onastvar to to the following:
    I did the same:
    php -q update.php
    I also chose to reconfigure services. And this landed me straight back to square 1:

    So what am I suppose to do next? "Cause I'm really lost over here... :confused:
    Last edited: Sep 9, 2010

Share This Page