Results 1 to 3 of 3

Thread: Install hung on upgrade/update

Hybrid View

  1. #1
    Join Date
    Mar 2007
    Rep Power

    Question Install hung on upgrade/update

    I am trying to move a Zimbra installation from one machine to another. The OS in both cases is identical (Debian/Etch).

    The process I have followed thus far (taken from relevant wiki entries) is:
    * Install base OS and prerequisites
    * Configured a local bind9 install to operate for split DNS (the new machine
    * Install Zimbra, from the same install set that was last used to upgrade the source machine
    * Remove /opt/zimbra
    * Replace /opt/zimbra with the latest backup of the current system
    * Run zmfixperms (in case uid/gids and perms had been upset in backup/restore)

    At this point all seemed well (the services seemed to start without error) but both the normal web interface ( and the admin interface ( responded with 404-page-not-found errors from Jetty.

    Following instructions for fixing problems in similar situations (foud by searching this forum) I reran the install/upgrade procedure. This worked as expected (it identified that it was "upgrading" 5.0.11 to 5.0.11) until right at the end when it stalled. The last screenful of information is:
    Setting up syslog.conf...done.
    You have the option of notifying Zimbra of your installation.
    This helps us to track the uptake of the Zimbra Collaboration Suite.
    The only information that will be transmitted is:
            The VERSION of zcs installed (5.0.11_GA_2695_DEBIAN4.0)
            The ADMIN EMAIL ADDRESS created (
    Notify Zimbra of your installation? [Yes]
    Notifying Zimbra of installation via
    Notification complete
    Starting servers...done.
    Checking for deprecated zimlets...done.
    Checking for network zimlets in LDAP...done.
    Removing network zimlets...
    Finished removing network zimlets.
    Installing common zimlets...
    Finished installing common zimlets.
    Upgrading Document templates...done.
    Restarting mailboxd...done.
    Setting up zimbra crontab...done.
    Looking at the running processes in pstree, the relevant line seems to be
    and that mv process shows up in pa ax as
     6758 pts/0    S+     0:00 mv /tmp/auth_keys.tmp /opt/zimbra/.ssh/authorized_keys
    which seems like an odd thing to stall (I've left it in that state overnight, so it has been doing nothing for 10 hours now).

    The files /tmp/auth_keys.tmp and /opt/zimbra/.ssh/authorized_keys both exist and contain identical content. I can now login to both the main user interface and the admin interface fine, though I've not tried sending/receiving mail yet.

    I am assuming that I'm safe to kill the currently stuck process, but do I need to rerun or anything else at this point?

  2. #2
    Join Date
    Mar 2007
    Rep Power


    A quick update:
    I killed the "mv" process, and the setup process seemed to continue to completion:
    Setting up zimbra crontab...done.
    Moving /tmp/zmsetup.01072009-002553.log to /opt/zimbra/log
    Configuration complete - press return to exit
    I'm assuming that this means all is well. I will give it a proper test tonight when the sub-domain MX entry I've added for testing purposes has had time to propagate.

  3. #3
    Join Date
    May 2006
    Rep Power


    One thing to check will be the admin console UI, navigate around and see if a 'remoteMmanager' error comes up - troubleshooting for that:
    Quote Originally Posted by mmorse View Post
    Substitute your port number in the below instructions:

    1. Check /etc/sshd_config and be sure it's set to 22 (or the port you're using)
    2. stop/start/restart sshd
    /etc/init.d/ssh restart
    3. su - zimbra
    4. Be sure zimbra's ssh port is set to 22 (or the port you're using)
    zmprov ms zimbraRemoteManagementPort 22
    5. Generate new ssh keys
    cd /opt/zimbra/bin/
    6. Deploy the keys
    To test this, you can run the following command:
    ssh -vi .ssh/zimbra_identity -o strictHostKeyChecking=no -p 22 (or the port you're using)
    -Shouldn't prompt you for a password at this point...then check the admin console mail queues area/servers/certificates.
    If you continue to get that error please see this doc for some more ideas Mail Queue Monitoring - Zimbra :: Wiki

Similar Threads

  1. install oss fc7 not run well
    By epelaez in forum Installation
    Replies: 1
    Last Post: 03-05-2008, 02:26 PM
  2. Replies: 21
    Last Post: 09-27-2007, 11:49 AM
  3. INSTALLATION PROBLEM ON Centos 4.3 x_86-64Bit.
    By in forum Installation
    Replies: 11
    Last Post: 07-09-2007, 08:09 AM
  4. Replies: 16
    Last Post: 11-29-2006, 09:36 AM

Posting Permissions

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