error during upgrade from 10 to 11

[eddieedwards]

I am running the installer from the CLI and it chokes with a fatal error during installation of the scalix_server rpm. The first time I tried, the log showed:

Code: Select all

2006-12-23 12:52:12,624 ERROR Operation failed: Install /bin/rpm -U -v --percent /root/scalix-11.0.0-GA/software/scalix_server/scalix-postgres-11.0.0.653-1.noarch.rpm /root/scalix-11.0.0-GA/software/scalix_server/scalix-tomcat-connector-11.0.0.130-1.rhel4.i386.rpm /root/scalix-11.0.0-GA/software/scalix_server/scalix-tomcat-5.5.16-140.noarch.rpm /root/scalix-11.0.0-GA/software/scalix_server/scalix-platform-11.0.0.653-1.noarch.rpm         /root/scalix-11.0.0-GA/software/scalix_server/scalix-sis-11.0.0.653-1.noarch.rpm /root/scalix-11.0.0-GA/software/scalix_server/scalix-mobile-11.0.0.653-1.noarch.rpm /root/scalix-11.0.0-GA/software/scalix_server/scalix-server-11.0.0.431-1.rhel4.i386.rpm /root/scalix-11.0.0-GA/software/scalix_server/scalix-res-11.0.0.653-1.noarch.rpm /root/scalix-11.0.0-GA/software/scalix_server/scalix-swa-11.0.0.653-1.noarch.rpm /root/scalix-11.0.0-GA/software/scalix_server/scalix-sac-11.0.0.653-1.noarch.rpm: (10) [        'Preparing packages for installation...', 'scalix-server-11.0.0.431-1.rhel4', 'Stopping Scalix LDAP mapper (ldapmapper): [FAILED]\r', 'Starting Scalix LDAP mapper (ldapmapper): [  OK  ]\r', 'error: %post(scalix-server-11.0.0.431-1.rhel4.i386) scriptlet failed, exit status 3', 'scalix-tomcat-connector-11.0.0.130-1.rhel4', 'scalix-tomcat-5.5.16-140', 'scalix-postgres-11.0.0.653-1', 'scalix-platform-11.0.0.653-1', 'scalix-sis-11.0.0.653-1', 'scalix-mobile-11.0.0.653-1', 'scalix-res-11.0.0.653-1', 'Remote Execution Service (RES) successfully installed.', 'Please configure Remote Execution Service (RES) using the', 'Scalix Installer for configuration.', 'Have fun!', 'scalix-swa-11.0.0.653-1', 'Please configure Scalix Web Access (SWA) using the Scalix Installer for configuration.', 'SWA minimally requires Tomcat 4.1.29 and JDK 1.4.2+.', 'After configuring SWA you may launch SWA within Tomcat with <Tomcat_installed_dir>/bin/startup.sh.', 'To check any startup errors look at <Tomcat_installed_dir>/logs/scalix-swa_log.<date>.txt', 'scalix-sac-11.0.0.653-1', 'Scalix Administration Console (SAC) successfully installed.', 'Please configure Scalix         Administration Console (SAC) using', 'the Scalix Installer for configuration.', 'Have fun!']
2006-12-23 12:52:12,624 CRITICAL Operation failed: Installation of packages failed
2006-12-23 12:52:12,624 ERROR Operation failed: Aborting installation thread.
2006-12-23 12:52:12,625 INFO Install end


I figured I would try running the install program again, but this time, the only component available to upgrade was Scalix Server. I went ahead and ran into a fatal error again. From the log:

Code: Select all

2006-12-23 12:56:50,722 INFO Starting installation of RPM files [('/root/scalix-11.0.0-GA/software/scalix_server/scalix-server-11.0.0.431-1.rhel4.i386.rpm', 'scalix-server')]
2006-12-23 12:56:53,847 ERROR Operation failed: Install /bin/rpm -U -v --percent /root/scalix-11.0.0-GA/software/scalix_server/scalix-server-11.0.0.431-1.rhel4.i386.rpm: (2) ['Preparing packages for installation...','\tpackage scalix-server-11.0.0.431-1.rhel4 is already installed']
2006-12-23 12:56:53,848 CRITICAL Operation failed: Installation of packages failed
2006-12-23 12:56:53,848 ERROR Operation failed: Aborting installation thread.
2006-12-23 12:56:53,848 INFO Install end


So rpm knows the scalix-server is installed, but the installer dies anyway--and doesn't think it is installed when I re-run the installer script. It looks like everything should be OK anyway, but I want to bve sure I'm not missing anything.

I started X on the box and ran the graphical installer, but got the same error. When the installer checks for installed components it thinks the upgrade needs to be done, but the rpm installer finds it and so an error is reported. I will try removing the rpm and running through it again.

I will keep digging, but if anyone can help, I would appreciate it.

Mike

[eddieedwards]

I tried uninstalling the rpm, but then got an error during the network check of the next install:

Code: Select all

2006-12-23 13:54:06,759 ERROR Operation failed: Network check failed -- Hosts file is not configured with fully qualified domain name (FQDN)
2006-12-23 13:54:06,768 ERROR Operation failed: Network check failed -- Invalid FQDN '/opt/scalix/bin/omcheckgc: error while loading shared libraries: libom_sis.so: cannot open shared object file: No such file or directory'
2006-12-23 13:54:06,780 ERROR Operation failed: Network check failed -- Could not find a host name in DNS that matches an IP address
2006-12-23 13:54:06,781 ERROR Operation failed: Network check failed -- Could not resolve host name '/opt/scalix/bin/omcheckgc: error while loading shared libraries: libom_sis.so: cannot open shared object file: No such file or directory' in DNS
2006-12-23 13:54:06,793 INFO Network interface is  eth0
2006-12-23 13:54:09,021 INFO Apache 2.x appears to be installed.
2006-12-23 13:54:09,021 INFO Operation passed: Dependency check passed.
2006-12-23 13:54:09,401 INFO Operation passed: Running services check passed.
2006-12-23 13:54:09,402 INFO System check end
2006-12-23 13:54:09,404 INFO System check result: fail


It looks like a file is missing after I uninstall the rpm. So I tried reinstalling it manually. Then I get this:

Code: Select all

rpm -U -v software/scalix_server/scalix-server-11.0.0.431-1.rhel4.i386.rpm
Preparing packages for installation...
omcheckgc: error while loading shared libraries: libom_sis.so: cannot open shared object file: No such file or directory
scalix-server-11.0.0.431-1.rhel4
warning: /etc/init.d/scalix saved as /etc/init.d/scalix.rpmsave
Stopping Scalix LDAP mapper (ldapmapper): [FAILED]
Starting Scalix LDAP mapper (ldapmapper): [  OK  ]
error: %post(scalix-server-11.0.0.431-1.rhel4.i386) scriptlet failed, exit status 3


And then rpm -qa shows two instances of scalix-server are installed.

[eddieedwards]

OK, I think something is happening during the postinstall scriptlet in the rpm. Here are the contents of ominstall.log:

Code: Select all

Update man config: MANPATH  /opt/scalix/share/man
Configuration not complete, re-run "ompatchom -a"
Cannot open Message catalog
You must load at least one message catalog
and have your enviroment set correctly
before installing or running Scalix


That doesn't sound good. I have a feeling that the server rpm is being installed, but not all the proper things are happening during the postinstall scriptlet, so that the scalix installer chokes.

[eddieedwards]

I think I've narrowed it down to a problem with omsendin not working and throwing the "can't open Message Catalog" error as shown above.

[eddieedwards]

Somehow, after I removed the manually-installed server rpm, the installer didn't bomb. But then, while it was trying to set up the web components, it choked complaining about the version of java I'm running:

Code: Select all

2006-12-24 14:58:10,933 CRITICAL Operation failed: Couldn't determine the version of Java installed in .
The output of java -version command is included below:
2006-12-24 14:58:10,933 ERROR Operation failed: Aborting web apps configuration


Tried the "reconfig" route in the installer and got the same problem. Here's what java -version reports:

Code: Select all

java version "1.4.2"
gcj (GCC) 3.4.6 20060404 (Red Hat 3.4.6-3)
Copyright (C) 2006 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.


Am I the only one with these problems? It will be a Christmas miracle if I get this back up and running before Wednesday!

[eddieedwards]

A little help here?

If I don't get any feedback by noon today, I'm going to try the Uninstall option in the installer and try to install all over again. I hope the message store and users are preserved. My fear is that the message store is currently in some partially-migrated state and it will get corrupted when I try the uninstall/reinstall.

[ScalixSupport]

We found the problem to be that /var/opt/scalix/temp was mounted on a seperate mount point. Removing the mount point and creating /var/opt/scalix/temp allowed the installation to complete. Once installation was successful we created /var/opt/scalix/instance_initials/s/temp.

I've entered a bug and asked for the release notes to be updated.

Regards,
Don

[mito]

I'm running into this problem right now, and I only have one partition on this system.

However, this is on a system where I have been testing (and failing, many many times) to upgrade from Scalix 10 to Scalix 11. I am currently getting this error after uninstalling everything Scalix related, and then trying to do a Scalix 10 install with my old data in the /var/opt/scalix dir.

[eddieedwards]

I am currently getting this error after uninstalling everything Scalix related, and then trying to do a Scalix 10 install with my old data in the /var/opt/scalix dir.

You mean "trying to do a Scalix 11 install" right? What is your OS? Can you post the pertinent parts of the log?

[mito]

You mean "trying to do a Scalix 11 install" right? What is your OS? Can you post the pertinent parts of the log?

No, I meant Scalix 10... it's complicated and totally off-subject for this thread, but this is on a CentOS 4.4 (RHEL4) system. This is a test system, so that I don't screw up my production server. I had a clean install of the OS, restored the backup from my production system (which is a FC4 system... changing from FC4 to CentOS during this upgrade) and then tried to install Scalix10 to get it to a functional system, and the plan was to then upgrade to Scalix 11. Well, that plan died many deaths. I came across this error multiple times through my process though. The most current occurrence was after removing everything, I tried to do a clean Scalix 10 install, and then got this message.

I would have posted the relevant log portion, but between my first post and now, I have done a format/reload on the system and am currently redownloading the scalix installers. (I think I did something early-on that screwed me up, and I want to try it from scratch again).

A guess though after thinking about it some more, you might try making sure that the /var/opt/scalix/temp dir is created and has the appropriate permissions (the permissions on my production Scalix 10 box for that dir are drwxrwx--x ).

Good luck... I'll help if I can, but I'm not the best Scalix guru yet...

[eddieedwards]

The Scalix folks fixed me up. I was going to try to help you out, but it looks like you're doing a clean install already. Good plan.

[kanderson]

Can you post what was done to fix the problem so that others searching this thread can resolve it for themselves?

I haven't hit it yet, but I suspect I likely will...

[dkelly]

There wasn't much that needed "repairing".

The biggest hurdle is to make sure that the pre-upgrade /var/opt/scalix was moved to /var/opt/scalix/NN/s

Make sure that /etc/opt/scalix/instance.cfg contains the correct details for the new location.

I've added this to the FAQs at http://www.scalix.com/wiki/index.php?title=Scalix_11_FAQs

When you know that instance.cfg matches with the message store location, you can either move /etc/opt/scalix/[caa|res|webmail] to /var/opt/scalix/NN/ or you can uninstall and re-install the components using the installer.

Cheers

Dave