Scalix Forums

Hello,

I try to migrate from Scalix 10.0.0.175 running FC4 to RHEL4. Therefore i have a brandnew machine here, installed with a plain RHEL without X or anywthing. My Problem is: the migration from FC4 to RHEL doesn't seem to work clean:

My strategy was this:

I installled version 10.0.0.175 to my new RHEL and copied /var/opt from oldsystem to the new system.

I ran omcheck -s -d to correct permissions (copyed using rsync but without -a parameter)

When starting, Scalix Server won't start because of inconsistence between version in sys and lib directories. So I copyed /opt/scalix to the new System.
When Scalix now starts, it complains about wrong Version of libssl, so I created a symlink from /lin/libssl.0.4 to /lib/libssl0.5 and restart the machine.

Now scalix "hangs" on booting and doesn't start up, no error message.
But, because V. 10.0.0.175 is a bit outdated I wanted to upgrade to 10.0.1.3.

Scalix recognizes the Mailfolder and upgrades. But the installer hangs and can't install the scalix-server rpm, because it conflicts with the 10.0.0.175 rpm. Using --force to the rpm installs also doen't work (scriptlet failed, exit status 3)

running ompatchom gives me the following output:

ERROR: [OMINST 225] Could not add queue or acl for service SMINTFC.

And here I stuck.

Is there a better way to migrate from FC4 to RHEL and upgrading to newwest scalix Version?

Kurt

FYI: selinux and firewall are diabled, so this can't be the problem

I'm now on a clean System, reinstalled whole server ...

After copyed the /var/opt directoy from oldserver to new server I installed Scalix 10.0.1 Which works without Problems.

Next I ran omcheck -s -d and corrected file permissons.

Scalix won't start up so I executed ompatchom

But the error is

[OM 15785] Internal error detected - see the log files.
ERROR: [OMINST 225] Could not add queue or acl for service SMINTFC.

The command omsearch -e ou1=* gives me all my users but the command

omshowu -m all gives
omshowu : [SYS 13] /var/opt/scalix/sys/omlocks file system error number 13

hmmmm

I chowned -R /var/opt/scalix/sys to scalix:scalix.
and the error was away.
Next I need to run ompatchom again and this time there was no error message.
Sendmail was rebuilded and needed a long time to get started but in the end the script ended succesful.

Scalix was now starting up (wooot) but I can't access webinterface or sac. Outlook gives me a permission denied when connecting to the server. (And after that asking for passwd)

So I chowned the whole /var/opt/scalix directoryx to scalix but this also didn't helped. I uninstalled scalix10.

And installed it again, but this didn't helped. When booting up the sendmail service needs a really long time to boot up but seems to start.

Now here I'm, stucked in the middle

I'm back again.

Needed 40 minutes to figure out, that hostname --fqdn just delivers the hostname, not the whole domainname because the line "search my.domain" was missing in /etc/resolv.conf.

The next problem (this was a long weekend though) is:

Ic an login to sac as sxadmin and see my users there, but the users can't login through webmail. The Username is wrong. resetting the username in sac doens't show up anything.

... tomorrow at 6'clock I will have this problem solved, but I will report it anyway

Got it.

But very very unclean.

After digging around I found the IMAP logs and pulled verbosity to high. There I found the errors "permission denied" to various lock and config files. I removed all *.lk and *.lock files and then i ran chmod -R a+rw /var/opt/scalix.

Its working now, but this is not a good solution. Perhaps Scalix support knows any more hints when waking up

Greetz Kurt

Well, I think if you had searched throgh these forums, you could have saved you a lot of trouble.

I'd add the following tips to summarize:
- before copying over the message store, make sure user/group scalix are created on the target system with the SAME UID and GID
- copy the message store before running the installer. ompatchom is executed as part of it, it fixes up version IDs and things.
- there should be no permission problems. If there are, omcheck should help you.
- you can copy /etc/opt/scalix, but probably shouldn't.
- you should never copy anything in /opt
- DNS/hostname configuration needs to be correct.
- you should check for filenames beginning with a capital "O" under /var/opt/scalix/sys - those files would be renamed before-upgrade config files.

Hope this helps,
Florian.

Hello Florian,

thanks a lot for reply. Although this is my first migration and I'm fairly new to Scalix Administration (Old System was setup by another person) I have to say: I learned a lot.

Thanks a lot for your summarize, hope this will help some people.


greets from Berlin

Kurt

I'm currently trying to move the OM store to a newer server as the older one is starting to become unreliable. The original server had a name that wasn't consistent with the server group and the new server has a different name.

After moving the store & installing the server, it stopped - the code is below from the dump:


Any suggestions?

I assume the server should be up and running anyway, admin console, etc., will not work.

- delete the sxqueryadmin user



Restart the installer, uninstall the webapplications (leave scalix server installed) and reinstall them again.

should work.

attention, you might need to use sxmodfqdn and edit a couple of config files to compensate for your hostname change.

Florian.

florian wrote:attention, you might need to use sxmodfqdn and edit a couple of config files to compensate for your hostname change.

There is now a howto on the wiki - http://www.scalix.com/wiki/index.php?ti ... ChangeFQDN - for doing this.

Cheers,

Chris