Migration from 11.x -> 11.x fails

Discuss installation of Scalix software

Moderators: ScalixSupport, admin

hollo
Posts: 10
Joined: Sat Aug 25, 2007 10:13 am

Migration from 11.x -> 11.x fails

Postby hollo » Sat Aug 25, 2007 11:35 am

Hi i have some serious problems with a scaling migration,

My goal is to move a already existing version (11.x) of scaling to a new computer (11.x).. but i have some serious problems... i hope that people can help me.. here is what i have done.. then i dont know if i have made anything wrong... this i what i have read i could do in this forum..

First I better make clear that i have changed the IP-addr, but the hostname is the same..

This is my "recipient":

I will first explain what i do in text, afterwards exactly my scripting, what i have done..

1. i make a clean install
2. i test the new install
3. everything works just perfect

4. i shut down all processes that has to do with scalix
5. i copy my new clean install to a bck dir
6. i override my /var/opt/scalix with my old install
7. i change ownership of my old installed dir, so that scalix owns everything
8. i run ompatchom
9. i start scalix up as it was before

summary:
POP3 - OK
Everything works apparently

IMAP - FAILURE
IMAP recognize my login but drop the connection right away

Webmail - FAILURE
I cannot login with the message wrong password or username

SAC - OK
I can login with the password from the other machine, and watch all users from that system, and see that they have a lot of mails

Thats all.. now my descripting EXEACTLY as i did it.. i really hope that someone can help me out here..


After my install i tested the installation if it did work.. and everything appears to be just fine.. and the webmail/sac works perfect..
hollo@wonder:~$ telnet 10.1.1.82 110
Trying 10.1.1.82...
Connected to 10.1.1.82.
Escape character is '^]'.
+OK Scalix POP3 interface ready on: aalzh.rdsas.com
user sxadmin
+OK sxadmin gets mail here
pass test
+OK User Signed on: sxadmin intray has 0 messages
list
+OK Listing
.
quit
+OK POP3 Server connection terminated
Connection closed by foreign host.
hollo@wonder:~$ telnet 10.1.1.82 143
Trying 10.1.1.82...
Connected to 10.1.1.82.
Escape character is '^]'.
* OK Scalix IMAP server 11.1.0.10849 ready on aalzh.rdsas.com
a login sxadmin test
a OK LOGIN completed, now connected to aalzh.rdsas.com
^]
telnet> Connection closed.
hollo@wonder:~$

Now i wanted to stop the current running scalix..

Connection to 10.1.1.82 closed.
hollo@wonder:~$ ssh root@10.1.1.82
root@10.1.1.82's password:
Last login: Sat Aug 25 16:49:28 2007 from 10.1.1.115
[root@aalzh ~]# omshut
Disabling 14 subsystem(s).
omoff : [OM 4835] Subsystem Item Structure Server is already off.
Disabling 8 subsystem(s).
[root@aalzh ~]# /etc/init.d/scalix stop
[root@aalzh ~]# /etc/init.d/scalix-tomcat stop
Stopping Tomcat service (aalzh): [ OK ]
[root@aalzh ~]# ps aux | grep scalix
root 4675 0.0 0.0 4680 396 ? Ss 16:50 0:00 /opt/scalix/bin/ldapmapper --pidfile /var/run/ldapmapper.pid
postgres 9865 0.0 0.0 20180 2052 ? S 16:54 0:00 /usr/bin/postmaster -i -h aalzh.rdsas.com -p 5733 -D /var/opt/scalix/ah/postgres/data -k /var/opt/scalix/ah/postgres/data
root 10461 0.0 0.0 3912 620 pts/1 S+ 17:09 0:00 grep scalix
[root@aalzh ~]# /etc/init.d/ldapmapper stop
Stopping Scalix LDAP mapper (ldapmapper): [ OK ]
[root@aalzh ~]# /etc/init.d/scalix-postgres stop
Stopping scalix-postgres service (aalzh): [ OK ]
[root@aalzh ~]# ps aux | grep scalix
root 10577 0.0 0.0 4872 620 pts/1 S+ 17:10 0:00 grep scalix
[root@aalzh ~]#

now no processes with the name of scalix seems to be running... so now i want to copy my old install to a bck dir.. and copy my old install to my /var/opt/scalix path:

[root@aalzh ~]# cd /var/opt/
[root@aalzh opt]# cp -r scalix /tmp/scalix_old
[root@aalzh opt]# rm -rf scalix
[root@aalzh opt]# cp -r /root/scalix /var/opt/
[root@aalzh opt]# chown -R scalix:scalix /var/opt/scalix
[root@aalzh opt]# ompatchom
ompatchom: Scalix instance running
[root@aalzh opt]# omshut
Disabling 17 subsystem(s).
omoff : [OM 4835] Subsystem Item Structure Server is already off.
Disabling 8 subsystem(s).
[root@aalzh opt]# ompatchom
ompatchom: Running patch-uninstall scripts ...
ompatchom: Running patch-preinstall scripts ...

###########################################################################
#
# ompatchom: Initializing Scalix components.
#
###########################################################################

ompatchom: Copying Scalix components, please wait
# Uncomment the following lines to enable the Submission and LMTP listeners
#LMTP=ON
# This listener is only active if LMTP=ON is above
[LMTP]

###########################################################################
#
# Check and clear out all old directory change logs if one or more
# directory requires conversion. To initiate a resync request, use:
# omresyncds -d <directory_name>
#
###########################################################################

No directories require conversion

###########################################################################
#
# Convert all hidden directories without password
#
###########################################################################

No directories require conversion

###########################################################################
#
# Convert all shared directories without password
#
###########################################################################

No directories require conversion

###########################################################################
#
# Check any other directory with password not converted.
# Manual conversion instruction using omcnvdirs utility:
# omcnvdirs -d <name> -t <type> -p <password> -a ''
#
###########################################################################

No directories require conversion

Updating conversion and language specific configuration files.
This may take some time, Please Wait ...


Updating data for new Scalix version. Please wait ...

About to copy values for some new SYSTEM directory attributes.
Searching for SYSTEM entries with ENTRY-TYPE=1/PRIMARY-S=/(UL-AUTHID= | UL-CAPS= | UL-IL=) ...
No update required for attribute UL-AUTHID/UL-CAPS/UL-IL.

About to generate values for some new SYSTEM directory attributes.
Searching for local entries without attribute GLOBAL-UNIQUE-ID ...
No update required for attribute GLOBAL-UNIQUE-ID.
Searching for local entries without attribute HOST-FQDN ...
No update required for attribute HOST-FQDN.

About to upgrade all Bulletin Board Access Control Lists (BB ACLs).
No upgrade required, file /var/opt/scalix/ah/s/acl/bb/upgrade.fin already exists.

About to update GUIDs in USERLIST to match those in SYSTEM directory.
Update GUID for authid: sxadmin
[OM 16961] Modified 1 entry in the Directory
Update completed, or modification is not required.


###########################################################################
#
# ompatchom: Scalix Initialization Complete
#
###########################################################################

ompatchom: Running patch-postinstall scripts ...
NOTE: The above output is also contained in the
file /var/opt/scalix/ah/s/sys/install/11.1.0.10849.log
[root@aalzh opt]# /etc/init.d/scalix start
Starting Scalix services (aalzh): [ OK ]
[root@aalzh opt]# /etc/init.d/scalix-tomcat start
Starting Tomcat service (aalzh): [ OK ]
[root@aalzh opt]# /etc/init.d/scalix-postgres start
Starting scalix-postgres service (aalzh): [FAILED]
[root@aalzh opt]# /etc/init.d/postgresql start
Starting postgresql service: [ OK ]
[root@aalzh opt]# /etc/init.d/scalix-postgres start
Starting scalix-postgres service (aalzh): [ OK ]
[root@aalzh opt]# omshut

Now things should be back to the old system... so i continue to check the setup... and i do the following:

hollo@wonder:~$ telnet 10.1.1.82 110
Trying 10.1.1.82...
Connected to 10.1.1.82.
Escape character is '^]'.
+OK Scalix POP3 interface ready on: aalzh.rdsas.com
user sxadmin
+OK sxadmin gets mail here
pass <SECRET>
+OK User Signed on: sxadmin intray has 357 messages
^]
telnet> Connection closed.
hollo@wonder:~$ telnet 10.1.1.82 143
Trying 10.1.1.82...
Connected to 10.1.1.82.
Escape character is '^]'.
* OK Scalix IMAP server 11.1.0.10849 ready on aalzh.rdsas.com
a login sxadmin <SECRET>
a NO LOGIN failure, service unavailable
* BYE IMAP Server not available
Connection closed by foreign host.
hollo@wonder:~$ telnet 10.1.1.82 143
Trying 10.1.1.82...
Connected to 10.1.1.82.
Escape character is '^]'.
* OK Scalix IMAP server 11.1.0.10849 ready on aalzh.rdsas.com
a login sxadmin badpassword
a NO LOGIN failure, user name or password rejected
^]
telnet> Connection closed.
hollo@wonder:~$

The Webmail interface still continue to fail login.. but the Admin (SAC) interface is successfull to let me in.

now as you can see what is happening is that when i log in to my mail folder with IMAP, i just get thrown away... i really dont know how to debug on that.. but please somebody help me... i dont even know where i can find any usefull logging.

Clemens
Posts: 68
Joined: Fri Dec 23, 2005 3:20 pm

Postby Clemens » Sun Aug 26, 2007 6:11 am

Hi,

the way I normally approach it is the following:

1. Install a new OS on another computer.
2. Create the Scalix user/group in the system, with the same IDs (important!).
3. Copy the old /var/opt/scalix as TAR to the new server (tar preserves ownerships).
4. Extract Scalix to the place it should be.
5. Install the new Scalix, it should detect the old message store and simply upgrade it.

Works for me without any problems, hope I didn't forget anything!

Clemens

jaime.pinto
Scalix Star
Scalix Star
Posts: 709
Joined: Fri Feb 23, 2007 6:50 pm
Location: Toronto - Canada

Postby jaime.pinto » Sun Aug 26, 2007 8:21 am

This sequence of events is for RHE4 to RHEL5, is just a little bit more elaborated. Other OS's may be very similar:

1) Stop Scalix on the old server (omshut -t3)
1a) optional - stop port forwarding port 25 through your firewall (or the old server) incoming mail should be queued up on the secondary server
2) Stop PostgreSQL on the old server (/etc/init.d/scalix-postgres stop)
3) Run (or rerun) the rsync of /var/opt/scalix. Tar is also ok.
4) Power down the old server (or change its name/IP to something else)
5) Disable eth1 from the new server, if you had a "backdoor". Leave only *one* interface (very important).
6) Change the IP and Hostname for the new server to match the old server. Put the system back on the proper network, with access to DNS and the internet in general.
7) Install Scalix on the new server.
7a) you may have to remove PostgreSQL from the original OS and install it again from scalix if there is version conflict. Be prepared to run the install/config script more than once.
7b) I suspect that when you install Scalix, it will create a new user, with the ID of 100, but if doesn't, create it yourself. If not, you'll see that many services will not start.
8.) Verify that all services are correctly running (omstat -s && omstat -a)
8a) Depending on which services fail, you can fix it with "omcheck -s d >> /tmp/perms.txt && bash /tmp/perms.txt". This will create a script which will fix all of the permissions. It will also take a very long time to run. Please be patient with it.
9) Test connectivity from a desktop client via sac and SWA (Windows and/or Linux)
9a) If you turned off port 25 in step 1a, restart that.
10) Test incoming and outgoing email with a test to a yahoo or Gmail, or whatever account. Test with the connector as well.
11) Reinstall/transfer the SSL certificates from the old server onto the new.
12) fix the https:// redirect as in the old server. Newer OS often change things around quite a bit. Be prepared to check httpd.conf line by line for discrepancies.
13) Test https on port 443, as well as SSL imap/smtp on ports 465/993
14) If you have problems check the install/config for stunnel. You may need a initscript.
15) update sendmail to include amavis and other milters as you had before.
16) remove the extra domains that are included in the license keys (via sac), but we don't use yet.
17) Hah, reboot the server and check if everything is normal when it boots up

We had a lot of assistance from Kev at Digital Adrenaline when we did our own upgrade.
If you notice steps on your installation that we missed, in particular problem situations, please insert them to the list.
Image Jaime
||||||||||||||||||||||||||||||||||||||||

hollo
Posts: 10
Joined: Sat Aug 25, 2007 10:13 am

Postby hollo » Mon Aug 27, 2007 2:55 am

Clemens wrote:Hi,

4. Extract Scalix to the place it should be.

Clemens


hmm.. what exactly do you mean by Extract Scalix? when i go run the installer... then i cannot start up httpd.. it will fail with some missing directories.. also i have found some missing directories in /opt/scalix/global..

so what do you mean exactly by point 4?

jlae

Postby jlae » Wed Aug 29, 2007 5:43 am

Hi folks,

After our server crashed and re installing with our backups, we ran into to permission related problems, that are mostly fixed now.

Only one problem remains, like our friend hollo has posted here: the imap daemon
accepts connections but returns a failure.

Excerpt from the telnet conversation with the imap daemon:

Code: Select all

. login someone@somewhere.tld password
. NO LOGIN failure, service unavailable
* BYE IMAP Server not available


Any help would be much appreciated, since without imap, there are no folders and no webmail for anyone.


By the way, I ram omcheck -d -s and it reports some wrong permissions and ownerships, but I'm not too keen to change those without at least knowing if that might be the problem.

Thanks in advance,
Joao

jlae

Postby jlae » Wed Aug 29, 2007 7:51 am

I have news.

I've managed to fix the problems by trying to login in the console and checking the omshowlog -s imap -F 00:00 output.

It complained about some permission denied, and I found out that the write permission for the group was missing on some files.

So, hollo, try to check your omshowlog -s imap -F 00:00 and chmod g+w <file> for each file that the imap service complains about.

Hope that helps,
Joao

obrodkin
Posts: 69
Joined: Mon Oct 16, 2006 9:22 pm
Location: Fremont CA

Postby obrodkin » Tue Sep 04, 2007 10:41 pm

I assume that in step 3) you are rsync-ing /var/opt/scalix from old to a new box? Then, during scalix installation it should recognize old DB? Please clarify. Thanks

William
Posts: 314
Joined: Fri Jun 02, 2006 8:28 am
Location: British Isles

Postby William » Fri Sep 14, 2007 9:01 am

any chance someone could http://www.scalix.com/wiki a fairly definitive set of instructions for transfering a scalix installation from one server to another?

KiwiStag
Posts: 42
Joined: Thu Mar 23, 2006 5:48 am
Location: New Zealand

Migrating from 11.1 to 11.4

Postby KiwiStag » Sat Nov 29, 2008 3:52 am

My old server was Fedora core 5 with Scalix 11.1 & tried to move to Centos5 box with 11.4
Followed above, copied the /var/opt/scalix folder & contents across & installed Scalix 11.4 on the Centos box. It ran through fine & as soon as it had I changed permissions as the group/user number differs from the Fedora box.

I can't start scalix & when I even try ompatchom it states Scalix isn't installed.
I also had to manually put on JRE 1.5.15 as 1.6 was previously installed on the CentOs box via RPM.

I also get the error:
Inconsistency found between version files in sys and lib directories. This means that the patch levels of the Scalix shared library and binary areas do not match. Scalix cannot be started until this is solved.
(hence why trying to run ompatchom)

KiwiStag
Posts: 42
Joined: Thu Mar 23, 2006 5:48 am
Location: New Zealand

Post subject: Migrating from 11.1 to 11.4 - semi-resolved

Postby KiwiStag » Sat Nov 29, 2008 5:08 am

After uninstalling scalix modules & running ompatchom, everything seems better. It still seems however that none of the web services are working however...

Issue with httpd config script for Tomcat (re web services)
see: viewtopic.php?t=10121&highlight=


Return to “Installation”



Who is online

Users browsing this forum: No registered users and 3 guests