Cent OS Upgrade from 4.8 to 5.5

Discuss the Scalix Server software

Moderators: ScalixSupport, admin

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 4:59 am

Hi Everyone

Here is a good question.

I am in the middle of upgrading my mail server OS from CentOS 4.8 to CentOS 5.5 so that we can use the ActiveSync module. Everything has gone smoothly so far with the exception of the Scalix part of the file system: /var/op/scalix.

Before the upgrade my directory looked like this:

Code: Select all

[root@grumpy scalix]# ls -al
total 64
drwxrwxr-x 10 scalix scalix 4096 Mar 17  2010 .
drwxr-xr-x  3 root          root     4096 Sep 20 01:29 ..
drwxr-xr-x  2 root          root     4096 Feb 20  2010 backup
-rw-r--r--  1 scalix scalix   24 Mar 31  2007 .bash_logout
-rw-r--r--  1 scalix scalix  191 Mar 31  2007 .bash_profile
-rw-r--r--  1 scalix scalix  124 Mar 31  2007 .bashrc
drwxr-xr-x  2 root          root     4096 Mar 17  2010 deleted
drwx------  2 scalix sclaix 4096 Jan 22  2006 .elinks
-rw-r--r--  1 scalix scalix  383 Mar 31  2007 .emacs
-rw-r--r--  1 scalix scalix  120 Mar 31  2007 .gtkrc
drwxrwxr-x 12 scalix scalix 4096 Mar 31  2007 gy
drwxr-xr-x  3 scalix scalix 4096 Mar 31  2007 .kde
drwxr-xr-x  3 root          root     4096 Apr  9  2009 lvmbackup
drwxr-xr-x  3 root          root     4096 Aug 30  2007 root
drwxr-xr-x 10 root          root     4096 Jan 27  2009 rootback
-rw-r--r--  1 scalix scalix  658 Mar 31  2007 .zshrc
[root@grumpy scalix]#


Now it looks like this:

Code: Select all

[root@grumpy scalix]# ls -al
total 64
drwxrwxr-x 10 avahi-autoipd ecryptfs 4096 Mar 17  2010 .
drwxr-xr-x  3 root          root     4096 Sep 20 01:29 ..
drwxr-xr-x  2 root          root     4096 Feb 20  2010 backup
-rw-r--r--  1 avahi-autoipd ecryptfs   24 Mar 31  2007 .bash_logout
-rw-r--r--  1 avahi-autoipd ecryptfs  191 Mar 31  2007 .bash_profile
-rw-r--r--  1 avahi-autoipd ecryptfs  124 Mar 31  2007 .bashrc
drwxr-xr-x  2 root          root     4096 Mar 17  2010 deleted
drwx------  2 avahi-autoipd ecryptfs 4096 Jan 22  2006 .elinks
-rw-r--r--  1 avahi-autoipd ecryptfs  383 Mar 31  2007 .emacs
-rw-r--r--  1 avahi-autoipd ecryptfs  120 Mar 31  2007 .gtkrc
drwxrwxr-x 12 avahi-autoipd ecryptfs 4096 Mar 31  2007 gy
drwxr-xr-x  3 avahi-autoipd ecryptfs 4096 Mar 31  2007 .kde
drwxr-xr-x  3 root          root     4096 Apr  9  2009 lvmbackup
drwxr-xr-x  3 root          root     4096 Aug 30  2007 root
drwxr-xr-x 10 root          root     4096 Jan 27  2009 rootback
-rw-r--r--  1 avahi-autoipd ecryptfs  658 Mar 31  2007 .zshrc
[root@grumpy scalix]#


What I have been able to discover is that the new version CentOS 5.5 add a new default user called "avahi-autoipd" and is in the slot "100" that scalix originally occupied. The scalix user id is now in slot "101" In addition, CentOS 5.5 has added a new group called "ecryptfs" in slot "101" and an additional new group called "avahi-autoipd" in slot "102". The new slot that the scalix group is now in is "103". This is why we see different user/group names in my scalix directory.

Changing these slot numbers to suit me is probably going to cause lots of problems with the OS later, so I guess I would like to fix this once and for all.

What is the best way to approach this? I wanted to be careful about just globally changing ownership of the files in case special permissions have been set for these files I don't know about.

One more question, I am pretty sure i have the mailnode name correct (sbgk) but how do I positively check that with the database? I dont' want the mail server to start hicupping in case this I have put in the wrong mailnode name in and I can fix it before I start the server.

My mail server is currently down right now so a quick answer to this would be most appreciated.

Sincerley,

Sideband Samurai

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 5:31 am

Hi everyone,

I have found this command:

Code: Select all

#pwd <enter>
/var/opt/scalix
#chown -cR --from=avahi-autoipd:ecryptfs scalix:scalix *


The command was updated from this example:

Code: Select all

chown -cR --from=olduser:oldgroup newuser:newgroup *

After looking at the man page for chown, then doing some searches on google, I came up with this solution. This looks like this might solve my issue, but now I want to know of any ramifications to issuing this command or is this a situation that I don't have anything to worry about.

Are there any other permissions with in this directory (/var/opt/scalix) that will require changing?

and I still have my mailnode naming question to answer.

Thanks for your comments.

Sincerely,

SidebandSamurai

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 7:03 am

-- Continuing on with this project,

I did use that command and it looks like my scalix directories all have the correct user and groups.

Here is a partial output of what that command did:

Code: Select all

changed ownership of `gy/s/msgpool/poolctrl' to scalix:scalix
changed ownership of `gy/s/msgpool/48QP' to scalix:scalix
changed ownership of `gy/s/msgpool/3QP' to scalix:scalix
changed ownership of `gy/s/msgpool/64QP' to scalix:scalix
changed ownership of `gy/s/msgpool/93QP' to scalix:scalix
changed ownership of `gy/s/msgpool/88QP' to scalix:scalix
changed ownership of `gy/s/msgpool/12QP' to scalix:scalix
changed ownership of `gy/s/msgpool/1QP' to scalix:scalix
changed ownership of `gy/s/msgpool/66QP' to scalix:scalix
changed ownership of `gy/s/msgpool/85QP' to scalix:scalix
changed ownership of `gy/s/msgpool/59QP' to scalix:scalix
changed ownership of `gy/s/msgpool/98QP' to scalix:scalix
changed ownership of `gy/s/msgpool/86QP' to scalix:scalix
changed ownership of `gy/s/msgpool/19QP' to scalix:scalix
changed ownership of `gy/s/msgpool/24QP' to scalix:scalix
changed ownership of `gy/s/msgpool/97QP' to scalix:scalix
changed ownership of `gy/s/msgpool/8QP' to scalix:scalix
changed ownership of `gy/s/msgpool/62QP' to scalix:scalix
changed ownership of `gy/s/msgpool/74QP' to scalix:scalix
changed ownership of `gy/s/msgpool' to scalix:scalix
changed ownership of `gy/s/bb/netsync.seq' to scalix:scalix
changed ownership of `gy/s/bb/sys.seq' to scalix:scalix
changed ownership of `gy/s/bb/modtimes' to scalix:scalix
changed ownership of `gy/s/bb/ctner.config' to scalix:scalix
changed ownership of `gy/s/bb' to scalix:scalix
changed ownership of `gy/s/req' to scalix:scalix
changed ownership of `gy/s' to scalix:scalix
changed ownership of `gy' to scalix:scalix
changed ownership of `lvmbackup/scalix/.bashrc' to scalix:scalix
changed ownership of `lvmbackup/scalix/.bash_profile' to scalix:scalix
changed ownership of `lvmbackup/scalix/.kde/Autostart/Autorun.desktop' to scalix:scalix
changed ownership of `lvmbackup/scalix/.kde/Autostart/.directory' to scalix:scalix
changed ownership of `lvmbackup/scalix/.kde/Autostart' to scalix:scalix
changed ownership of `lvmbackup/scalix/.kde' to scalix:scalix
changed ownership of `lvmbackup/scalix/.zshrc' to scalix:scalix
changed ownership of `lvmbackup/scalix/.elinks' to scalix:scalix
changed ownership of `lvmbackup/scalix/.gtkrc' to scalix:scalix
changed ownership of `lvmbackup/scalix/.bash_logout' to scalix:scalix
changed ownership of `lvmbackup/scalix/.emacs' to scalix:scalix
changed ownership of `lvmbackup/scalix' to scalix:scalix


So that is fixed, but I don't have the answer to my last questions and i have to get this up very soon.

I think the Nodename is correct but how to I actually verify it?

I am just about ready to mount this in the proper spot, set my mount tab file and reboot. Any input on this question would be most appreciated.

Sincerely,

Sideband Samurai

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 7:47 am

Hi Everyone,

Ok, I decided to take this one step at a time.

I have decided to see if I can bring the server online. The following are the commands I am entering.

1. Start sendmail:

Code: Select all

[root@grumpy log]# service sendmail start
Starting sendmail:                                         [  OK  ]
Starting sm-client:                                        [  OK  ]



So far so good:

Next command is to start the postmaster:

Code: Select all

service scalix-postgres start
<a pause> then this response:
Starting scalix-postgres service (grumpy):                 [FAILED]


I am sure this has something to do with the os upgrade, so how do I troubleshoot this problem, nothing is logged to /var/log/messages

Thanks for the help.

Sincerely,

Sideband Samurai

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 8:07 am

Hi everyone,

I have dug a little bit further and found the postgres logs and here is what they say:

PGSTARTUP.LOG:

Code: Select all

LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down
LOG:  logger shutting down

... and that is all that is in there.

PGACTIVITY.LOG

Code: Select all

 2010-03-05 16:57:48.201 PSTLOG:  logger shutting down
 2010-03-06 16:57:33.634 PSTLOG:  logger shutting down
 2010-03-07 16:57:30.560 PSTLOG:  logger shutting down
 2010-03-08 16:57:48.073 PSTLOG:  logger shutting down
 2010-03-15 22:21:29.306 PDTLOG:  logger shutting down
 2010-04-06 19:22:55.466 PDTLOG:  logger shutting down
 2010-04-06 19:34:17.775 PDTLOG:  logger shutting down
 2010-04-06 19:34:24.734 PDTLOG:  logger shutting down
 2010-04-06 19:34:35.345 PDTLOG:  logger shutting down
 2010-09-19 19:48:09.045 PDTLOG:  logger shutting down
 2010-09-19 20:24:27.813 PDTLOG:  logger shutting down
 2010-09-19 21:00:24.637 PDTLOG:  logger shutting down
 2010-09-20 04:31:08.544 PDTFATAL:  incorrect checksum in control file
 2010-09-20 04:40:46.702 PDTFATAL:  incorrect checksum in control file
 2010-09-20 04:42:53.844 PDTFATAL:  incorrect checksum in control file


You will see that at 9:00pm last night I shut the server down, did the OS upgrade and then you see at 04:31 incorrect checksum in control file.

I will post more when I find more.

Sideband Samurai

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 9:09 am

Well I think this is good news,

I decided to reboot the server and start from a clean reboot. My first problem was I had coded my fstab incorrectly so my server would not even boot! I have managed to fix that (change an 01 to an 02 did the trick).

I was able to successfully get the server running. Not totally, and I don't know how well its working.

I am not able to log into webmail AT all, but I was able to get into SAC and the server looks normal to me here.

With that partial success, I then started outlook, and I was able to log in successfully and look at my inbox. I even successfully sent a test message to my gmail account also successfully.

I have not enabled port 25 on my router to let the flood in if you know what I mean. What will happen if I do this? (maybe it will start receiving mail to?) but I was concerned with mail getting caught up in the system and not getting delivered correctly. If Postgres supports only the webmail portion and the search, I can at least let the users start looking at their email. The webmail is used but not frequently so I can take my time troubleshooting this issue provided that everyone can receive their Email.
So I think the server is about 3/4 up at this point and I want to know, if postgres is not running what happens? does that mean that webmail is down? does that also mean that searching is not available?

Sincerely,

Sideband Samurai

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 11:54 am

I still need help with this problem, my smartphones are not getting any email, and I am not able to get my blackberry which works on IMAP to have messages pushed to it.

Sideband Samurai

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 12:03 pm

Further information for my problem is as follows: As I look at the IMAP daemon I see the following errors:

Code: Select all

ERROR          IMAP Server Da(IMAP server Pr) 09.20.10 08:46:38
[OM 24070 Debug message for Lab Use:
imapSatAuthenticate:Could not register with Session Monitor. usr_Signon() returned SIGNON LIMIT REACHED, errorno: Success
User Name: Harold L. Robinson / sbgk,us/CN=Harold Robinson


These errors repeat though out the whole log with various other users also.

Sideband Samurai

BaldBoy
Posts: 141
Joined: Fri May 19, 2006 12:45 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby BaldBoy » Mon Sep 20, 2010 12:06 pm

To test, and fix, permissions on Scalix's filesystem the best option for you is to run omcheck

Code: Select all

omcheck -s -d > fix.sh
sh fix.sh


This will detect and replace inaccurate permission settings.

BaldBoy
Posts: 141
Joined: Fri May 19, 2006 12:45 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby BaldBoy » Mon Sep 20, 2010 12:08 pm

Postgres is an essential component of the IMAP/MAPI cache part of Scalix.

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 12:48 pm

Thanks BaldBoy, the first stage is running right now. (been running for about 30 min) I will post back with any info I have after your steps have been completed.

I am seeing these warnings come up during this first stage:

Code: Select all

find: /var/opt/scalix/gy/s/temp/0bgclo3.ofd: No such file or directory
find: /var/opt/scalix/gy/s/temp/0bgclo3.ofd: No such file or directory
find: /var/opt/scalix/gy/s/temp/0bgclo3.ofd: No such file or directory
find: /var/opt/scalix/gy/s/temp/0bgcli2.ofa: No such file or directory
find: /var/opt/scalix/gy/s/temp/0bgcli2.ofa: No such file or directory
find: /var/opt/scalix/gy/s/temp/0bgcli2.ofa: No such file or directory
find: /var/opt/scalix/gy/s/temp/0bgclo6: No such file or directory
find: /var/opt/scalix/gy/s/temp/0bgclo6: No such file or directory
find: /var/opt/scalix/gy/s/temp/0bgclo6: No such file or directory


Sideband Samurai

** Edit update, after reading the man page I found out that these errors are normal and is an indicaton that I should shutdown scalix and re-run the command.

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 1:58 pm

@baldboy -> I did as you instructed. Postgres still will not start.

Sideband Samurai

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Mon Sep 20, 2010 7:02 pm

Hi Everyone,

Another update!

I managed to get my imap services to start authenticating. As you remember, the imap service would not authenticate any users, thus shutting out all our smartphones. The last thing I remember doing was on a whim I decided to update my user record on the scalix server, I added my cell phone number, thinking that maybe there was some corruption in that area that deals with user authentication. About 5 minutes later I started getting reports that the phones were getting their email. I also checked the Imap service and its no longer spewing errors.

I am still having problems troubleshooting Postges. Is there a forum post or a document anyone can point me to that will help me with this process? Maybe it will not start because it can not open the database. Though the postgres logs do not indicate that.

Sincerely,

SidebandSamurai

SidebandSamurai
Posts: 236
Joined: Sun Jan 08, 2006 10:57 pm

Re: Cent OS Upgrade from 4.8 to 5.5

Postby SidebandSamurai » Tue Sep 21, 2010 1:47 pm

Hi Everyone,

I still need help with this, as my server is not fully operational.

Thanks.

Sideband Samurai

ls-al
Scalix Star
Scalix Star
Posts: 510
Joined: Tue Jun 29, 2004 8:28 am
Location: Leipzig, Germany
Contact:

Re: Cent OS Upgrade from 4.8 to 5.5

Postby ls-al » Wed Sep 22, 2010 2:00 am

Regarding you Postgres issues: CentOS 4 was using Postgres 7, CentOS 5 is using Postgres 8. The Scalix installer and documentation will not handle the problem of upgrading this database as it is just a cache. In theory, as it is the only place where the sync status for AS is stored.
The easiest way to solve:
Use the installer and remove the DB along with the messaging services (API). Reinstall them with the installer. Finally let your AS-users recreate their sync profiles.

For larger installations read: http://www.scalix.com/wiki/index.php?title=PopulatingPlatformDb


Return to “Scalix Server”



Who is online

Users browsing this forum: No registered users and 9 guests