omshowmn: [OM 16065] Possible user database error

[easysoft]

Hi All,

A running Scalix 11.4.5 installation on CentOS 5 x86_64 suffered a disk failure, and the data was recovered from the RAID5 by some data recovery dudes.
I have rebuilt the server using some new disks, but when I copy over the recovered '/var/opt/scalix' and then attempt to start Scalix I get this :
# /etc/init.d/scalix start
Starting Scalix services (anas): omshowmn : [OM 16065] Possible user database error
[ OK ]

and scalix isn't running properly :
# /etc/init.d/scalix status
*** MyServer
PC Monitor Started NON-STOP 0
Directory Relay Server Started 18:23:09
Notification Server Started 18:23:09 0
Shared memory daemon Started NON-STOP
Notification Monitor Started NON-STOP
Session Monitor Aborted NON-STOP
Indexer Started NON-STOP
Stats Daemon Aborted NON-STOP
Container Access Monitor Started NON-STOP
Item Structure Server Stopped 06.24.10
Database Monitor Started 18:23:09
Licence Monitor Daemon Aborted NON-STOP
LDAP Daemon Started 18:23:09
Queue Manager Started NON-STOP
Item Delete Daemon Started NON-STOP
IMAP Server Daemon Started 18:23:09
SMTP Relay Started 18:23:09
Mime Browser Controller Started 18:23:09
Event Server Started 18:23:09
Service Router Stopped 18:20:18 0
Local Delivery Stopped 18:20:18 0
Internet Mail Gateway Stopped 18:20:18 0
Local Client Interface Disabled 18:20:18 -
Remote Client Interface Disabled 18:20:18 -
Test Server Stopped 18:20:18 0
Request Server Stopped 18:20:18 0
Print Server Stopped 18:20:18 0
Bulletin Board Server Stopped 18:20:18 0
Background Search Service Stopped 18:20:18 0
CDA Server Stopped 18:20:18 0
POP3 interface Stopped 18:20:18 -
Omscan Server Stopped 18:20:18 0
Proxy Folder Service Stopped 18:20:18 0

I have already used omcheck to fix the file permissions.

Any ideas what I can do to try and fix this? Any help would be greatly appreciated.

TIA,
Regards
Arif

[ls-al]

Using "omrc" instead of "/etc/init.d/scalix start" might show additional informations during the start.
Furthermore the fatal log or the filtered output of omshowlog would be needed.

[easysoft]

Thanks for your quick response!
omrc gives the same error.

The fatal log has :

SERIOUS ERROR LDAP Daemon (LDAP Engine ) Thu Jan 6 09:23:45 2011
[OM 16065] The requested Directory does not exist.
Pid of logging process: 9811

ERROR Session Monito(Session Monito) Thu Jan 6 10:34:46 2011
[OM 16065] The requested Directory does not exist.
Pid of logging process: 10813

SERIOUS ERROR Administration(Scalix Admin C) Thu Jan 6 10:34:46 2011
[OM 16065] The requested Directory does not exist.
Pid of logging process: 10830

SERIOUS ERROR Stats Daemon (Stats Daemon ) Thu Jan 6 10:34:46 2011
[OM 16065] The requested Directory does not exist.
Pid of logging process: 10832

SERIOUS ERROR Licence Monito(Licence Monito) Thu Jan 6 10:34:46 2011
[OM 16065] The requested Directory does not exist.
Pid of logging process: 10842

omshowlog just gives an error when I run it :
# omshowlog
[OM 4871] Incomplete record in the log files.

Any ideas?

Again your help is much appreciated.
Regards
Arif

[ls-al]

This looks like the database files containing the directories are not fully recovered. This is critical for the operation of the server backend.

By default, a Scalix installation has 3 directories: SYSTEM, USERLIST and FREEBUSY. The systems knows about these from /var/opt/scalix/??/s/sys/dir.index.

Code: Select all

[root@sys0 sys]# cat dir.index
5
C   4   1000
H   0   USERLIST         1   VISTA     1/1/vi_dir   1293115825   0         4.0
S   0   SYSTEM   D      2   VISTA  1/2/vi_dir   1293115837   604800         4.0
S   0   FREEBUSY         3   VISTA     1/3/vi_dir   1293115852   0         4.0


This is referring to the vista_DB files in the /var/opt/scalix/??/dir/ structure.

I'd start to see if these files and directories are complete.
Example:

Code: Select all

[root@sys0 1]# pwd
/var/opt/scalix/s0/s/dir/1
[root@sys0 1]# tree
.
|-- 1
|   |-- dir_1024.dat
|   |-- dir_128.dat
|   |-- dir_128.key
|   |-- dir_16.dat
|   |-- dir_16.key
|   |-- dir_24.key
|   |-- dir_256.dat
|   |-- dir_32.dat
|   |-- dir_32.key
|   |-- dir_4.key
|   |-- dir_40.key
|   |-- dir_512.dat
|   |-- dir_64.dat
|   |-- dir_64.key
|   |-- dir_8.key
|   |-- dir_keys.dat
|   |-- vi_dir.dbd
|   `-- vi_dir.lock
|-- 2
|   |-- dir_1024.dat
|   |-- dir_128.dat
|   |-- dir_128.key
|   |-- dir_16.dat
|   |-- dir_16.key
|   |-- dir_24.key
|   |-- dir_256.dat
|   |-- dir_32.dat
|   |-- dir_32.key
|   |-- dir_4.key
|   |-- dir_40.key
|   |-- dir_512.dat
|   |-- dir_64.dat
|   |-- dir_64.key
|   |-- dir_8.key
|   |-- dir_keys.dat
|   |-- vi_dir.dbd
|   `-- vi_dir.lock
`-- 3
    |-- dir_1024.dat
    |-- dir_128.dat
    |-- dir_128.key
    |-- dir_16.dat
    |-- dir_16.key
    |-- dir_24.key
    |-- dir_256.dat
    |-- dir_32.dat
    |-- dir_32.key
    |-- dir_4.key
    |-- dir_40.key
    |-- dir_512.dat
    |-- dir_64.dat
    |-- dir_64.key
    |-- dir_8.key
    |-- dir_keys.dat
    |-- vi_dir.dbd
    `-- vi_dir.lock

3 directories, 54 files


Consistency check ist possible with dbcheck (from/opt/scalix/diag)

Code: Select all

[root@sys0 1]# pwd
/var/opt/scalix/s0/s/dir/1/1
[root@sys0 1]# dbcheck vi_dir
db_VISTA Version 3.10
Database Consistency Check Utility
Copyright (C) 1985, 1986, 1987 by Raima Corporation
...


hth

[easysoft]

Thank you very much for your help - your post gives me somewhere to start in trying to fix this mess!

Regards
Arif

[mikethebike]

you could also try an omdiropt, and rekey all the entries...it may show where the problem lies.

Mick