Problem backing up Exchange 2003 on w2003

[tekkanet]

legato 7.2 server on Linux (build 172).
Legato 7.2 client (build 172) on W2003 with 4.1 module for exchange(build 4.1.0.148)
manually started backup with saveset MSEXCH:MB

06/27/05 20:07:11 savegrp: command 'nsrxchsv -s bcksrv.peviani.it -g mailserver -LL -m mailserver -t 1119052983 -l incr -q -W 78 -N MSEXCH:MB MSEXCH:MB ' for client mailserver exited with return code 227.
* mailserver:MSEXCH:MB 1 retry attempted
* mailserver:MSEXCH:MB [3176] 06/27/05 18:45:55 nsrxchsv_ese.exe Version: 4.1.0.148 Supporting Exchange 2000 and Exchange 2003
* mailserver:MSEXCH:MB [3176] 06/27/05 18:45:55 System Version: 5.2 Build 3790
* mailserver:MSEXCH:MB [3176] 06/27/05 18:45:55 ESEBCLI2.DLL Version: 6.5.7226.0 Service Pack 1
* mailserver:MSEXCH:MB [3176] 06/27/05 18:45:56 Performing an Exchange save
operation; save level is Incremental (Legato Backup Level: incr).
* mailserver:MSEXCH:MB [3176] 06/27/05 18:45:56 Parsing parameters from the command line...
* mailserver:MSEXCH:MB [3176] 06/27/05 18:45:56 Opening connection to NetWorker Server...
* mailserver:MSEXCH:MB [3176] 06/27/05 18:45:59 Opening connection to Exchange Server...
* mailserver:MSEXCH:MB [3176] 06/27/05 18:45:59 Backing up Microsoft Exchange Private Mailboxes ...
* mailserver:MSEXCH:MB [3176] 06/27/05 18:46:00 Starting: Administrator
* mailserver:MSEXCH:MB [3176] 06/27/05 18:46:34 Success: Administrator 26 MB 801134 bytes/second
...

After some successful mailboxes saved I get:
* mailserver:MSEXCH:MB [3176] 06/27/05 19:41:02 Starting: FOO BAR
* mailserver:MSEXCH:MB [3176] 06/27/05 20:10:53 This program has been asked to exit.
* mailserver:MSEXCH:MB [3176] 06/27/05 20:10:53 Performing orderly shutdown of the backup environment.
* mailserver:MSEXCH:MB [3176] 06/27/05 20:10:53 The mailbox backup received a request to abort while creating temporary file.
* mailserver:MSEXCH:MB [3176] [E] 06/27/05 0:10:55 FOO BAR: Error: status = 0x000003e3.
* mailserver:MSEXCH:MB [3176] [E] 06/27/05 20:10:56 The XBSA transaction was not committed; the save operation did not complete.

In nsrxchsv.log on exchange server I see:
[3176] 06/27/05 20:10:53 main_handler received signal: 9
and in xbsa_messages nothoing particular.
Any hints? How to debug this?
it seems I have not space on disk problems...
Thanks in advance.
Gianluca

 

[NWNinja]

Check the Inactivity timeout of the group, if it´s 0 increase it to 60 or something. Is it only MB that's giving you the problem? Does IS work fine? Does regular scheduled backup of MB work?

Cheers!

 

[nokker]

In Exchange 2003 you do not have permission to read the mailboxes as administrator. You should try to create a special user just for backing up your exchange server. And then set the correct permissions on the information store.

 

[tekkanet]

For NWNinja:
Timeout was 30 and I set it to 60.
I'll try to run a full backup and see what happens.

For nokker:
The permission should not be a problem.
I log in as a special user and infact the error comes out
after successfully back up of other 7 mailboxes.
BTW: if I don't specify the user in client setup window inside nwadmin console, what is the user used by default?
Does the password pass clear through network causing security concerns? Other kind of settings at client level to bypass setting password at server level?
Gianluca

 

[nokker]

Okay.
Try using the Exmerge tool from Microsoft and se if it backup correct.
For further information from Legato take at look in the nsr\applogs folder and look in the nsrxchsv.log which usually provides more debug info than the administator console.

If you don't provide the password I think that the process is started as SYSTEM. Not sure thoug ...

Best Regards
Tommy

 

[NWNinja]

Like nokker said, it actually might be a permission problem.
Could be that the mailbox that causes the failure has some special permissions set. It might not be enough to just log onto the client with a special account, try starting the NetWorker services on the client using an account that has full control to the filsystem as well as Exchange. A domain admin account would be preferable. Also, does the backup fail on the same mailbox everytime? In that case, try to just backuup that mailbox and see if you get some other error message that might help isolate the problem.

Cheers!

 

[nokker]

NWNinja > Pr default the permission in Exchange 2003 is set to deny "Send As" and "Recieve As" for Domain and Enterprise Admins. These permission does somehow prevent the backup from make a connection to the mailboxes. Therefore you are also prevented from making an export if you are running Exmerge with an Administrator account. I use a user account with full permission to the MB and PF store.

Best Regards
Tommy

 

[tekkanet]

The mailbox affected was always the same.
But it is the first MB encountered with about 2Gb of size.
Between the previous ones, successfully backed up, the biggest was 750MB.
I started a session for full backup of only this mailbox and it is running (I changed actual Name Surname with "Problematic User"):

Jun 28 17:52:26 bcksrv logger: NetWorker savegroup: (info) mailserver:MSEXCH:MB/Problematic User: No full backups of this save set were found in the media database; performing a full backup

I think the problem could be related with the size and with the fact that running "dmesg" and also in /var/log/messages
I find many messages like this

Jun 28 17:52:29 bcksrv kernel: st0: Failed to read 65536 byte block with 32768 byte read.

So the backup process is very slow and with 2Gb size it could timeout as NWninja said.
If I backup the same host at a file system level (es C:\) I haven't these kind of problems and the throughput is between 3000KB/s and 7000KB/s (1GB saved in 5 minutes).
So the problem seems limited to exchange agent...

Any particular setup at tape device block size level when I back up windows clients with Linux/Unix servers/storage nodes?
Gianluca

 

[tekkanet]

Ok, now it works.
thanks to all for the tips.
The action was to set inactivity timeout to 120.
It seems I don't get now the message:

This program has been asked to exit...

All MBs are processed.
The only problem is that the process is too slow:
about 16 hours to backup 28Gb of mailboxes (saveset was MSEXCH:MB)
The file system backup, for the same amount of data on the same server, takes 2 hours and a half.
Is it normal this ratio of about 1/6 for exchange mailboxes backup? Is it due to the module inefficience or something else I can tune?
Thanks in advance.
Gianluca