Novell help

cwilloug

ADSM.ORG Senior Member
Joined
Sep 13, 2006
Messages
388
Reaction score
11
Points
0
Location
North Dakota
Website
Visit site
We have a novell server that starts backups at 6pm, and sometimes does not finish until noon or later, thusly effecting the performance of our email system in the morning hours - causing users to complain.

Would running two tsm schedules on this system help speed things up? (one for the OS, and one for the email folders)

So basically I'm asking..... HHHEEEELLLLPPP!!!!!

Thanks bunches. :)
 
This server takes 18 hours to backup? What mail system are you using GroupWise? You need to go over your TSM logs, cwilloug, and check the speed (bandwidth) been used here, also what settings you have for backing up open files. That is to start.

Most likely you need to do a trace as well.
 
1. Check NIC settings like full duplex or auto setting. The NIC and switch should be matched.
2. Set resouceutilization to a high number, like 5 or 6

This should improve things significantly. If the backup window is still not acceptable, you can setup two schedules as what you stated above.

Our Groupwise backup takes only 1 hour at most for 900+ e-mail accounts.
 
Nope, thats all good, I thought I had fixed this problem, but it has showed up again. In the sched.log there is a 9 hour pause between backing up the os and starting the groupwise folders.

Has anyone else seen this problem? Would removing and reinstalling the client help? I'm just about at the big hammer stage of things with this machine.....
 
Does the 9 hours pause apperas every days?
or just once a while or a special day?

Which value did you set fo the "Commetimeout" parameter on your TSM server?
 
The client does pause every day, anywhere from 6 hours or more, Tuesday nights backup just happened to be 9 hours.

Commetimeout does not sound familiar, I've looked on the node f=d, and in the dsm.opt file, where is that set? Or is that similar to idle timeout? idle time out is set to 60 minutes, however this node NEVER shows idle, even during the pause time in the dsmsched.log, on the server side the session always shows a RUN sess state, or sometimes IdleW for no more then 5 minutes.
 
Statistics

The statistics apply to the client node's last session with the server

Bytes received MB 24070.30%
Bytes sent MB 289.9


Duration seconds 72960.5
Session time spent idle 75.97%


Session time spent waiting for media to be mounted % 0
Session time spent waiting for the server to contact the client node %10.31%
 
Still looking for some help here!!!!!!!!!!!

Anyone have any ideas? We have changed the dsm.opt file from:

domain sys: grpwise: nds: to - -
domain grpwise: sys: nds: the pause was between the sys and grpwise, moved grpwise to BU first and see if we still have the lag, but we need to restart the box to take the change (stoping the TSM scheduler locks up groupwise, forcing a restart)

Also, management is getting hot and heavy about restores for the legal department - if we need to do a restore from say a week ago 1) it takes forever 24 to 36 hours and 2) most times the files they need are not there
Management class is set up for 30 day retention and 7 versions, so it should be there, but its not...
 
What version of the Client is installed on the Novell server? What is your TSM Server version? Netware 5?
 
TSM Server is 5.4.1.2, Novell server is 6.5 sp2 I think, the novell guys are out of the office right now, and the TSM client is 5.4, I think we are going to go back to 5.3 for the clinet on these boxes. That client had no problems stopping/restarting the tsm schedule.

Would doing archiving help with restoring all files needed for legal?
 
Can you update the TSM BA client to a later version? I know there are some fixes applied on a later BA client version.
 
Hi
I had a similar issue but it was restore
can you please post the version of netware
just type
PHP:
version
on the server console

also I need to know version of your tsafs
PHP:
m tsa*
will tell you

can you also tell me how much data daily roughly do you backup?

Vitek
 
NetWare SP6

There is a bug with the c-libraries with NetWare 6.5 that has been fixed with service pack 7 for NetWare that can blow out the backup times - see Novell TID 3325975

Client 5.3.0.12 uses the "old" libraries that actually work

Hope this helps
 
After moving the grpwise volumes to backup first, then the sys: and nds: volumes, and a system reboot the other morning, backups ran and finished by 3AM this morning with no lags.

Vitek -
Version = Netware 6.5 Revision 7
Novell Directory ver. 8.7.3.9
NDS Ver. 10883.73 10-9-06
m tsa* = tsafs.nlm ver 6.52.08 8-14-07
tsands.nlm ver 10551.71 7-19-06

We have 4 novell servers, the largest post office (and the one with the longest lag, before the change on the dsm.opt and system restart) was backing up around 24 GB.

We are, of course, going to be watching this system, and timing its backups, and doing some test restores.

thanks
 
Can anyone explain the lag that I am seeing ............. Again, no errors in the dsmerr, no system errors.

Executing scheduled command now.
04/07/2008 18:18:28 Node Name: BIGBIRD
04/07/2008 18:18:28 Session established with server TSM: AIX-RS/6000
04/07/2008 18:18:28 Server Version 5, Release 4, Level 1.2
04/07/2008 18:18:28 Server date/time: 04/07/2008 18:14:22 Last access: 04/07/2008 15:01:23

04/07/2008 18:18:28 --- SCHEDULEREC OBJECT BEGIN 7REV3592 04/07/2008 18:00:00
04/07/2008 18:18:28 Incremental backup of volume 'BIGBIRD\GRPWISE:'
04/07/2008 18:18:28 Incremental backup of volume 'BIGBIRD\NDS:'
04/07/2008 18:18:28 Incremental backup of volume 'BIGBIRD\SYS:' ~~~~???? Why does it take 7 hours for this process to start?????~~~~~
04/08/2008 01:27:42 Directory--> 0 GRPWISE:/ [Sent]
04/08/2008 01:27:42 Directory--> 0 GRPWISE:/ACADEMIC [Sent]
04/08/2008 01:27:51 ANS1898I ***** Processed 500 files *****
04/08/2008 01:27:51 Normal File--> 126,016 GRPWISE:/VOLDATA.TDF [Sent]
04/08/2008 01:27:51 Directory--> 0 GRPWISE:/ACADEMIC/ofmsg [Sent]
04/08/2008 01:27:51 Directory--> 0 GRPWISE:/ACADEMIC/ofuser [Sent]
04/08/2008 01:27:51 Directory--> 0 GRPWISE:/ACADEMIC/wpcsin [Sent]
04/08/2008 01:27:51 Expiring--> 358 GRPWISE:/ACADEMIC/0330CHK.LOG [Sent]
04/08/2008 01:28:35 ANS1898I ***** Processed 1,000 files *****
04/08/2008 01:28:35 Normal File--> 1,236 GRPWISE:/ACADEMIC/0407CHK.LOG [Sent]
04/08/2008 01:28:35 Normal File--> 86,016 GRPWISE:/ACADEMIC/NGWCHECK.DB [Sent]
04/08/2008 01:28:36 Normal File--> 452,608 GRPWISE:/ACADEMIC/NGWGUARD.DB [Sent]
04/08/2008 01:28:36 Normal File--> 452,608 GRPWISE:/ACADEMIC/NGWGUARD.FBK [Sent]
04/08/2008 01:28:36 Normal File--> 24 GRPWISE:/ACADEMIC/NGWGUARD.RFL [Sent]
04/08/2008 01:28:38 Normal File--> 4,911,104 GRPWISE:/ACADEMIC/wphost.db [Sent]
04/08/2008 01:28:38 Normal File--> 75,776 GRPWISE:/ACADEMIC/GWDMS/DMSH.DB [Sent]
04/08/2008 01:30:22 ANS1898I ***** Processed 4,000 files *****
 
Cwilloug, i had a similiar issue awhile back; after weeks of painfull digging it was related to our symantec antivirus client compressing/decompressing files durinig backups. Not that this is your issue but thought i would throw it out there.
 
Thanks rpt, I just asked the admin of the boxes I'm having trouble with, they are not running any antivirus on the novell boxes.

Back to square one..........................again.........................
 
This is a slight stab in the dark but try it with memoryefficientbackup turned on.

Sometimes this will actually increase performance significantly rather than decrease it.

I'm just wondering if that delay is caused by the time its taking to sort through the directory tree before actually starting the file examination/backups.
 
BBB - thats the first time I've heard that memoryefficientbackup increases performance, I havent tried that yet because everything I've read says that it has a negative impact on performance, I'll try it and see though.

Thanks
 
I did say *sometimes* :) But if its taking that long I don't see how it can get much worse :)

ps. you can always use the tracing features of the tsm client to figure out why the big delay. not sure if it works under netware though.
 
Back
Top