Problems with Networker and DDS. 1

[warlock0199]

Hallo there everybody.

We have a bit of the problem. We are running a windows 2000 and a SAN environment, with about 12 W2K clients.

When the backup starts, there are a few processes that start as well. (NSRMON.EXE and NSRINDEXD.EXE and so on…) The problem is that when the one client’s backup is done, the processes don’t kill automatically. After a while, there are about 100 processes running. Then everything fails.

Any advice will be appreciated.

 

[605]

This maybe true but ... what does this have to do with DDS?

Start the group from the command line with added verbosity: savegrp -v -G group_name

Then look at the output and ..\nsr\logs\daemon.log . This should tell you more. This is all i can help you at this moment with this kind of input.

 

[wallace88]

Which processes that don't kill automatically are you talking about?

Are you seeing this on the server or on the client, or both?

What version of NetWorker are you using?

 

[warlock0199]

We are running networker 6.1.1 build 238.

I have only checked the backup server for the processes, so as far as I know it is only the backup server giving the process problem. The processes are:

NSRINDEXD.EXE
NSRMMD.EXE
NSREXEC.EXE
NSRJB.EXE
NSRMMD.EXE
NSRMMDBD.EXE
NSRMON.EXE

thnx

 

[wallace88]

My first suggestion would be to upgrade your server and storage nodes to NetWorker V6.1.4. After which, monitor to see if your observations are repeated.

On the backup server, there should only be one nsrmmdbd proces running, similarily to nsrd. There can be a number of nsrmmd's during the backup job and nsrjb's will occur depending on whether tape operations are required.

Similarily it is not uncommon to see multiple nsrindexd and nsrexec's during backups.

Check the event logs and look for messages regarding system or application problems during backups.

Also, as 605 suggested, run the savegroup backup from command line and see if you can generate any more detailed information relating to the problem.

So far, there is not enough information to show that a problem exists.

 

[warlock0199]

Thanks for the comments. Just a another question.

Does it make sense to you guys if I say that the server cant cope with all the SCSI commands when a the backup starts with the 5 clients?

When I run the inquire command I get the following:

scsidev@14.0.1:COMPAQ DLT8000 022B|Tape, \\.\Tape0
scsidev@14.0.2EC TL800 (C) DEC0519|Auto-changer (Jukebox), \\.\Changer0
scsidev@14.0.3:COMPAQ DLT8000 022B|Tape, \\.\Tape1

Does it make a difference that Tape0 is picked up before the auto-changer?

 

[605]

1. Disable the Removable Storage Manager service.

2. Make sure that the OS does not use the jukebox. In other words,
"deinstall" the jukebox is such way that the handle "\\.\Changer0"
will disappear - just this entry, not the jukebox as well.
You most likely need to restart the computer.

Then run inquire and reinstall it again.

 

[warlock0199]

Thanx...

We changed the MDR config. Everything seems to be working much better now

 

[jjrinconbayon]

warlock
did you have the problem with the server or with the client, and how do you solve the problem
Thanks a lot,
Juanjo

 

[HelpNow]

I suspect the problem might have occured due to confuson between scsi id between the jukebox and standalone tape drive, in my openioun may not be good idea to mix external tape drive with another library on the same server.
In other words, when networker issues command to tape0 the changer 0 may be receiving the cmd.

 

[605]

No - as long as the device can be uniquely addressed, you may have as many devices of whatever mix as you want. You should expect the same problems when actiong with OS tools.