DR networker server

[santhas]

Hi Gurus,

We are running production backup server on Sun Fire 480r, I am setting up DR server on a Sun Fire v240. DR server will be used for restore purpose only. Both hostnames are different. I have setup rsync to sync the /nsr/index from prod to DR. What else I have to do? or does anyone have a better way to accomplish the task?

Thanks
San

 

[patrikjelinko]

Hi,


If you make a cold standby server like solution, you will need the mm (media database) and the res (resource definitions) directories to be replicated as well. Also ask EMC licensing to authorize the software for both hostIDs, so you wont have any problems running the software from the other machine. (This is for free, called fail-over hostid authorization).

Keep in mind, that you may have to recreate/reconfigure your backup devices on the cold standby host.


Other solution:
You can cluster NetWorker with Sun Cluster or with EMC AutoStart.

I hope this helps.

BR,
Patrik

 

[santhas]

Thanks for your replay.

My senario is different:
I have a different server and a different JB for DR. If copy the mm and res directories, the DR server thinks as a client of prod server. How could I overcome this?

Thanks
San

 

[patrikjelinko]

San,

So you mean you have to different servers licensed with two separate jukeboxes. You want to be able to take out a tape from server1's jukebox and put it into server2's jukebox and restore file from it. Is that right?

If yes, than I advise you when you move the tape, you should run a scanner process to put that particular tape into server2's media database and rebuild the indexes.

Steps:
load media
scanner devicename
check if all clients on the media are created on the DR server, if not create them
scanner -i devicename

You can find more info about this kind of stuff in the admin guide, chapter 13, section 20. (At least in the Admin Guide for NetWorker 7.3.2).

I hope this helps.

Best regards,
Patrik

 

[santhas]

Thanks Patrikjelinko.

That is what I am after. I won't be in the office next two days. I will try your method when I get back to work.

 

[santhas]

Hi Patrikjelinko,

I am in the process trying your options:

1. Loaded a tape from Prod JB to DR JB
2. Ran nsrjb -d -S -P to deposit the tape
3. An now when I load the tape (nsrjb -I -S) to the drive I get the following error message:
nsrjb: Jukebox error: Mon 15:18:14 Hardware Error, Diagnostic Failure on Component 0x01

Any ideas?

Thanks
San

 

[605]

Look up your hardware manuals. They should tell you more.

You may also run the jbexercise command to see whether your jukebox is fine.

 

[santhas]

Summary report of jbverify
======= ====== == ========

Hostname Device Handle Blocksize Jukebox Drv No. Status
-------- ------------- --------- ------- ------- ------
backup002 /dev/rmt/0hbn N/A STK-L20 1 Fail

**********************************************************************


Exiting jbverify successfully.

 

[patrikjelinko]

San,

It looks like that you have a drive failure. Was this drive working before?
If yes, then I would advise you to talk to the manufacturer to replace the failed drive.
If not, that maybe the configuration is improper.

BR,
Patrik

 

[santhas]

OK, sorted out the drive issue. Now the issue How can load the tape from prod server to dr server jb?.
nsrjb -l -f /dev/rmt/1cbn -S15
slot `15' (024) will not be loaded since pool `Unknown pool' is not available.

Then only I can run scanner to rebuild the index.

Thanks
San

 

[TDun]

Try using nsrjb -ln -f /dev/rmt/1cbn -S15
the n tells it mount but no load.
The difference is that the OS can see it to allow scanner to work, but nsrwatch would not show it on hte drive as Networker proper does not have it mounted.