Recover of Catalog to a different server/Netbackup instance

[croftwm]

Software: Netbackup 6.0MP4
Server: Linux OS

Hello everyone,

I currently have Netbackup performing nightly backups on an Oracle 10g instance.

In trying to make my disaster recover more robust (read: incase my server room burns down) I have set up a dupilcate system in another building under a different subnet. In trying to make sure that I have all that I need to bring our systems back in an emergency, I have set another instance of Netbackup.

I was hoping to to restore my catalog backups from my production Netback onto this DR Netbackup instance but I am not understanding how this is performed.

Ultimately I was hoping to get the catalog(s) restored so I could restore the Hot backups of my Oracle instance to this DR system.

Has anyone attempted this before and would be willing to give me some pointers.
My experience with Netbackup is limited to installation and configuring backups. I do not have any experience (yet) with restoring catalogs or hot backups.

The pervious admin here warned me that hot restores are tricky (there was an accident were data was restored to the production server and not the intended test server)

Thank you for your time,

William

[Sriram]

You can very well restore HOT catalog backup to your DR systems. Provided ensure the following points:

1.  Hardware whould be exactly the same in DR and production
2.  Dosent matter how many media servers you have, even single master server can do the job for you.
3.  Install netbackup in the DR site with all proper license that of production.
4. Have the OS configured with same system parameter of that of production
5.  Ensure the DR site master server sees tape drives.  Ensure the same tape drives are installed in DR as of production
6.  Now take the hot netbackup catalog backup in production
7.  Once you verify you have your system ready with tape drives and netbackup install in the DR then
8.  Load the catalog backup tape to the tape library or in the drvie which is seen by DR master server
9.  Install the JAVA GUI in the DR site
10. Login to JAVA GUI, and on the right pane click on catalog recovery.
11. This will prompt you for restoring the catalog backup taken in production.
12. When you took the catalog backup ata production site you would have noted that it creates a DR file after successful completion of backup.  You have that DR file copied to your CD or DVD or thumb drive
13. Now in the catalog backup window in DR site it will prompt you to give the location of the DR file which you took from production.
14.  You have to copy that DR file before starting the catalog backup in the DR master server somewhere and make a note of the location of DR file and give the location in point 13.
15.  As recommended by symantec restore the entire catalog.
16.  Now the catalog restore will go and search the tape you have inserted to the DR library connected to the master server.  Emsure you have inventories the library so that the tape is recognised by netbackup.
17. Set the barcode rule for the tape to read the lable correctly
18.  Once you do point 17 and 18, point 15 will start successfully to restore the entire catalog
19. Once the catalog is recovered you will now have the production recovered to DR site.  It means that all your device configuration in DR has overwritten by production site.
20.  You have to run the device configuration wizard again to recognize the DR devices.
21.  Once done you will have DR devices seen and all the policies seen of production.
22.  Ensure the db/class folder is renamed because all the scueduled backups will start to run and fail.
23.  Netbackup catalog recovery is now completed and you are ready to restore the oracle backups taken from production site.
24.  Ensure the oracle backup tapes are there in DR site.

[Nicolai]

HI William.

A restore should be possible if you are restoring to the the same type of master server (clustered VS non-clustered).

You can save time by not re-installing Netbackup on DR-site if you know this article :

How to create a blank NetBackup NBDB database for the recovery of the NetBackup NBDB database from a...

I have used the article before with great success. If you have concerns of overwriting data, then make sure the DR master server can't resolve the IP of the production server.

Have fun :)

[quebek]

Keep in mind that the hostname of Your Netbackup Master server at DR site should be the same as the production one. This is extremly important!!
If hostnames would not match probably catalog recovery will fail.
And as Nicolai said - do not let for any communication in IP terms between DR NBU and PROD NBU!!!

[Nicolai]

Very good point quebek. Master server names must match (one vote from me) . You can also use the Linux built-in firewall to deny traffic to the production master server. Depending on flavor of Linux - look for iptables, ipchains or ipfilter.

[Sriram]

That is a very basic check which should be taken care of by any one who is preparing for DR.

;P

[LeeClayton]

I was speaking to a "netbackup" guy the other day and he said when recovering a catalog on a DR server, that DR master server does NOT need to be the same name as the original one. Maybe I;ve missed something, we do hot catalog backups as well and hope to test DR soon. I thought the name should be the same plus, can someone confirm, should all the drive letters on the server be the same?

[Sriram]

No not at all

[LeeClayton]

what do you mean, the server name has to be the same? and the drive letters have to be the same?
using windows 6.5

[croftwm]

Thanks for the information everyone I will get back to reading and start testing.

Thanks,
William

[Sriram]

All the best for your DR.

before you go vote and mark it as a solution

Thanks

[croftwm]

I just got the Catalog restored and everything looked good. Policies where there.
I reconfigured my tape drive (which on the DR system is only a standalone external 448).

When I started to get ready to begin the oracle restore I found that I could not access the drive.
Each time I attempted to go to the Device Monitor in the GUI I recieved this error
"Could not connect to vmd on host myserver.mydomain (70)"

I searched for this error but did not find anything that seemed to relate to what is happening with this system

I attempted to reconfigure the storage device again, using the GUI, and got this message
"Server connection Failed. Verify services of daemons are up and running."

I then went to the Activity monitor and checked on the processes and daemons and it shows them as running.

I rebooted the system just to make sure but nothing has changed.

Is this error a result of the restore and not configuring the drive correcty?

Thanks again for all of your help,
William

[Sriram]

As i have mentioned earlier in my post, normally when you recover the catalog from production it will overwrite the DR.  So you have to do the device configuration again after successful completion of catalog recovery in DR.

Hope this helps.

[karthikcs]

I have tested this in my environment; you can restore your production netbackup catalog to different server. Once your catalog backup was restore to alternate server or different server next thing you need to do is removing from network other wise if it is same network it will have an effect on your production environment.