NBU catalog Recovery on new server

[pylej]

We are currently running NBU 6.0 mp7 on Solaris 10 sparc hardware. We are trying to recover onto a Solare 10 X86 platform with NBU 6.5.  I did some testing on a single node and was able to take a cold backup of the older environment restore it using the bprecover commands, then uninstall the SYMCnetbp and when it asked "is this part of a upgrade process" I said yes so if left the DB info from the 6.0 environment on the machine.

I then reinstalled NBU and during the reinstall it found the old 6.0 DB and upgraded it to 6.5, so everything work great.

I now need to do this on a clustered environment, I have followed basically the same process. However when I run the cluster_config script it drops the Emm database and create a blank new one. So when I bring up the system I can see all the client, but no tape history info.

Does anyone know how I could get around it dropping the database? The second problem is then the files in the /db/data folder are from the 6.0 enviroment and never get upgraded. So is there a way that I can upgrade the EMM database?

Thanks for any insight or help in advance

[Ed_Wilts]

http://www.symantec.com/business/support/index?page=content&id=TECH77448

[Marianne]

Only Symantec Consulting services can convert non-clustered to clustered master....

Have a look at this discussion and the TN mentioned:

https://www-secure.symantec.com/connect/forums/upgrade-nbu-master-server-ha-cluster-server

[pylej]

Hi Marianne, I only did some testing which none clustered servers.

In this case the old enviroment is clustered using NBU 6.0 mp7 on Solaris sparc hardware.

The new environment is also clustered using NBU 6.5.4 on Solaris X86.  I have basically followed the techs notes suggested, however the files that get restored to the /global file system that normally live in /opt/openv/db/data are being blown out and and empty db is put in its place, after I run the /cluster_config during the installation.

I know that we should use the Symantec consulting to do this because it is clustered, but I am sure other people have been able to do this.

Do you think the transfer from the sparc hardware to the x86 hardware is causing any problems?

Can I extract just the EmmDB parts from the cold backup I took and just try to recover that?

Do you think it would help if I used the GUI and restore it from the DSR files? I have only done the cold catalog method so far.

Thanks alot for your help.

[pylej]

Hi Ed,

I have seen this tech note and have followed it basically, however I am dealing with a cluster system which is causing the problem, but I am sure someone out there has done this before and know how to get this to work.

I posted more details to marianne post, maybe you have some other ideas. My main issue is witht he restore of the files that normally live in /opt/openv/db/data

Thanks for the help