Can some one please share your thoughts on the below situation and help me.
I have a Netbackup 7.1 master server ABC with 2 media servers and have implemeted puredisk. There is a Netbackup server at DR location. Since there are few servers to be backed up in DR location i am planning configure DR netbackup server as media server(with SL500 Tape library) and Master server as ABC. We have the catalog on shared storage which is replicated to DR location every 6 hours.
In case of disaster, can i promote media server in DR location as master server(change media server name to ABC) and use the replicated catalog to bring up Netbackup in DR.
If yes, then is there anything specific settings that needs to enabled on master server before the disaster actually happens.
My answer is based on the assumption you have only copied the images DB.
If you have replicated only the db/images part of catalog, you can't just use this alone, as you are missing the EMM database part. You have not confirmed exactly what you have taken across.
You could take a catalog tape and use it for a normal restore to just restore the files in /usr/openv/db/data to an alternate location, and then recover the DB using nbdb_restore -recover <dir> (and class dir to get the policies). Problem is, how do you guarantee any other files restored by a 'normal' catalog restore are going to be correct.
You would have to configure the media server as the same hostname as the 'old' and probably easiest to just reinstall NBU to make it the master. You can take the replicated catalog to recreate the image db, but need to recreate the EMM database, which you cannot do from a catalog restore without recovering the image db as well.
My concern, given that I do not know how you have replicated things, is if you have copied across the db/data files (the EMM database) how are you going to confirm it is consistent ?
I'm old fashioned, because I deal very regularly with restore attempts that are 'not supported' - and very very often they go wrong, not necessarily on the day they are done, but issues are hit when it has 'gone live'.
My official answer is that if it is not in the book, do not do it. Quite often, systems get into such a mess that the only way is to do a proper supported restore - meaning you have one very unhappy customer, they have to do the supported method that they were trying to avoid + they ''lost' all the time they were trying to get the other method to work. There are a number of doc floating around about using, for example srdf replication or simliar - but I am not aware you have this (or equiv).
To be honest, more info is needed, need to know exactly what you have copied across for a start.
My answer so far is based on the assumption you have only copied the images DB.
There are means and ways - https://www-secure.symantec.com/connect/sites/default/files/NetBackup%20Catalog%20Replication%20with...
Thanks Marianne for supplying that link.
As showwn in the TN, coping the DB every x hours is NOT supported.
I broke my golden rule in my first reply - I showed a way that is unsupported, this was done to show you why what I think you are suggestiong will not work, I am not showing this to give you a way to work around it.
If you ended up with a non working system, suppport would request that you performed a proper catalog recovery.
Thank you for your responses and sharing your thoughts.
Well If i do a RSYNC of entire /opt/openv to the DR location will i have to the issues you had mentioned. I mean will i still have to recreate the EMM database on DR server.
Probably not, it will probably work fine. Note I use the words "probably" that's because it's not supported to use this method. However, I have copied /usr/openv before and it's worked without issues. Just make sure nbu is down when trying your sync and remember, it's not supported.