Simple DR scenario

[rake1596]

Hi,

I have 2 Windows 2008 R2 servers on which I plan to install netbackup V7.0.1.

I would like 1 server to be the primary Master server and the other will be used as a standby server in the event of a problem with the primary server. No tapes are involved and this is purely a disk to disk backup solution.

Ive read the 7.0.1 HA Admin guide (chapter 5) which basically goes through the procedure of creating an alias, installing Netbackup using the alias name then installing Netbackup on the standby server etc etc. I presume I will need to create the same storage units and clients on both servers. 

All I bascally want is a simple and efficient way of using the standby server in the event of a serious issue with the primary server. Im pretty certain most of what I need to concentrate on is the resilience of the catalog between the servers which im still reading up on.

Appreciate this could be a complex subject but hopefully someone could point me in the right direction of some documentation or share their experiences. I would have liked to go down the full cluster route which we may still do (comfortable for VCS) but thats in the future,

Any thoughts would be appreciated..

[RiaanBadenhorst]

Hi,

Since you're not going to cluster it, there is no need to use a virtual name. You can simply install the first node, install netbackup, patch it, configure it, and run your backups.

Install the standby node with the same name (not connected to the network), install netbackup, and patch it.

Now, when you want to "fail over" to the second node you have two options after you shutdown the active node (or if its fails)

If your catalog is on a SAN, deassign it from the active node and assign it to the passive node. Do the same for your backup to disk location if you're doing backups on the master (using it as a master / media server).

or

You can use the catalog backup and do a recovery on the standby master. This might be a little more complicated as you're using backup to disk (so you'll need to move the backup to disk location too)

Also make sure to export your registry on the active server and import it on the standby server regularly. Some settings are kept in there, the rest of the information is in the catalog/EMM databases.

Those are some general points. Hope it helps.

[bfcsaus123]

Thanks Riaan some helpful information,

Im interested to know why the hostname has to be the same on the primary and secondary server. I presume this is due to the Catalog requiring the same hostname or is there another reason??

Thanks...

[RiaanBadenhorst]

Hi,

Yes, the master server name has to be the same. If you were clustering it, VCS would configure a virtual name that would move between servers. NBU would be installed using the virtual name, and not the physical. In your case it that isn't feasible (unless it unix where its easier to have a virtual name). But still more complication than you need. What you're actually doing is a DR exercise (manual failover).

Do take care when making the decision to cluster or not to cluster. NBU these days is "cluster aware" so you can't just move from clustered to un-clustered and vice versa (it requires consulting). That sucks, i know......

[bfcsaus123]

Ive read that Netbackup is MSCS aware (with an enterprise license) so I may see if theres a possibility of installing MS cluster which will make life a lot easier...

Thanks again for your assistance, much appreicated...

[RiaanBadenhorst]

Hi,

Yes it works with MSCS, installed one last week. It works fine.

Good luck!

[Marianne]

Correct - Chapter 2 of NBU HA guide covers 'NetBackup in a Microsoft Cluster Server'.

Principals remain the same - you need shared storage for the catalogs as well as virtual hostname and IP address.