Disaster Recovery suggestions

[James_Butler_2]

Hi

I was hoping to get some suggestions or comments on experience from people who have had experience restoring NetBackup in a disaster recovery scenario. We are currently running NBU5.1 MP2 but and planning a new installation of NBU6.0 MP2

Our organisation has a requirement to test recovery of all systems at the DR site once per year, and I wanted to make this a painless as possible for the NetBackup system.

The production environment
====================

The current NBU5.1 system runs on two VCS cluster nodes (one active, one passive), and only one node at a time runs both master and media services. This automatically fails over to the other node. One SAN-attached MSL6060 is used at head office, and mirrored to another MSL6060 at DR

The dr environment
==============

Another two node VCS cluster - cold standby. Each time a catalog backup is run in production, it is also copied to dr. When VCS starts NBU at DR, a script changes entries in the catalog database to match the new server name. This was setup by veritas, but it now non-functional.

In normal production the head office NBU cluster can see both tape libraries, but in a dr test, the netbackup servers can only see the dr tape library.

Thanks in advance for any suggestions. There is no need to save any changes during the dr test (ie. no failback.)

Thanks
James

[G_S]

My suggestion is not to have a disaster.

/sorry...I had to....:)

[Chia_Tan_Beng]

> The dr environment
> ==============
>
> Another two node VCS cluster - cold standby. Each
> time a catalog backup is run in production, it is
> also copied to dr. When VCS starts NBU at DR, a
> script changes entries in the catalog database to
> match the new server name. This was setup by veritas,
> but it now non-functional.
>

James, do you intend to preserve the current way (i.e. a script changes entries in the catalog database to match the new server name)? If yes, I think the only way is to get the professional service team to correct the problem in order to have your DR procedure officially support (and I think the script may do more, such as changing the storage unit of the backup policies, catalog backup configuration, etc).

If not you can have the 2-node cluster at DR site configure with the same name as your production, since they're cold standby. It will make a lot of thing much simplier.

[Raghava_Chary]

One of the important thing is to make a note of catalog media ID.

Incase if DR setup doesn't work you can restore every thing from catalog in 3 to 4hrs and the environment will be up and running...

Raghava.