Backup in a ops/service guard cluster and recovery

[Justo Exposito]

Hello,

We have two boxes in a cluster with 11.00. And we need to know which kind of backup we must to do in order to in case of a disaster recover the cluster in a diferent place and a diferent boxes, and how to recover.

We have omniback II 3.5 to make a backup and the database used in this cluster is Oracle 8.1.6., we have a disk array too.

We are doing an omnibackII full backup for the system, an omnibackII hot backup by tablespace for Oracle and a make_recovery for the system. And we extract information about the system by the script attached and then we send this information to a PC by ftp everyday.

This is ok?

Thanks in advance.
Justo.

[Ian Dennison_1]

Sounds like you have all the components needed. How about Redo Logs from Oracle to bring the system up to date?

The best way to ensure you have all you need for a recovery? Perform an trial failover and / or rebuild if you can! This will highlight anything you are missing pretty quickly, in terms of files, and also in terms of experience and procedures.

Share and enjoy! Ian

[Duncan Edmonstone]

Justo,

Well a quick look says you are doing all the right things, but the bottom line is that no-one can tell you your DR strategy is OK based on this info! The *only* way to know your DR stratgey is right is to TEST, TEST & TEST again!

If you haven't actually done a recovery to another system, how can you have the confidence it will work?

Regards

Duncan

[Justo Exposito]

Hi,

The problem is when you don't have two boxes to probe, then you need somebody that probe it and tell you which kind of backup operation he is doing.

Many thanks,

Justo.