Saturday 15 June 2013

Restoring PRD Instances into QAS for DR Purposes

As Basis, you would be in the situation to provide DR planning activities for business sustainability during unforeseen disaster. 

Common SAP DR solution:

Option 1: Main SAP systems (DEV,QAS,PRD) locate in HQ with connectivity to Backup system (PRD) in DR Centre.

Option 2: Low cost solution , where a set of PRD instance in the PRD systems be clone / configure in QAS server that act as primary system during DR (ex: PRD server hardware failure)

In this sample, we'll focus on the steps of restoring PRD data into QAS (PRD instances) 
As highlighted, the QAS server contains both QAS and PRD instances

The specifications of QAS and PRD servers of this restoring process are running on HP-UX, Oracle 11g and HP Data Protector as the backup / restore software

1) QAS system been set up with a PRD instances which is the exact mimic of PRD server (sample configuration will be cover soon)

2) Complete filesystem backup of the PRD system with Ultrium tape

Steps to restore with HP Data Protector:

1) Only restore the PRD oracle database. Oracle application does not required

2) In the "Restore" dropdown box. Right click Filesystem, source server, “properties”

3) Click “restore version” to check the file to be restore

4) Ensure the data is with the latest backup set

5) Select “destination” tab and ENSURE the destination is “QAS SERVER” (any incorrect target selection will cause disaster)

Ensure “options” tab the “display statistical information” check box selected and click “restore” button

7)  Pop-up of the restore windows

8) Click “finish” button to start the restore process

9) The restoring screen

10) Once restoration complete, use putty to ensure the data is restored in the QAS server (PRD instance) 

11) Repeat all the data required in step 1.

* Recommended to perform the restoring of each folders manually one after another manually for each of the folders restoration. By running all the restoring tasks in queue will cause possible error that due to the natural behaviour of HP Data Protector. Ex: if the 1st restoring task taken long time to be complete, the 2nd task in the queue will be time out and causing all the subsequence tasks go into pause stage and required user to resume the 2nd time out task to continue restoration process.

No comments:

Post a Comment