1Aug/09Off
Disaster-recovery quality characteristics
Today's tip comes from Klaus Schmidt's High Availability and Disaster Recovery. In the book Schmidt outlines a useful list for being able to test a disaster-recovery (DR) installation:
- Being able to activate the DR system while the primary system is active.
- Being able to test the DR system without requiring changes to the primary system.
- Being able to component test different aspects of the DR system.
- Availability of workload generators, regression tests, and monitoring on the DR installation.
- A short time frame for resynchronization after testing is completed.
August 26th, 2009 - 02:20
Hi Karen,
i attended your seesion in STANZ 09. and found some points you said interesting.
Anyway, one quick point on DR testing.
Many testers when they do DR does it one way. What i mean is, in the scenarios i mentioned below, testers often forget 6 and 7! They dont consider bringing back DR to Production.
1. Take down the production environment on a pre-determined time.
2. Ensure that the DR environment can be bought up and running in the agreed time
3. Do smoke tests to ensure that the DR environment is operational and all BAU activities can be continued. All standard Add/Edit/Delete functions on the application should work as normal. (e.g. Create a Quote or Contract/Supplier functions)
4. Do smoke test to ensure that the data created in the production environment is mirrored to the DR environment (e.g. a Quote created is there in the DR environment)
5. Ensure that the Production environment can be bought up and running and service can be resumed
6. Do smoke tests to ensure that the Production environment is operational and all BAU activities can be continued. All standard Add/Edit/Delete functions on the application should work as normal. (e.g. Create a Quote or Contract/Supplier functions)
7. Do smoke test to ensure that the data created in the DR environment is mirrored to the Production environment (e.g. a Quote created is there in the production environment) in the agreed time (10ms)
8. Ensure that service can be resumed within specified time (as per SLA)