| || |
configuration& fail-over testing
Can somebody please explain how you approach and plan for configuration testing, failover testing and recovery testing. If you provide details with any examples is truly appreciated.
Re: configuration& fail-over testing
Failover testing can be fun. Ahhh. Intentionally crashing boxes when they are under load.
Most of our failover tests are conducted like any performance test. We put a normal load on the system (normal loads, not stress loads, though you could do that, too). A few minutes in to the test, we force an event that _should_ cause a failover.
- Pull the network cable out of the back of a system
- Pull a power cable
- Pull the system SCSI drive
- Pull a SCSI cable if attached to an external array
There are other ways, such as executing a program that will crash the system. But, I find the above list is the most fun.