Database Reference
In-Depth Information
Conclusion
RAC is an HA solution; however, it can only be HA when all the components that comprise the clustered configuration
are HA.
Almost always when project timelines slip, it's the server-level tests that are comprised to meet the production
go live dates. To ensure HA, these components have to be tested. Irrespective of how many clusters have been
configured and how many times a RAC configuration has been deployed into production. We cannot stress enough on
the fact of how important these tests are. None of these tests can be ignored. Moving from a single instance to a RAC
configuration and a few days into production if RAC crashes due to a hardware failure or a component failure, RAC
cannot be blamed: it's the lack of systematic testing that create such failures.
We discuss other tests around scalability and throughput in the next chapter. Availability testing has a final phase,
which has not been discussed here. To maintain the flow of the various tests, we discuss the last phase of availability
testing along with the scalability tests in the next chapter.
Search WWH ::
Custom Search