Eliminate database caching from skewing your performance test results

If you repeat the execution of short, say 1 hour bursts, of performance tests, you run the risk of skewing your performance results as a result of database caching. This can happen, as we often want to repeat a test utilising the same data, for baselining and then comparison benchmark runs. Using the same data in repeated tests may retrieve data from the database internal cache, reducing or even eliminating disk reads, which can give better preceived performance results.

Continue reading