The test expect sub-test not to run in parallel.
This option tells ctest to run each test N times until the test fails or the N times have run. This is useful for finding random failing tests.