Determining if a core dump is from a process spawned by a given mongo shell based solely on PIDs is not reliable enough. 

Since introduced, we've seen 3 BFs where the unexpected core dump is from a different, concurrently-running jstest that creates an expected crash. BF-36046, BF-36549, BF-36765

There are no BFs with this syndrome that contain real issues.

We must keep the cleanup for tests with intentional crashes to avoid the overhead of running the core analyzer when we don't need to.