jstests/core/splitvector.js occasionally times out on slower buildvariants in the retryable_writes_jscore_stepdown_suite. Even when successful, this is consistently the slowest test in this suite by a long shot (only ~8 min on Enterprise RHEL 6.2, > 1 hour on Linux DEBUG, ~ 1 hour on Windows 2008R2 DEBUG).

This is expected because the test performs 10s of thousands of inserts and at least a dozen splitvectors. We can blacklist the test to avoid these timeouts, but the test itself isn't failing, and we'd lose some test coverage and some stepdown stress testing (longer runs can have ~400 failovers).

The purpose of this ticket is to determine if its worth blacklisting splitvector.js or if we should find another way to avoid the timeouts.