* Re-acquire pool at the beginning of Protocol loop * Fix query router + add tests for recycling behavior
* Explicit shard selection; Rails tests * try running ruby tests * try without lockfile * aha * ok