SERVER-20447 added the ability to define a $config.skip(cluster) function to avoid running an FSM workload against a cluster with an incompatible configuration (e.g. the balancer being enabled). We can use this mechanism to clean up the explicit blacklisting of the sharded_moveChunk_drop_shard_key_index.js FSM workload in all runners except fsm_all_sharded_replication.js.

Additionally, SERVER-20447 introduced chunk_helper.js for performing the "split", "moveChunk", and "mergeChunks" commands. It handles errors returned by the server from another sharding metadata operation already in-progress by retrying the operation with exponential backoff. We should see if it possible to use these helper functions for the "moveChunk" command or the "split" command performed in sharded_moveChunk_drop_shard_key_index.js. Note: Given that a worker thread doesn't own a particular chunk range, there may not be any stronger assertions that can be made even with the retrying.