commands_that_write_accept_wc_configRS.js and commands_that_write_accept_wc_shards.js still have fairly low wtimeouts of 25 seconds. These should be increased to 5 minutes so that timeouts truly indicate bugs.

We should also add calls to awaitReplication at the beginning of each call to "dropTestData" on each replicaset in the cluster to make sure that any non-replicated writes get replicated before calling dropUser (which requires writes to replicate on the config servers) or testing a new command.