operation_time_read_and_write_concern.js disables replication to secondaries using a failpoint, makes a local write to the primary, re-enables replication, then waits for the write to replicate to the secondaries before issuing a read with readConcern majority, expecting the locally written document to be returned. Currently the test uses ReplTest.awaitReplication, which blocks until the secondaries last applied optime matches the primary's, but it should use ReplTest.awaitLastOpCommitted, which blocks until their last committed optime is caught up. This will ensure the majority read returns the expected documents.