Today, there is no need to hold the critical section across failovers during a moveRange commit because the commit is performed in a single step on the config server.

However, with authoritative shards, we must hold the critical section throughout the entire process of committing the chunk migration in both the global catalog and the shard catalog. Therefore, we need to ensure that the critical section is preserved across failovers so that it is not released in the middle of the process.