When movePrimary fails during the clone phase due to any reason, the operation will be aborted and cleaned up on the next iteration.

During the movePrimary cleanup, the donor sends a drop for every collection that was expected to be cloned to the recipient shard. One of these collections could be <dbName>.system.views, which can't be dropped if there is any view bucket collection on the node. Therefore, if there is any sharded view collection on the recipient shard while <dbName>.system.views collection is being dropped, the movePrimary cleanup process will fail and the movePrimary will be retried forever.