The writeConcern setting w:2
means that the write will be acknowledged when exactly two members of the replica set has acknowledged that the write happened (see https://docs.mongodb.com/v3.2/reference/write-concern/#w-option). In other words, it will wait until the write has replicated (via the oplog) to one other node, since the Primary is counted as one node.
This means that the "speed" of the insert/update query will be subject to your network speed. If the network is slow or congested, then the insert will appear to be "slow". This is not due to replication blocking anything, it is simply the effect of specifying w:2
in a congested network.
There may be a network congestion that triggers both the sync source change and the slow insert, but the replication process by itself does not block any insert operation.