Write operations in MongoDB replica sets will, by default, get acknowledged once the write has been acknowledged by the primary server. If we want to change this behavior we can do it in two different ways.

We can request a different write concern per operation in cases where we want to make sure that a write has propagated to multiple members of our replica set before marking it as completed as follows:

> db.mongo_books.insert(
  { name: "Mastering MongoDB", isbn: "1001" },
  { writeConcern: { w: 2, wtimeout: 5000 } }
)

In the preceding example, we are waiting for the write to be confirmed by two servers (primary plus any one of the secondaries). We are also setting a timeout of 5000 milliseconds to avoid our write from blocking in cases where the network is slow or we just don't have enough servers to acknowledge the request.

We can also change the default write concern across the entire replica set, as follows:

> cfg = rs.conf()
> cfg.settings.getLastErrorDefaults = { w: "majority", wtimeout: 5000 }
> rs.reconfig(cfg)

Here, we set the write concern to majority with a timeout, again, of 5 seconds. The write concern majority makes sure that our writes will propagate to at least n/2+1 servers, where n is the number of our replica set members.