Some users have a misconception that if an update with a write concern of w>1) (e.g. w:3 or w:majority) times out, that we rollback the changes to all nodes that they have occurred on.
We should document that this is not the case.
Some users have a misconception that if an update with a write concern of w>1) (e.g. w:3 or w:majority) times out, that we rollback the changes to all nodes that they have occurred on.
We should document that this is not the case.