Hello there,
I have a question about a possible scenario happened with a replica set, involving a network partition. Mongo version is 3.2.11

Let's suppose to have 3 nodes: A, B, C. A is the primary. B, C are secondary.
Suddendly a network partition happens and it continues for lot of hours. During the network partition A and B can connect, as well as B and C.
If I correctly understand how replication works:

• A still see a majority of nodes and remains primary
• C don't see A and thus attempts to start an election
• C is disconnected from A and thus its data become older and older
• B still see A and thus ignores the request for election from B
• B continues to replicate from A

Now let's suppose that, after a lot of hours, without fixing the network partition, node A crashes.
If I'm correct B and C will start and election.

The question is: are we 100% sure that the election will result in the node B (which is synchronized with A) to be elected as primary ?
Or there is a possibility that instead the node C (which has very old data).