[0xdeadbeefbabe]

I'm still misunderstanding what happens when a network partition creates a two leader scenario. It is in a slide here http://thesecretlivesofdata.com/raft/ under "log replication" after creating a network partition that separates CDE from AB. Doesn't that mean a client of node B could read stale data? Or maybe an implementation like etcd would let the client choose to read or not in that case?

[ongardie]

If the network is split into AB and CDE, then only CDE will be able to reach a quorum. If a client is on the AB side of that split, no consensus algorithm could guarantee freshness on a read. LogCabin clients will continually retry until they can talk to a functioning leader (and clients can put a timeout on that) so that you get a guaranteed fresh read every time; other systems could choose to return a stale result right away instead. It's up to the implementation, but I strongly recommend people not do stale reads by default.