Why CRDB use both of raft and gossip protocol?

I know that all consensus happens through raft, and gossip takes a role in managing cluster nodes.

Why crdb uses both protocols together not only use the raft protocol?

Hey @joowon,

We use Raft in order to maintain consensus in order to replicate data with consistency and high availability. Gossip on the other hand is how new nodes announce themselves when joining a cluster, as well as metadata that doesn’t have to be consistent (like the amount of used and available disk space).

Let me know if this answers your question.