Sharding in Manticore Search: automatic distribution and replication

Manticore can split a table into shards and spread them across a cluster with a single CREATE TABLE statement. On one node, sharding parallelizes work across CPU cores. Across many nodes, it does the harder job: distributing data, replicating each shard to a configurable replication factor, and rebalancing automatically when nodes fail or join. This article explains both use cases, how it works under the hood, how to operate it, and where the current limits are.