Skip to content
kv

Guides

Task-oriented walkthroughs for building on kv: how the storage engine works, choosing a durability mode, sizing a store, and running the Redis-protocol server.

Each guide is built around a job rather than an API call: understanding the storage engine your data lands in, deciding how much durability to pay for, sizing a store with the Options knobs, and serving it over the Redis protocol. They assume you have worked through the quick start.

Storage engine The storage core behind kv: a sharded hash index over a hybrid hot-tier and cold-log core, why it makes point lookups fast and flat, how it runs larger than memory, and what it does not do. Durability The two durability modes kv offers, the bounded loss window under the default, when to pick which, and what Sync and Close guarantee. Sizing a store How to size a kv store with Options: KeyCapacity as the main memory knob, HotBytes and HotKeys for the write buffer, ResidentBytes for the cold read window, and ReadCacheCells for the read cache. Running the server Serving one kv store over the Redis wire protocol: the flags, unix socket versus TCP, the supported commands, and driving it with redis-cli and Redis client libraries.