You are currently browsing legacy 4.1 version of documentation. Click here to switch to the newest 5.1 version.
Rachis - RavenDB's Raft Implementation
Rachis is the RavenDB's Raft implementation.
Definition of Rachis: Spinal column, also the distal part of the shaft of a feather that bears the web.
What is Raft ?
Raft is a simple and easy-to-understand distributed consensus protocol.
It allows you to execute an ordered set of operations across your entire cluster.
This means that you can apply a set of operations on a state machine, and have the same final state machine in all the cluster's nodes.
So any series of events (called a Raft Commands) will be executed in the same order on each node.
The Leader Node accepts all the Raft Commands requests for the cluster and handles committing them cluster-wide.
These Raft Commands are done only if the majority of the nodes in the cluster agreed to it !
What is Rachis ?
Rachis is RavenDB's Raft implementation with the following added features:
- Support for in memory and persistent large multi-tasks state machines.
- Reliably committing updates to a distributed set of state machines.
- Support for voting & non-voting cluster members, see Cluster Toplogy.
- Dynamic topology, nodes can be added and removed from the cluster on the fly.
- Managing situations such as handling a Leader timeout and forcing a leader to step down.
- ACID local log using the Voron Storage Engine.