Home

HA-Ops

Tools for deploying and managing a Neo4j High Availability cluster.

As a labratory project, this wiki has developer-level notes.

Neo4j High-Availability Cluster

A Neo4j High-Availability cluster actually consists of two clusters - a Neo4j GraphDB cluster and a Neo4j Coordinator cluster.

A client application performs read and write operations using any member of the GraphDB cluster.

The GraphDB cluster operates in a master-slave configuration, using the master to synchronize write operations. Membership in the GraphDB cluster is coordinated with the help of the Coordinator cluster.

The coordinator cluster acts as a fault-tolerant registry for membership and master election in the GraphDB cluster.

Deploy What?

For the purpose of these tools, deployment means a client application that uses a Neo4j GraphDB. The tools will keep in mind (though not necessarily support) these considerations:

• Embedded & Server
• Multiple Clients
• Multiple Graph Databases
• Single System & High-Availability Cluster

Exploring the implications:

• Deployment Maths - units and operations
• Deployment Configurations - variations of a GraphDB
• Deployment Evolution - migrations and scalability

Deploy How?

• By Hand - deploying using bash scripts
• With Chef - using Chef to provision Neo4j