Database Reference
In-Depth Information
• Enabling data replication from the master node by polling at regular intervals
• Electing the master node
• Exposing system health / monitoring via JMX beans
Neo4j architecture is a self-driven and independent architecture and requires minimum
human intervention or management.
All the nodes in a cluster are self-sufficient and any node, either slave or master, can be
used to read data from the database. Writing is an exception where all writes have to be
persisted or relayed through the master node. Although you can still connect to the slave
nodes and relay your write requests, before returning success, the slave nodes will direct
all requests to the master node, and once the master node confirms the success, only then
will the client receive a success response. So, it is always advisable to relay all writes to
the master node, and reads can be performed from any node.
Neo4j also provides linear scalability where we can add nodes to the existing cluster and
data is asynchronously replicated to the new nodes.
Note
To enable HA / cluster mode, modify the value of
org.neo4j.server.database.mode
as
HA
in
<$NEO4J_HOME>/conf/
neo4j-server.properties
.
