Multimaster clustering, synchronous replication, low latency
MariaDB Galera Cluster, an open source active-active, multimaster synchronous replication solution for MariaDB Enterprise Server in combination with MariaDB MaxScale, provides high availability, scalability and automated failover for mission-critical transactional workloads. Galera Cluster provides parallel replication and data consistency across all nodes and automatically manages the identification and removal of failed nodes as well as rejoining new or repaired nodes. With Galera Cluster, there’s no chance of lost transactions and significant replica lag or client latency.
If one or more MariaDB Enterprise Server nodes fail, the remaining nodes will continue to operate and failed nodes are automatically taken offline, repaired and reintroduced into the cluster – all without intervention from the application level or any developer intervention.
Unlike MariaDB primary/replica topologies that optimize for read-intensive applications, Galera Cluster is multimaster – all nodes accept write and read operations and synchronize to ensure data consistency across all nodes.
While MaxScale provides read/write splitting for parallel read operations with a single primary and multiple read replicas, Galera Cluster extends distributed load balancing and parallel processing to read and write operations, improving horizontal scale out.
Replicated nodes can be placed across a wide area network (WAN), allowing clients to be located closer to node connections thereby improving performance and reducing response times for end-user web services, particularly read operations.
In the example topology below, a mission-critical transactional workload is supported by MariaDB through MariaDB MaxScale, Enterprise Server and Galera Cluster, deployed as a multimaster cluster, consisting of five database nodes and MaxScale to support front-end load balancing and automatic failover.