Distributed SQL with MariaDB Xpand

Overview
Term Definitions
Examples
Resources
FAQ
Reference

MariaDB Enterprise Server 10.5 introduces MariaDB Xpand, which provides distributed SQL, high availability, fault tolerance, write scaling, and horizontal scale-out for transactional workloads with MariaDB Enterprise Server. MariaDB Xpand consists of the Xpand storage engine and the MariaDB Xpand service.

Overview

MariaDB Xpand provides distributed SQL with the MariaDB Xpand service and the Xpand storage engine included with MariaDB Enterprise Server 10.5:

Xpand supports transactional workloads
Xpand is strongly consistent.
Xpand leverages a shared-nothing architecture to efficiently execute distributed SQL.
Xpand provides high availability (HA) and fault tolerance by design.
Xpand scales out horizontally for both reads and writes.

Note

Want to use MariaDB Xpand immediately? Get Started with MariaDB Xpand.

Xpand Architecture

Architecture	Architecture of MariaDB Xpand is designed to provide distributed SQL, high availability, fault tolerance, write scaling, and horizontal scaleout for transactional workloads.
Availability	Xpand stores redundant copies of data on multiple nodes to provide high availability for distributed SQL.
Data Distribution	Xpand stores table and index data separately, and both are split into slices independently. Xpand distributes replicas of these slices through the deployment, ensuring the data remains highly available and fault tolerant.
Fault Tolerance	Xpand provides fault tolerance by maintaining multiple copes of data across the deployment.
Query Evaluation	Data stored on Xpand tables is distributed across all Xpand Nodes in the deployment. When you issue an SQL statement to an Xpand Node, Xpand evaluates the query, optimizes it to form a plan, then retrieves data from slices on relevant Xpand Nodes.
Query Optimizer	Xpand has a sophisticated distributed query planner, compiler, and distributed shared-nothing execution engine.
Rebalancer	The rebalancer process maintains a healthy distribution of data by modifying the placement of data with minimal interruption to user operations and relieving the DBA from the burden of manual data placement.
Strong Consistency	Xpand uses a consistency model that sclaes with a combination of intelligent data distribution, MVCC, and the Paxos Protocol, allowing Xpand to scale writes, scale reads in the presence of write workloads, and provide strong ACID semantics.

Xpand Administration

Administration Tools	How to use Xpand's administration tools.
Allocating Disk Space for Fault Tolerance and Availability	How to allocate disk space that can be used to recover from failures.
Decrease device1 Size	How to shrink `device1`, which is the primary file used by Xpand.
Deploying MariaDB Xpand	How to install and configure an Xpand deployment.
Email Alerts	How to configure Xpand to send email alerts.
Failure and Recovery	How to recover from various failures.
Fast Backup and Restore	How to backup an Xpand deployment.
Group Changes	How various node changes affect the deployment.
Licensing	How to obtain and apply your license.
Network Security	How to configure your network security.
Replication	How to use MariaDB Replication with Xpand.
Scale-In	How to reduce your deployment's capacy by removing nodes.
Scale-Out	How to increase your deployment's capacity by adding nodes.
Zones	How to configure your deployment to use zones.

Monitoring Xpand

Monitoring Overview	How to monitor Xpand.
Grafana	How to monitor Xpand with Grafana.
Zabbix	How to monitor Xpand with Zabbix.

Term Definitions

Term	Definition
System	One running operating system. A system typically answers one or more IP addresses.
Service	One set of processes running on a system. A service typically answers one or more TCP port or UNIX socket. A service is typically started and stopped using systemd.
Process	One or more pid running on a system, associated with a service. A service may include one or more process. A process typically serves one or more TCP connection or socket connection.
Node	One service, or multiple services working in concert, on a system.
Deployment	One service, or more than one service connected by a multi-node technology. A deployment may operate on one system or multiple. A deployment may contain one node or multiple.
Plugin	A component of MariaDB Enterprise Server which may be included in the server function optionally, either at runtime or compile-time, and which adheres to an established plugin architecture within Enterprise Server.
Multi-Node	The case where multiple nodes exist in a deployment.
MariaDB Enterprise Server service	The MariaDB Enterprise Server service that runs on each node in a deployment.
MariaDB Xpand service	The MariaDB Xpand service that runs on each node in a deployment.

Examples

Creating an Xpand Table

To create an Xpand table, use the CREATE TABLE statement with the ENGINE=Xpand option:

CREATE DATABASE hq_sales;

CREATE TABLE hq_sales.invoices (
   invoice_id BIGINT UNSIGNED AUTO_INCREMENT NOT NULL,
   branch_id INT NOT NULL,
   customer_id INT,
   invoice_date DATETIME(6),
   invoice_total DECIMAL(13, 2),
   payment_method ENUM('NONE', 'CASH', 'WIRE_TRANSFER', 'CREDIT_CARD', 'GIFT_CARD'),
   PRIMARY KEY(invoice_id)
) ENGINE = Xpand;

SELECT TABLE_SCHEMA, TABLE_NAME, ENGINE
FROM information_schema.TABLES
WHERE TABLE_SCHEMA='hq_sales'
AND TABLE_NAME='invoices';

Resources

Features

MariaDB Xpand Features

Storage Engines

FAQ

How many nodes should my MariaDB Xpand deployment have?

A MariaDB Xpand deployment should have 3 or more nodes to provide high availability and to avoid a split-brain scenario.

Distributed SQL with MariaDB Xpand

Overview

Xpand Architecture

Xpand Administration

Monitoring Xpand

Term Definitions

Examples

Creating an Xpand Table

Resources

Features

Storage Engines

FAQ

How many nodes should my MariaDB Xpand deployment have?

Reference

Xpand Reference

MariaDB Enterprise Server Reference