Overview of the split database solution
ÃÛ¶¹ÊÓƵ Commerce offers several scalability advantages, including the ability to use three separate master databases for different functional areas of the Commerce application.
Checkout, orders, and product data can each use a separate master database that you can optionally replicate. This separation independently scales load from website checkouts, order management activities, website browsing, and merchandising activities, depending on your needs. These changes provide considerable flexibility in how the database tier can be scaled.
The ResourceConnections class provides the unified MySQL database connection to the Commerce application. For queries to the master databases, we implement the Command Query Responsibility Segregation (CQRS) database pattern. This pattern handles the logic for routing the read and write queries to the appropriate databases. Developers do not need to know which configuration is being used and there are no separate read and write database connections.
If you set up optional database replication, you get the following advantages:
- Data backup
- Data analysis without affecting the master database
- Scalability
MySQL databases replicate asynchronously, which means slaves do not need to be connected permanently to receive updates from the master.
The following figure shows how this feature works.
In Magento Open Source, only one master database is used.
ÃÛ¶¹ÊÓƵ Commerce uses three master databases and a configurable number of slave databases for replication. ÃÛ¶¹ÊÓƵ Commerce has a single interface for database connections, resulting in faster performance and better scalability.
Configuration options
Because of the way the split database performance solution is designed, your custom code and installed components cannot do any of the following:
- Write directly to the database (instead, you must use the ÃÛ¶¹ÊÓƵ Commerce database interface)
- Use JOINs that affect the sales or quote databases
- Use foreign keys to tables in the checkout, sales, or main databases
- Ask the component developers to update their components.
- Use the components as-is without the split database solution.
- Remove the components so you can use the split database solution.
This also means you can either:
-
Configure the split database solution before putting Commerce into production.
ÃÛ¶¹ÊÓƵ recommends configuring split databases as soon as possible after you install the Commerce software.
-
Manually configure the split database solution.
You must perform this task if you have already installed components or if Commerce is already in production. (Do not update a production system; make the updates in a development system and synchronize the changes after you have tested them.)
note warning WARNING You must back up the two additional database instances manually. Commerce backs up only the main database instance. The magento setup:backup --dbcommand and Admin options do not back up the additional tables.
Prerequisites
The split database requires you to set up three MySQL master databases on any host (all three on the Commerce server, each database on a separate server, and so on). These are the master databases and they are used as follows:
- One master database for checkout tables
- One master database for sales tables (also referred to as Order Management System, or OMS, tables)
- One master database for the remainder of the Commerce 2 application tables
In addition, you can optionally set up any number of slave databases that serve as load balancers and backups.
This guide discusses how to set up the master databases only. We provide sample configurations and references for you to set up slave databases if you wish.
In this guide, the three master databases are named:
magento_quotemagento_salesmagento
(You can name your databases anything you wish.)