Erlang OTP: Distributed & Fault-Tolerant Systems Programming · Lesson

Distributed Mnesia & Replication

Configure Mnesia for distributed operation, including data replication and fault-tolerant storage across a cluster.

Distributed Mnesia & Replication is a free Erlang OTP: Distributed & Fault-Tolerant Systems Programming lesson on CoddyKit — lesson 3 of 4. You can read the complete lesson below for free — then practise it hands-on in the browser with a built-in code editor and a 24/7 AI tutor. It is part of the Erlang OTP: Distributed & Fault-Tolerant Systems Programming learning path, one of 4 lessons in the course, and your progress syncs across the web and the CoddyKit app.

Intro to Distributed Mnesia

Welcome to the final lesson on Mnesia! So far, we've explored Mnesia's fundamentals and how to manage data with transactions. Now, let's unlock its true power: distributed operation.

Distributed Mnesia allows your database to span multiple Erlang nodes, offering incredible benefits for fault tolerance and scalability. Imagine your data staying available even if some servers go down!

Core Concept: Node List

For Mnesia to operate across multiple nodes, it needs to know which nodes are part of its cluster. This is managed through an explicit node list.

Each Mnesia instance on a node is aware of the other nodes.
It uses Erlang's built-in distribution mechanism to communicate.
Data can be replicated to or stored on specific nodes in this list.

Without this configuration, Mnesia would only run locally on a single node.

Initializing a Distributed Database

Setting up a distributed Mnesia cluster involves a few key steps:

Start Erlang Nodes: Launch each Erlang VM with a unique name (e.g., erl -sname node1@host -setcookie mysecret).
Connect Nodes: Ensure all nodes can communicate using net_adm:ping/1.
Create Schema: Use mnesia:create_schema/1 on all participating nodes, passing a list of all node names.
Start Mnesia: Call mnesia:start/0 on each node.

This establishes the foundation for your distributed database.

Code: Basic Mnesia Start

This example shows the basic steps to initialize and start Mnesia. In a real distributed setup, you'd run these on each connected node after creating the schema.

-module(mnesia_starter).
-export([start/0, stop/0, create_schema/0]).

% To run this in a shell:
% erl -sname mynode@localhost -setcookie mysecret
% c(mnesia_starter).
% mnesia_starter:create_schema().
% mnesia_starter:start().

create_schema() ->
    io:format("Creating Mnesia schema on ~p...\n", [node()]),
    % For a multi-node setup, list all node names here:
    % mnesia:create_schema(['node1@host', 'node2@host']).
    mnesia:create_schema([node()]).

start() ->
    io:format("Starting Mnesia on ~p...\n", [node()]),
    mnesia:start().

stop() ->
    io:format("Stopping Mnesia on ~p...\n", [node()]),
    mnesia:stop().

Creating Distributed Tables

Once your Mnesia cluster is set up, you define tables. The key difference for distributed tables is specifying the node_list when creating them.

The node_list attribute determines which nodes will store copies of the table.
You can specify different types of copies (disc_copies, ram_copies, disc_only_copies) for each node.
This allows fine-grained control over data placement and replication.

Mnesia ensures that the table schema is consistent across all nodes in the node_list.

Code: Distributed Table Definition

This example shows how to define a table, specifying that it should have disk copies on specific nodes. In a real scenario, 'node1@host' and 'node2@host' would be actual Erlang node names.

-module(distributed_table_def).
-export([create_user_table/0, start_mnesia/0, stop_mnesia/0]).

create_user_table() ->
    io:format("Attempting to create 'user_info' table...\n"),
    % In a distributed system, this list would contain
    % actual node names, e.g., ['node1@host', 'node2@host'].
    % For this runnable demo, we'll use the current node.
    NodeList = [node()], 

    mnesia:create_table(user_info, [
        {attributes, [id, name, email]},
        {disc_copies, NodeList} % Data stored on these nodes
    ]),
    io:format("Table 'user_info' created (or already exists) \n    with disc_copies on ~p.\n", [NodeList]).

start_mnesia() ->
    mnesia:start().

stop_mnesia() ->
    mnesia:stop().

% To run this:
% 1. Start Erlang shell: erl -sname mynode@localhost -setcookie mysecret
% 2. Compile: c(distributed_table_def).
% 3. Run: distributed_table_def:start_mnesia().
% 4. Run: distributed_table_def:create_user_table().

Replication Types: Disc vs. RAM

Mnesia offers different replication strategies for your table copies:

disc_copies: Data is stored on disk and loaded into RAM on startup. Provides persistence and fault tolerance.
ram_copies: Data is only stored in RAM. Fastest access but data is lost if the node crashes (unless replicated elsewhere).
disc_only_copies: Data is only stored on disk and not loaded into RAM. Slowest access, but uses minimal RAM.

Choosing the right type depends on your persistence, performance, and memory requirements.

Distributed Data Consistency

When you write data to a distributed Mnesia table, Mnesia ensures atomicity across all nodes involved in the transaction. This means:

A transaction either commits successfully on all relevant nodes, or it rolls back on all of them.
Mnesia handles the complexities of two-phase commit protocols behind the scenes.

This guarantees that your data remains consistent, even when spread across a cluster.

Managing Cluster Membership

Erlang's dynamic nature extends to Mnesia clusters. You can add or remove nodes from a running system without downtime.

Use mnesia:change_table_copy_type/3 or mnesia:add_table_copy/2 to add new copies of a table to a node.
Use mnesia:delete_table_copy/2 to remove a copy from a node.

These operations allow you to scale your Mnesia cluster horizontally or perform maintenance.

Fault Tolerance through Replication

This is where distributed Mnesia truly shines! By replicating data across multiple nodes, your system becomes highly fault-tolerant.

If a node with a disc_copies table fails, other nodes with copies can continue serving requests.
Mnesia automatically synchronizes data when a failed node recovers and rejoins the cluster.
You can design your system to withstand multiple node failures based on your replication strategy.

This 'always-on' capability is crucial for critical applications.

Check Your Understanding

Which Mnesia table copy type offers the fastest read/write access but loses data if the node crashes and no other copies exist?

Recap: Distributed Mnesia

Congratulations! You've now learned about configuring and managing distributed Mnesia.

We covered how to set up Mnesia across multiple Erlang nodes.
You saw how to define tables with node_list for distribution.
We explored different replication types: disc_copies, ram_copies, and disc_only_copies.
Finally, you understand how Mnesia ensures consistency and provides fault tolerance through replication.

Mnesia is a powerful tool for building robust, scalable, and fault-tolerant distributed applications in Erlang.

Frequently asked questions

Is the “Distributed Mnesia & Replication” lesson free?

Yes — the full text of “Distributed Mnesia & Replication” is free to read here on the web, and the Erlang OTP: Distributed & Fault-Tolerant Systems Programming course includes 4 lessons in total. To practise it interactively (a built-in code editor and a 24/7 AI tutor) and unlock the rest of the Erlang OTP: Distributed & Fault-Tolerant Systems Programming course, upgrade to CoddyKit PRO.

What will I learn in “Distributed Mnesia & Replication”?

Configure Mnesia for distributed operation, including data replication and fault-tolerant storage across a cluster. You practise Erlang OTP: Distributed & Fault-Tolerant Systems Programming with hands-on code you run directly in the browser, and a 24/7 AI tutor answers your questions as you work through the lesson.

Do I need any experience to start Erlang OTP: Distributed & Fault-Tolerant Systems Programming?

No prior experience is required. Erlang OTP: Distributed & Fault-Tolerant Systems Programming on CoddyKit is structured for beginners through advanced learners; this is — lesson 3 of 4, so you can start here or from the beginning and move at your own pace.

How long does the “Distributed Mnesia & Replication” lesson take?

Most CoddyKit lessons take about 5–10 minutes. Each one is bite-sized and interactive, so you make steady progress and pick up exactly where you left off across the web and the app.

Can I write and run code in this Erlang OTP: Distributed & Fault-Tolerant Systems Programming lesson?

Yes. Every Erlang OTP: Distributed & Fault-Tolerant Systems Programming lesson includes a built-in code editor, so you write and run real code right in your browser and get instant AI feedback — no local setup required.

All lessons in this course

← Back to Erlang OTP: Distributed & Fault-Tolerant Systems Programming