Last updated Sep 28, 2021

Clustering in AMPS

Introduction

AMPS 8.3 introduces the ability for Data Center app developers to start up Atlassian Data Center products in clustered mode on their own machine.

In theory, a cluster could be of any desired size. With each node consuming resources (CPU, memory, and more), we recommend that clusters be kept as small as possible, typically two nodes.

How it works

To understand how clustered mode works, let's first review non-clustered mode.

Non-clustered mode

In non-clustered mode:

  • Product: AMPS runs each product in a single Tomcat instance, managed by the Cargo Maven plugin.
  • Database: The product is preconfigured to use an embedded HSQL or H2 database instance, with no user intervention required.
  • License: The product is preconfigured with a license that may or may not be for Data Center.

Clustered mode

By contrast, in clustered mode:

  • Product: AMPS runs each node of each product in its own Tomcat instance, managed by the Cargo Maven plugin. AMPS starts these nodes sequentially.
  • Database: You must provision a supported, non-embedded database for use by all nodes, as well as a home ZIP configured to connect to that database.
  • License: If the default license is not for Data Center, you must provide a Data Center license via your AMPS configuration.

Setting up a cluster

The process of setting up a cluster depends on the Atlassian product you’re using:

Rate this page: