Starting a Confluence cluster on a single machine

As part of the work to ensure your plugin is cluster compatible you will need to functionally test it on a running Confluence cluster. One way to do this is to configure a 2 node cluster on your development machine.

1. Download and install the clustered artifact

  1. From the download site, download one of the 'cluster' installations (you can also find EAP versions)
  2. Follow the Confluence Cluster Installation instructions to install Confluence. When you select an interface for the cluster the loopback interface (lo1) is typically a reasonable choice. When you select a shared home directory (new in Confluence 5.5-m19), a local directory can be used.
  3. At this point you will have a single instance running and should have three relevant directories: 
    • a directory containing the distribution, which we will refer to as 'confluence-1' 
    • a home directory which we will refer to as 'home-1'.
    • new in Confluence 5.5-m19: A shared home directory which we will refer to as 'shared-1'.
  4. Stop the server.

When setting up Confluence you can use the following clustered evaluation license. 

This license is provided for the sole purposes of testing your plugin with a Confluence cluster. 


This license is a 1000 user developer license and will expire 72 hours after its initial activation.

2. Create the second node

  1. Copy 'confluence-1' and 'home-1', and name them 'confluence-2' and 'home-2'.
  2. Set the home directory for the second node to 'home-2': 
    1. Edit confluence-2/confluence/WEB-INF/classes/
    2. Alter the confluence.home setting to point to the 'home-2' directory.
  3. Alter the Tomcat server for 'confluence-2' to use different ports from 'confluence-1':
    1. Edit confluence-2/conf/server.xml
    2. The file starts:

      <Server port="8000" shutdown="SHUTDOWN" debug="0">
          <Service name="Tomcat-Standalone">
              <Connector className="org.apache.coyote.tomcat4.CoyoteConnector" port="8090" minProcessors="5"
                         enableLookups="false" redirectPort="8443" acceptCount="10" debug="0" connectionTimeout="20000"
                         useURIValidationHack="false" URIEncoding="UTF-8"/>

      alter the ports and redirectPort attributes so it reads:

      <Server port="9000" shutdown="SHUTDOWN" debug="0">
          <Service name="Tomcat-Standalone">
              <Connector className="org.apache.coyote.tomcat4.CoyoteConnector" port="9090" minProcessors="5"
                         enableLookups="false" redirectPort="9443" acceptCount="10" debug="0" connectionTimeout="20000"
                         useURIValidationHack="false" URIEncoding="UTF-8"/>

      (in this example the 8's have been changed to 9's).

  4. The second server also needs to be configured with a different Hazelcast port from the first. This is done via a system property:

Start the cluster

On some operating systems there can be trouble with the Coherence multicast traffic and IPv6 support. To work around this you need to start each server with the system property

To start the cluster:

    1. Start the first server and wait for start up to complete. This is signified by the message in the log init Confluence is ready to serve.
    2. Start the second server.
    3. On either node, go to General Configuration > Clustering to check your cluster is running.  You should see both nodes reported, similar to this:

Was this page helpful?

Have a question about this article?

See questions about this article

Powered by Confluence and Scroll Viewport