On this page we'll guide you through the process of setting up a standalone Synchrony cluster, hosted on your own infrastructure. The ability to run your own Synchrony cluster is only available with a Data Center license. 

Architecture overview

Here's a simplified view of the architecture when you manage Synchrony yourself, in a seperate cluster. Note that this diagram doesn't show communication between nodes. 

Set up a Synchrony standalone cluster

This page will guide you through setting up a Synchrony standalone cluster on your own infrastructure. 

If you're using AWS or Azure, using one of our templates may be a more efficient way to set up Confluence with a standalone Synchrony cluster. 

1 Provision your Synchrony nodes

For the purposes of this guide, we assume you have already provisioned the hardware or virtual instances for your Synchrony nodes. We recommend starting with 2 Synchrony nodes. 

You should allow 2GB memory for Synchrony, and enough disk space for the Synchrony application and logs.   

2 Create the Synchrony home directory

To create the Synchrony directory on your first Synchrony node:

1. Grab the <install-directory>/bin/synchrony directory from one of your Confluence nodes and move it to your new Synchrony node.  We'll call this your <synchrony-home> directory.
2. Copy synchrony-standalone.jar from your Confluence local home directory to your <synchrony-home> directory. 
3. Copy your database driver from your Confluence <install-directory>/confluence/web-inf/lib to your <synchrony-home> directory or other appropriate location on your Synchrony node.

3 Edit the start and stop scripts

We provide scripts to start and stop Synchrony on each node. These need to be edited to add information about your environment:

1. Edit the <synchrony-home>/start-synchrony.sh or start-synchrony.bat file 
2. Enter details for all of the required parameters listed under Configure parameters.
   See Required properties below, for a description of each. 
3. Enter detail for any optional properties you may want to specify. 
   See Optional properties below for a description of each. 
4. Save the file. 
5. Start Synchrony by running the start-synchrony script.
6. Visit http://<SERVER_IP>:<SYNCHRONY_PORT>/synchrony/heartbeat  to check Synchrony is running. 

4 Add additional Synchrony nodes and configure your load balancer

To create your second Synchrony node:

1. Copy your <synchrony-home> directory to the second Synchrony node. 

2. Start Synchrony on that node using the start-synchrony script. As each node joins you'll see something like this in your console.

   Members [2] {
   	Member [172.22.52.12]:5701
   	Member [172.22.49.34]:5701 
   }
   

3. Configure your load balancer for Synchrony traffic. 
   For best results, your load balancer should allow WebSocket connections. SSL connections must be terminated at your load balancer, as Synchrony can't accept HTTPS requests. 
   
   You can choose to use the same load balancer for both Confluence and Synchrony, or two seperate load balancers. When we refer to the Synchrony load balancer, we mean whichever load balancer is handling Synchrony traffic. 
   
   
4. Make sure the Synchrony port (8091) is open. Ports used by Atlassian Applications has a good summary of all ports Synchrony uses in Data Center. This is the only one that needs to be open. 

5 Start Confluence one node at a time

Now that Synchrony is running in a cluster, it's time to get Confluence involved. It is essential that you stop Confluence on all nodes before continuing. 

1. Stop Confluence on all nodes. 

2. Start Confluence on one node with the following system property. 
   This property is used to tell Confluence where to find Synchrony, and prevents Confluence from automatically launching a Synchrony process on your Confluence node.

   -Dsynchrony.service.url=http://<synchrony-load-balancer-url>/synchrony/v1

   For example  http://42.42.42.42/synchrony/v1 or 
   http://synchrony.example.com/synchrony/v1

3. Check that Confluence can connect to Synchrony.  Head to  > General Configuration > Clustering then choose  > Collaborative editing beside the Confluence node you just started.
   
   The Synchrony mode should be Standalone Synchrony cluster. 
   
   
   
   If the mode is 'Managed by Confluence', your Confluence node is not connected to your Synchrony cluster. Make sure you're passing the Synchrony service URL system property correctly. 
4. Repeat this process, starting each Confluence node, one at a time, with the synchrony.service.url. 

See How to check the status of Synchrony for Confluence Data Center for more info on how to check Synchrony is running. 

6 Enable collaborative editing

If you're installing Confluence for the first time, collaborative editing is enabled by default. If you've upgraded from an earlier Confluence version, or have disabled it in the past, collaborative editing may still be disabled. 

To enable collaborative editing:

1. Head to  > General Configuration > Collaborative editing.
2. Choose Change mode.
3. Select On and choose Change.

You can now try editing a page. You'll need to access Confluence via your load balancer.  You can't create or edit pages when accessing a node directly.

Any users who had the editor open before you made this change will need to refresh in order to continue editing, as the Synchrony URL they're connected to will have changed.

Required  properties for Synchrony standalone

These properties only apply when you're running Synchrony standalone in its own cluster. If Synchrony is managed by Confluence (Server or Data Center) these properties don't apply. 

The following properties must be provided in the start-synchrony script. 

Optional properties for Synchrony standalone

These properties only apply if you're running Synchrony standalone in a cluster. 

When you start Synchrony, we pass default values for the properties listed below. You can choose to override these values by specifying any of these properties when you start Synchrony. 

hazelcast.prefer.ipv4.stack

Run Synchrony standalone in an IPv6 environment

If you're running a Synchrony standalone in a cluster in an IPv6 environment, you will need to start Synchrony with the following JVM argument:

-Dhazelcast.prefer.ipv4.stack=false

If you're using the start-synchrony scripts, simply uncomment this line in the script.  

Run Synchrony standalone as a service 

If you're running Synchrony standalone in a cluster, and you'd prefer to run Synchrony as a service on each node, see Run Synchrony-standalone as a service on Linux. 

It's not possible to run Synchrony standalone as a service on Windows. Consider switching to managed Synchrony instead. 

Provide credentials to Synchrony standalone using environment variables 

If you're running Synchrony standalone in a cluster, and you prefer to store sensitive information in your environment, rather than directly in the Synchrony startup scripts you can create a synchronyenv file, and use it to provide your database credentials.  This is only available in Linux environments. 

See Provide credentials to Synchrony standalone using environment variables (Linux)