Configuring a MySQL Datasource in Apache Tomcat

This page tells you how to set up a MySQL datasource connection for Confluence.

Step 1. Shut down Tomcat

  1. Run bin/ or bin/shutdown.bat to bring Tomcat down while you are making these changes.
  2. Make a backup of your <CONFLUENCE_HOME>/confluence.cfg.xml file and your <CONFLUENCE_INSTALLATION>/conf/server.xml file, so that you can easily revert if you have a problem.

Step 2. Install the MySQL database driver

  1. Download the MySQL JDBC driver. Links are available on this page: Database JDBC Drivers.
  2. Unpack the archive file you have downloaded, and find the JAR file called something like this: mysql-connector-java-x.x.x-bin.jar, where x.x.x is a version number.
  3. Copy the JAR file into the lib folder of your Confluence installation: <CONFLUENCE_INSTALLATION>/lib.

Step 3. Configure Tomcat

  1. Edit the conf/server.xml file in your Tomcat installation.
  2. Find the following lines:

  3. Insert the DataSource Resource element within the Context element, directly after the opening <Context.../> line,  before Manager:

    • Replace the username and password parameters with the correct values for your database.
    • In the url parameter, replace the word 'confluence' with the name of the database your Confluence data will be stored in.
  • If switching from a direct JDBC connection to a datasource connection, you can find the above details in your <CONFLUENCE_HOME>/confluence.cfg.xml file.
  • The configuration properties for Tomcat's standard datasource resource factory (org.apache.tomcat.dbcp.dbcp.BasicDataSourceFactory) are as follows:
    • driverClassName – Fully qualified Java class name of the JDBC driver to be used.
    • maxTotal – The maximum number of database connections in the pool at the same time. 
    • maxIdle – The maximum number of connections that can sit idle in this pool at the same time.
    • maxWaitMillis – The maximum number of milliseconds that the pool will wait (when there are no available connections) for a connection to be returned before throwing an exception.
    • password – Database password to be passed to your JDBC driver.
    • url – Connection URL to be passed to your JDBC driver. (For backwards compatibility, the property driverName is also recognised.)
    • user – Database username to be passed to your JDBC driver.
    • validationQuery – SQL query that can be used by the pool to validate connections before they are returned to the application. If specified, this query must be an SQL SELECT statement that returns at least one row.
  • Why is the validationQuery element needed? When a database server reboots, or there is a network failure, all the connections in the connection pool are broken and this normally requires an application server reboot. However, the Commons DBCP (Database Connection Pool) which is used by the Tomcat application server can validate connections before issuing them by running a simple SQL query, and if a broken connection is detected, a new one is created to replace it. To do this, you will need to set the validationQuery option on the database connection pool.

Step 4. Configure the Confluence web application

  1. Edit this file in your Confluence installation: <CONFLUENCE_INSTALLATION>/confluence/WEB-INF/web.xml.
  2. Insert the following element just before </web-app> near the end of the file:

If you are changing an existing Confluence installation over to using a Tomcat datasource:

  1. Edit the <CONFLUENCE_HOME>/confluence.cfg.xml file.
  2. Delete any line that contains a property that begins with hibernate.
  3. Insert the following at the start of the <properties> section.

Step 5. Restart Tomcat

Run bin/ or bin/startup.bat to start Tomcat with the new settings.

Was this helpful?

Thanks for your feedback!

Why was this unhelpful?

15 Archived comments

  1. User avatar

    William Crighton

    The Configure the Confluence web application instructions give me an error with Tomcat. Specifically, when I add my resource-ref section such that it appears directly under the <description> line I get an error on startup saying the xml file could not be parsed. To resolve this I moved the <resource-ref> section to the end of the file and it starts up/works fine.

    19 Aug 2004
  2. User avatar


    I have a few databases, confluencedb, which stores the confluence stuff, and webdb, which stores some web-related stuff. I would like to put some of the things in the webdb on the wiki, too. I thought/think that's possible with the sql plugin.

    However, I keep getting "sql: Security restricted macro with parameter 'datasource' and parameter value 'databasename' is not allowed on this page".

    Is there a way to find out more about the error? The logs don't look like Confluence is even attempting to access webdb. I tried changing the names of the datasource to various configurations (for example, connecting to confluencedb, which I know it's using b/c I can login and see content on the wiki), and commented out the restrictions in just for testing, but I keep getting the same error.

    05 Dec 2006
    1. User avatar

      Don Willis [Atlassian]

      Because running arbitrary SQL on a server is not something you want to allow everybody to do, the sql macro, when installed, is disabled by default for security.


      11 Dec 2006
  3. User avatar

    Benjamin White

    When configuring Confluence, the instructions direct to remove all properties that begin with '''hibernate'''. We have properties begining with '''hibernate.3cp0'''. Will the '''3cp0''' statements removal cause a problem?

    26 Sep 2008
  4. User avatar

    Erik Ellestad

    After upgrading from confluence 2.9.2 to 2.10.3 this seems to no longer be working.

    I just get the following error about, "sql: Unexpected program error: com.atlassian.renderer.v2.macro.MacroException: org.apache.tomcat.dbcp.dbcp.SQLNestedException: Cannot create PoolableConnectionFactory (Could not create connection to database server. Attempted reconnect 3 times. Giving up.)"

    20 May 2009
  5. User avatar

    Tim Eddelbuettel

    Have you copied the MySQL Java Connector to \...\confluence\WEB-INF\lib ?


    20 May 2009
    1. User avatar

      Erik Ellestad

      Yes, we are using a MySQL backend for confluence which is working fine.  It's just the data source that isn't.

      22 May 2009
      1. User avatar

        Erik Ellestad

        I figured it out.  The MySQL host cache needed to be cleared on the database server.

        22 May 2009
      1. User avatar


        any solution to that ?

        I did step by step this manual on 2 fresh freebsd8 VMs and:

        any ideas here ? tomcat v6 , mysql5 what a reason of not working this could be ?

        19 May 2010
        1. User avatar


          Ever found a solution?

          21 May 2011
  6. User avatar


    <res-type>javax.sql.Datasource</res-type> should be <res-type>javax.sql.DataSource</res-type> with a capital 'S'.

    22 Jun 2011
  7. User avatar


    how can i do this without shutting down tomcat or interrupting the service?

    I have a SaaS application with several  db, one for each client and i can't shutdown everytime i add a new client.


    27 Nov 2013
    1. User avatar

      Niraj Bhawnani

      Is using a direct JDBC connection an option? That wouldn't require a restart of Tomcat for each new client.

      Also note that deploying multiple Atlassian applications in the same container is not supported: Deploying Multiple Atlassian Applications in a Single Tomcat Container.

      02 Dec 2013
  8. User avatar

    Mark Symons

    Are the Tomcat datasource instructions in step 4 still correct for Tomcat v7?

    I followed the steps exactly as detailed and, after startup of Confluence v5.6.1, I saw that Confluence had moved the the 3 "hibernate" properties within the file, changed them,  and added a 4th "hibernate" property to the file.

    There are no errors.  But the difference cost me an hour of double-checking and also means that the active copy of confluence.cfg.xml does not match what's been committed to our Subversion.

    13 Sep 2014
Powered by Confluence and Scroll Viewport