Installation and Administration

JDBC Datasource Configuration

Datasource Definition

Nuxeo code and Nuxeo configuration uses JDBC connections for a number of purposes, and therefore defines different datasource names for them. Examples are:

  • jdbc/nxsqldirectory
  • jdbc/nxaudit-logs
  • jdbc/NuxeoDS

Additional datasources can be used, for instance when defining a Directory you have to specify which datasource it uses, and you may want to use something else than the standard  jdbc/nxsqldirectory if you want to store data elsewhere.

These datasources are all defined in Tomcat conf/Catalina/localhost/nuxeo.xml (which is generated from templates/common-base/conf/Catalina/localhost/nuxeo.xml.nxftl — to defined new datasource names you should copy this template and override it). The nuxeo.xml datasources are defined like this:

<ResourceLink name="jdbc/NuxeoDS" global="jdbc/nuxeo" type="javax.sql.DataSource" />
<ResourceLink name="jdbc/nxsqldirectory" global="jdbc/nuxeo" type="javax.sql.DataSource" />
...

So by default they are actually links to a global resource defined in Tomcat's conf/server.xml (generated from templates/common-base/conf/server.xml.nxftl):

<Resource name="jdbc/nuxeo" auth="Container" type="javax.sql.DataSource"
    maxActive="${nuxeo.db["max-pool-size"]}" maxIdle="30" maxWait="10000" driverClassName="${nuxeo.db.driver}"
    url="${nuxeo.db.jdbc.url}" validationQuery="${nuxeo.db.validationQuery}"
    username="${nuxeo.db.user}" password="${nuxeo.db.password}"
    accessToUnderlyingConnectionAllowed="true" />

The properties used in this file come from the ones defined in bin/nuxeo.conf and the template chosen for your database.

Single-Datasource Mode (non-XA)

By default a mode called single-datasource is activated (see NXP-10308 for technical details).

In this mode, every database connection made by the Nuxeo Platform is funneled through a single physical datasource, including:

  • Directory connections
  • VCS connections
  • Hibernate connections

Even if you defined a different datasource name for a directory, with this mode activated a single shared datasource will actually be used instead. This allows the use of a regular datasource for everything, and avoids the use of XA, which is a big performance boost for some databases (like Oracle in RAC mode).

Single-datasource mode applies even to VCS, which means that in this mode the <xa-datasource> configuration and the various database <property> defined by nxserver/config/default-repository-config.xml will not be used.

To activate single-datasource mode, the following is defined in templates/common-base/nuxeo.defaults:

nuxeo.db.singleDataSource=jdbc/NuxeoDS

The datasource name can of course be different than jdbc/NuxeoDS, but it must be a valid datasource already defined elsewhere.

In addition, the maximum pool size for the datasource must be enough for all the VCS connections, which means that you must have nuxeo.db.max-pool-sizenuxeo.vcs.max-pool-size + 2. (The + 2 comes from two reserved connections for the VCS lock manager and the cluster node handler.)

As an advanced feature, if you want single-datasource mode except for some specific datasources (that need to hit a separate database for instance), then you can use:

nuxeo.db.singleDataSource.exclude=jdbc/myExcludedDatasource, jdbc/anotherExcludedDatasource

This feature can be interesting if you want to define a vocabulary whose values should be retrieved from a database, which is not the one where Nuxeo will store the documents.

The excluded datasources will not participate in any global XA transaction so there will be no two-phase commit for them, and on error and rollback (which admittedly should not happen) they may become inconsistent.

Disabling Single-Datasource Mode

You will have to disable single-datasource mode in these situations:

  • You have more than one SQL repository
  • You want to go back to the old XA mechanism for multiple datasource

If either of this is the case, then you will have to disable single-datasource mode, by redefining the property in bin/nuxeo.conf with an empty value:

nuxeo.db.singleDataSource=

Using Single-Datasource Mode in Java Code

If you code a new Java component that needs to access JDBC directly and wish to use single-datasource mode, then you should get your connections through:

        // try single-datasource mode first
        Connection connection = ConnectionHelper.getConnection(dataSourceName);
        try {
            if (connection == null) {
                // standard mode, get the connection through regular JDBC lookup
                connection = DataSourceHelper.getDataSource(dataSourceName).getConnection();
            }
            ... use the connection ...
        } finally {
            if (connection != null) {
                connection.close();
            }
        }

3 months ago manonlumeau Relocated assets to fit new URL structure for 60 - fix
5 years ago Solen Guitter 8 | ormattin
5 years ago Thierry Martins 7 | example on single datasource exclusion
6 years ago Solen Guitter 6 | Added TOC
6 years ago Florent Guillaume 5
6 years ago Florent Guillaume 4
9 years ago Florent Guillaume 2
9 years ago Florent Guillaume 3 | Migrated to Confluence 4.0
9 years ago Admin name placeholder 1
History: Created by Admin name placeholder

We'd love to hear your thoughts!

All fields required