# Java Persistence API (JPA) Persistence
This service allows you to persist state updates using a SQL or NoSQL database through the Java Persistence API (opens new window). The service uses an abstraction layer that theoretically allows it to support many available SQL or NoSQL databases.
It will create one table named historic_item where all item states are stored.
The item state is stored in a string representation.
The service currently supports Apache Derby, MariaDB, MySQL and PostgreSQL databases. Only the embedded Apache Derby database driver is included. Other drivers must be installed manually. (See below for more information on that.)
# Configuration
This service can be configured in the file services/jpa.cfg.
| Property | Default | Required | Description |
|---|---|---|---|
| url | Yes | JDBC connection URL. Examples:jdbc:derby://hab.local:1527/openhab;create=truejdbc:mariadb://localhost:3306/openhabjdbc:mysql://localhost:3306/openhabjdbc:postgresql://hab.local:5432/openhab | |
| driver | Yes | database driver. Examples:com.mysql.jdbc.Driverorg.apache.derby.jdbc.ClientDriver``org.mariadb.jdbc.Driverorg.postgresql.DriverOnly the Apache Derby driver is included with the service. Drivers for other databases must be installed manually. This is a trivial process. Normally JDBC database drivers are packaged as OSGi bundles and can just be dropped into the addons folder. This has the advantage that users can update their drivers as needed. The following database drivers are known to work:postgresql-9.4-1203-jdbc41.jarpostgresql-9.4-1206-jdbc41.jar | |
| user | if needed | database user name for connection | |
| password | if needed | database user password for connection | |
| syncmappings | if needed | The OpenJPA synchronize mappings configuration |
# Adding support for other JPA supported databases
All item- and event-related configuration is done in the file persistence/jpa.persist.
If a database driver is not an OSGi bundle, the technique below can be used to extend the openHAB classpath.
Other database drivers can be added by expanding the openHAB classpath.
Use the following classpath setup in start.sh (opens new window) / start_debug.sh of openHAB:
cp=$(echo lib/*.jar | tr ' ' ':'):$(find $eclipsehome -name "org.eclipse.equinox.launcher_*.jar" | sort | tail -1);
This will add all .jar files in a folder "lib" in the root of openhab.
All databases that are supported by JPA can be added.
Define driver and url according to the database definitions.