要約

Occasionally a Confluence admin may need to use a custom schema on PostgreSQL to store Confluence's objects - a DB migration, for example.   Out-of-the-box, PostgreSQL will place/look for any database objects on its default schema: public.  An admin can make both PostgreSQL and Confluence use a custom schema to place those objects.

環境

PostgreSQL (9.6, 10, 11)

Confluence Server/DC 6.13, 7.4

ソリューション

Confluence side

To ensure your application (Confluence) will have its objects stored on a different schema, you need to add a parameter to the JDBC connection string.

 Add the property below to your <confluence_home_dir>/confluence.cfg.xml (or to your <confluence_install_dir>/conf/server.xml if you're delegating Confluence's database connection pool to Tomcat's datasource):

?currentSchema=<schema name>

例:

Confluence new installation:

Confluence existing installation:

Change confluence.cfg.xml file to use the custom schema:

<property name="hibernate.connection.url">jdbc:postgresql://localhost:5432/conf1317?currentSchema=osi</property>

Or server.xml, if that's the case:

url="jdbc:postgresql://localhost:5432/confluence?currentSchema=osi"

PostgreSQL side

search_path

However, to make it work you'll need also to inform PostgreSQL of where it should store and look for the application's database objects.
To do so you'll want to change your database's default search path:

conf1317=# ALTER DATABASE conf1317 SET search_path TO osi;

要約

Essentially you need to ensure that:

1. All Confluence objects are residing on the schema of your choice (if you're migrating your objects to a custom schema). In a new installation if all settings above are correct, they'll be created there.
2. Confluence knows that it should use this schema (currentSchema parameter)
3. PostgreSQL knows that it should use this schema (search_path adjustment)