jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alexander Klimetschek" <aklim...@day.com>
Subject Re: Jackrabbit not working on Oracle
Date Thu, 28 Aug 2008 13:35:17 GMT
The interesting exception is this one, created when the
DatabaseFileSystem loads the schema oracle.ddl (in
org/apache/jackrabbit/core/fs/db/ in jackrabbit-core.jar):

java.sql.SQLException: Non supported SQL92 token at position:
213: tableSpace

Not sure (need the workspace.xml files, too), but your config is not
quite right, and this might trigger the problem. In the <Workspace>
element you should always use the placeholder "${wsp.name}" for the
workspace name, so you should set the attribute name to "${wsp.name}"
and you also need a unique name for the db tables: Db file system and
db persistence manager thus need a schemaObjectPrefix that includes
${wsp.name}, for example "JCR_FS_${wsp.name}_" and "JCR_${wsp.name}_".
(The persistence manager inside <Versioning> should be named
"JCR_VERSION_" - it is like a fixed, separate workspace that holds the
global versioning data).

You will also need to delete your workspace.xml files if you want to
start from scratch again. The reason is that Jackrabbit, when it
creates a new workspace, eg. the default workspace during startup,
will take the workspace element, replace all ${wsp.name} with the name
of the new workspace, write this down as workspace.xml into
<repository-home-dir>/workspaces/<workspace-name> (configurable in
Workspaces.rootPath) - now with the replaced names - and load this
configuration for the workspace and eg. open the connections to the
database with those parameters. On the next startup it will find
existing workspace.xml files under <repository-home-dir>/workspaces/*
and provide those workspaces. The workspace element in repository.xml
won't be usd for existing workspaces, but only as a template for new
ones then.

Hope this helps,

Alexander Klimetschek

View raw message