hi warner On Wed, Sep 17, 2008 at 11:51 PM, Warner Onstine <[EMAIL PROTECTED]> wrote: > Hi all, > Moving along on the project but now we need to make this a clustered app so > I have been looking into getting this setup on our Oracle databases for the > journaling and the main repository and our DBA has a few questions that he > would like answered before we move ahead. > > 1) In the oracle.ddl file for the repository under > jackrabbit-core/src/main/resources/org/apache/jackrabbit/core/persistence/db > there appear to be no foreign key relationships. Why?
because none are needed ;) referential integrity of the jcr data model is handled on a higher level. the tables are just used to store key/value pairs, much like a hashmap or a berkeley db. > 2) In the oracle.ddl file for the journal under > jackrabbit-core/src/main/resources/org/apache/jackrabbit/core/journal why > have a table with just a single column. what's wrong with single column tables? this table just contains a single record with the highest available revision number. > 3) In the Journal schema the Revision table name seems to be a bad choice > for a name as it is a unique number (and I'm guessing a UUID number). While > it may be a revision you should be able to have a revision 1 of one document > and a revision 1 of another document and not confuse the two. this table stores a single integer value (not an UUID), i.e. the highest available revision number. see also: http://jackrabbit.apache.org/api/1.4/org/apache/jackrabbit/core/journal/DatabaseJournal.html#doUnlock(boolean) http://jackrabbit.apache.org/api/1.4/org/apache/jackrabbit/core/journal/DatabaseJournal.html hope this helps, cheers stefan > > Thank you for any clarification on this to help get this moving forward, we > just want some documentation relating to the schema before we push it into > our DBs. > > -warner >
