Hi Julian, On 5/7/07, Julian <[EMAIL PROTECTED]> wrote:
There was a statement made in this thread about Jackrabbit requiring complete control over the database connection. Does this mean that a JNDI connection cannot be used for clustering.
Sorry, I was not precise. By "complete" control, I meant that the database connection - either acquired through JDBC DriverManager calls or via JNDI datasource - must not be manipulated by some other component outside Jackrabbit. For instance, rolling back some change that Jackrabbit's PersistenceManager thinks has already been committed to the database will render inconsistent Jackrabbit's in-memory representation.
that the BLOBs be stored in the database for a cluster (not looking to start a flame about BLOBs in a database vs. on the FS, but some may prefer FS)? I don't mean to threadjack, but this seems relevant to the subject.
There is some other thread on the developer list focusing on this subject: http://www.nabble.com/Jackrabbit-1.2.3,-RecordInput-DatabaseJournal-t3655088.html Kind regards Dominique
Thanks, Julian __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
