Hi, there's a file called revision.log inside each cluster's node repository directory (it should be just 8 bytes long). We usually delete that and usually works (using the DB backed journal). Alessandro
On Wed, Mar 26, 2008 at 11:42 AM, Chris Schmidt <[EMAIL PROTECTED]> wrote: > Hi, > > > > We are using an internal Jackrabbit repository to store metadata > information for a product. Currently, we use clustering to allow > multiple processes to interact with the information. I've noticed an > issue when users wipe out their database without deleting the installed > product. It seems information in the workspace has some information > (possibly the journal revision number?) that becomes out of sync when > the JCR tables are recreated. I cannot add or find information through > jackrabbit afterwards. The only way to get beyond this problem is the > manually delete the repository directory as well. We are configured to > have everything in the database, and thought the only information on the > filesystem would be the lucene index. > > > > Is there a way to detect and reset the information on the filesystem > without having to delete it completely? > > > > Thanks, > > > > Chris > > > >
