Thanks for the response. As it seems this problem is not realy fixed in 1.3.1. I removed the complete index. Also the node node types definitions. If I start the Repository aggain there is only 1 segment. For the deletion there are 3 segments. After the new start the searches are not functioning correct. My XPath search looks like: //node1//[EMAIL PROTECTED]'true'] order by @date descending
Any ideas? Many thanks 2007/9/13, Christoph Kiehl <[EMAIL PROTECTED]>: > > Stefan Ivanov wrote: > > > I use 1.3.1 of Jack Rabbit. I have the problem to loose some nodes and > not > > have them listed in the query results. > > If I access the node directly (per UUID) I can get it. Searches do not > > include the node. A workaround may be to save the node once aggain. > After > > this the node is back in the search result. > > Reproducing this problems is relative easy. Just stop the repository and > > delete the lucene index. Also if you have autoRecover='true'; dosn't > > help.</div> > > The questions are: > > is there a good way to recover the lucene index (manual or > automatically). I > > didn't found somthing useful in web till now? > > Why happens this? We've seen it on a running repository? > > There was a problem with version 1.3 > (https://issues.apache.org/jira/browse/JCR-778) where index segment might > get > lost on recovery after unclean shutdowns. But this is fixed in 1.3.1 which > you > stated you use. I'm not aware of any other problems which lead to nodes > not > being available through the index. > But you can always shutdown your repository and delete your index > directories > (e.g. <repo-home>/workspaces/default/index or > <repo-home>/repository/index). > Jackrabbit will then re-index all nodes. > > Cheers, > Christoph > >
