about lucene

[jllort]

If you have a large repository what seems is your case you might restore Lucene index backup. Then you have 2 days old backup, what happens with files updated or modified, well, in this case, should be created a crontab task to force indexing only these documents again based in the changelog information ( not reindexing the whole repository what might take hours or some day in your scenario ).

Another option what we are working in version 7 is solr search engine, what behind is also a Lucene search engine, but what can be configured into a cluster. But version 7 it is not still available for the customer, we consider from September we will start officially releasing it to few numbers of customers and wide from the beginning of the next year.

[jllort]

You should have all the documents what have been changed in the change log.

Make a query to iterate among the list, get the NodeBase object and update with same values ( like a rename operation but without doing it, setting the same name ). This is an approach. Should also consider metadata changes what I'm not 100% sure if when the action I suggested before will also be refreshed, should be made some test ( Otherwise should make two iterations, one for the main object covered with the previous action and another specific for metadata ).