At the moment our lucene index is "per server", so searches in a multi node environment will become out of date without constant reconstruction. We should switch to a JGroups/JMS based indexing queue style index, shared via java's internal infinispan (or something similar).
At the moment our lucene index is "per server", so searches in a multi node environment will become out of date without constant reconstruction. We should switch to a JGroups/JMS based indexing queue style index, shared via java's internal infinispan (or something similar).