Hi Lucene Users, I work on a product with several thousand clients. We use Lucene to index various client data and make the functionality available as part of our product. Currently, each client has its own index for security reasons, but I am wondering if this is the best way to handle the indexing moving forward. Right now we don't take advantage of the near real-time indexing. As we scale up (and potentially incorporate NRT), I am wondering if an index per client is the best way to architect the system. Maintaining hundreds or more IndexWriters in memory to support NRT seems potentially problematic. Perhaps a better way would be to create one large index (or several large indices) and use a BitSet to to limit the results to only the relevant client. Has any one worked on a system similar to this one and can provide some architecture advice?
-Sean --------------------------------------------------------------------- To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
