Ultimately it depends on your specific usage patterns. Generally speaking, if you have IndexReaders (and do not use their delete functionality) you don't need locking at all; you can use a no-op lock factory, in which case you'll pretty much only be constrained by your storage subsystem.
Kay Kay-3 wrote: > > For one of our projects - we were planning to have the system of > multiple individual Lucene readers (just read-only instances and no > writes whatsoever ) in different physical machines having their > IndexReader-s warmed up from the same directory for the indices and > working on the same. > > I was reading about locks (implemented as files) that Lucene uses > internally. I am just curious if using multiple readers would be a > feasible option here, all sharing the same index directory (across NFS / > similar network mounted storage ) in terms of locking etc. > > Would there be a performance hit ( ignoring the NFS related performance > of course) that would hinder multiple readers to serve query search > simultaneously from the same set of index files. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org > For additional commands, e-mail: java-user-h...@lucene.apache.org > > > ----- -- http://www.tomergabel.com Tomer Gabel -- View this message in context: http://www.nabble.com/Multiple-IndexReaders-from-the-same-Index-Directory---issues-with-Locks---performance-tp21136262p21142273.html Sent from the Lucene - Java Users mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: java-user-unsubscr...@lucene.apache.org For additional commands, e-mail: java-user-h...@lucene.apache.org
