Hi, There are many more alternatives available to the JMS bridge. There is also the abilty to do incremental copy of the index over a shared filesystem for instance. That being said, 5 seconds seems really short to me.
I read all this in "Hibernate Search In Action" but I suppose the online material has the reference as well. HTH, Stéphane On Mon, May 4, 2009 at 3:00 AM, no spam <mrs.nos...@gmail.com> wrote: > We are using hibernate search that is an abstraction on top of lucene. Has > anyone used this in a clustered environment? There is built-in JMS ( > master > / slave ) support for this but the assumption is that the slaves get > updates > every 30 minutes or more from the master. Our application requires that > all > indexes be up to date within ~5 seconds after metadata changes. Our index > is not huge, < 500k rows and the amount of index on disk is currently only > a > few hundred megs. > -- Large Systems Suck: This rule is 100% transitive. If you build one, you suck" -- S.Yegge
