Just your normal JVM memory constraints... :-) Seriously, is there no defined maximum. But, the size of the cache needs to be experimented with. Just making the cache larger doesn't necessarily make things run faster. At a certain point, depending on your application, you will reach a point of diminishing returns and you'll just be wasting memory and processing cycles.
Kevin On Fri, Jul 16, 2010 at 6:27 AM, Kelvin Law <[email protected]> wrote: > Thanks, one more question about data cache. Is there a limit on the maximum > CacheSize on a single JVM? > Thanks > > > > > > > http://ehcache.org/documentation/openjpa_provider.html > > > > On Wed, Jul 14, 2010 at 9:09 PM, Rick Curtis <[email protected]> wrote: > > > >> This is something that I've always wanted to experiment with, just have > >> never found the time. AFIK, OpenJPA proper doesn't have support for a > >> shared > >> cache but you could use the OpenJPA TCPRemoteCommitProvider to have > >> distributed cache evictions. This *should* work if you get all of the > >> pieces > >> leaning properly against each other. This method would be quite > problematic > >> as you would have to update your p.xm file each time you wanted to add a > >> new > >> cluster member (in addition to managing the ports). > >> > >> One other option would be to explore using ehcache and openjpa. I can't > >> find > >> a link right now, but hopefully someone can point you in the right > >> direction > >> in the morning. > >> > >> HTH, > >> Rick > >> >
