Sergio, First, sorry for not replying earlier to the initial email you sent me. As mentioned before, I still know very little about CDSM, while it is necessary for me to dive into that code like... now! Anyways, one thing worth mentioning here is that CDSM has a local cache enabled by default (iirc). Might be something you'd want to look into as well for your use case. Not much more detail I can give right now though :/ Alex On Friday 8 April 2011 at 10:19, Sergio Bossa wrote: > On Fri, Apr 8, 2011 at 12:34 AM, Tim Eck <t...@terracottatech.com> wrote: > > > For distribution purposes for sure concurrency == 1 is fine when there is > > a single active L1 in your cluster. You still might potentially see some > > more throughput with slightly higher concurrency levels with something > 1 > > though. > > Great, thanks. > > > I'm pretty sure there is a task to look into whether the current default > > (2048) is a good choice since it does have some downsides at the moment > > (in particular I think initial faulting is slower because of this). > > Indeed, that was the reason I started looking at the source code: due > to the 2048 maps, my application startup was very slow and ended up in > OOM! > > Cheers, > > Sergio B. > > -- > Sergio Bossa > http://www.linkedin.com/in/sergiob > _______________________________________________ > tc-dev mailing list > tc-dev@lists.terracotta.org > http://lists.terracotta.org/mailman/listinfo/tc-dev >
_______________________________________________ tc-dev mailing list tc-dev@lists.terracotta.org http://lists.terracotta.org/mailman/listinfo/tc-dev
