If you have many documents (say > 10M documents, probably a larger threshold) then you will benefit from sharding your index, i.e. splitting your index up into multiple cores and using distributed searches. You could use one VM and multiple cores just fine, assuming you have multiple CPUs.
If not, then I see no point in using more Java VMs. Java is pretty scalable in the enterprise, you know. ~ David Smiley Author: http://www.packtpub.com/solr-1-4-enterprise-search-server/ On Dec 6, 2009, at 11:56 PM, Amit Nithian wrote: > This may be a silly question but is there any capacity gain if I run > multiple jetty instances each having their own SOLR_HOME where each jetty > instance/solr will replicate their index from a separate cluster of masters? > I have a couple powerful multi-core servers and am not sure if/how a single > JVM takes advantage of multi-cores and feel that I could increase my > resource usage and hence search capacity by running multiple jetty instances > per server as opposed to adding more machines. > > Physical redundancy aside, is this acceptable practice? > > Thanks! > Amit
