On 8/31/2011 2:39 PM, Rick Bullotta wrote: > What is probably happening is that the relationship references might be > concentrated around a couple of nodes, forcing a large number of them to be > in memory. > > I can't really tell without understanding the model and how you're > "randomizing" relationships. > well, with 1,000,000 nodes and each node having exactly one relations to a randomly chosen other node, it would be very unlikely for a node to have >10 to 20 relationships. In any event, is there an explanation some place of exactly what stays in memory and what is on the disk in the datastore when my graph gets big like this? What can I do minimize the amount of java heap used?
Gautam _______________________________________________ Neo4j mailing list [email protected] https://lists.neo4j.org/mailman/listinfo/user
