Without answering your post directly, I've also been doing some experimenting with a 5 node cluster built on openSUSE (your close cousin).
One of my experiments has been to restart the cluster decreasing memory capacities and surprisingly I was able to squeeze memory available to where substantial performance latency was observed but at no time was the swap touched ( except one time when only a couple megabytes was swapped out of the 2+GB total on a couple nodes). And this was with completely default JVM settings (including no mlockall setting). All activity was related to indexing and shard allocation, no heavy queries were performed. Tony -- You received this message because you are subscribed to the Google Groups "elasticsearch" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/c02b6375-51cb-49f1-9d6b-5fc9390da3d9%40googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.
