What you see if possibly longer latency of a single request, and that
happens because now a single search request will be broken down to be
executed on several shards, with some executing on the other node (network).
In a single node case, there is no network.
On Fri, Oct 14, 2011 at 2:44 AM, Liyu email@example.com wrote:
I am running 2 nodes with 1 index 5 shards 1 replica. My query returns
several facets. I just noticed the query always take longer with 2
Is this behavior expected? Shouldn't the 2nd node takes some of work
in parallel and make the search faster?