I have a single index with ~1M documents at 3.2GB total. During development
I've been using a single ES node with default VM and ES settings on a box
with 4GB of RAM. I've been happy with performance (most queries under
100ms) but haven't tested this setup under load. I'm preparing to start a
phased roll out to production and need advice about what tweaks I should
make. Production traffic should be relatively low.
If things are good, and dev is similar to production, then go for it. I would suggest setting the same memory MIN/MAX memory value. By default, min is 256m and max is 1g, so set both to 1g.
On Wednesday, January 25, 2012 at 5:49 PM, Shane Witbeck wrote:
I have a single index with ~1M documents at 3.2GB total. During development I've been using a single ES node with default VM and ES settings on a box with 4GB of RAM. I've been happy with performance (most queries under 100ms) but haven't tested this setup under load. I'm preparing to start a phased roll out to production and need advice about what tweaks I should make. Production traffic should be relatively low.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.