Running ElasticSearch in Production

08/20/2013

ElasticSearch is a very awesome tool with very awful documentation.

One thing you’ll need to be sure to do is update the ES_HEAP_SIZE in your /etc/init.d/elasticsearch to use half your system’s memory in production. The default will give you awful performance.

TOTAL_MEM_K=`grep MemTotal /proc/meminfo | awk '{print $2}' `
HALF_MEM_K=$((TOTAL_MEM_K / 2))
ES_HEAP_SIZE="${HALF_MEM_K}k"