Elasticsearch - ilya-khadykin/notes-outdated GitHub Wiki
Elasticsearch is based on Apache Lucene - an open-source search engine. Elasticsearch is dependent on Java
Elasticsearch divides data into shards and replicates those shards across the cluster of multiple servers
Production setup:
- at least 3 master nodes;
- as many data nodes as needed;
- if cluster becomes big (more than 6 or 8 nodes) you could add load balancing nodes (usually 2 for high availability);
Common tasks
Starting Elasticsearch
# starting using shell script
/bin/elasticsearch
Configuration
# JVM
ES_HEAP_SIZE=256m; # usually half of the size of the host
/opt/elasticsearch/config/elasticsearch.yml:
# data serving node
node.master: false
node.data: true
# master node
node.master: true
node.data: false
# load balancing node
node.master: false
node.data: false
Elasticsearch as a Time Series Data Store
Source: https://www.elastic.co/blog/elasticsearch-as-a-time-series-data-store
Moving Averages
https://www.elastic.co/blog/staying-in-control-with-moving-averages-part-1
https://www.elastic.co/blog/staying-in-control-with-moving-averages-part-2