Setting up an ElasticSearch container

Introduction

elastic spits on ports 9200 (http) & 9300 (elastic)
based on upstream image and tagged version 2.x.x (https://hub.docker.com/_/elasticsearch/)
re-run without troubles (just keep the volume)
shared accross different applications

Fetch / Update the official image

Fetch/update the Elastic Search image and check, CAREFUL about the version needed for the application e.g. ES v2.x.x,

docker pull elasticsearch:2
docker images

Launch the container

Make sure the data folder doesn’t exist yet otherwise the init part that follows doesn’t make sense,

ls -ld /data/elasticprod/
#sudo rm -rf /data/elasticprod/

Make sure the 9200 & 9300 ports are available on the docker host if you want to expose them,

    netstat -antpe --inet --inet6 | grep LISTEN | egrep '9200|9300'

Launch the docker instance for ES and map the data volume to /data/elasticprod/,

app=elasticprod
volume=/data/$app
docker ps -a | grep $app
docker run -d --name $app -h $app -v $volume:/usr/share/elasticsearch/data elasticsearch:2
docker ps -a | grep $app
docker logs $app

No need for -p 9200:9200 -p 9300:9300 even from the host. Check with,

nmap -p 9200,9300 172.17.0.2

If you need to further tweak the thing,

docker exec -ti $app bash
cd /usr/share/elasticsearch/config/
cp elasticsearch.yml elasticsearch.yml.dist
apt update
apt install vim-tiny
vi elasticsearch.yml

Ready to go

Eventually check that you can access the service from another container,

docker run -ti --link elasticprod:elastic ubuntu
apt -y update && apt -y install netcat
ping -c1 elasticprod
nc -z -v elasticprod 9200 9300
#nmap -p 9200,9300 elasticprod

You can now inject the indexes to be used for the application(s).

Now check that you can see the indexes from that other container,

apt -y update && apt -y install curl
curl 'elasticprod:9200/_cat/indices?v'
curl -XGET 'http://elasticprod:9200/_cluster/health?pretty=true'

Note. the indices command just lists the indexes, it doesn’t actually check if there is something in it. Note. the health command shows you the cluster status. It is green although we only have one node, which is weird (should be yellow).

Operations

Operating Docker Containers

Resources