Elasticsearch 7.x on Docker Swarm

# Maintainer: Jakub Hajek, jakub.hajek@cometari.com
#
# docker stack deploy -c stack-elastic.yml elastic
#
# The stack creates Elasticsearch cluster consiting of
# - 3 dedicated master nodes in order to keep quorum
# - 4 dedicated data nodes to manage CRUD,
#
# Docker compose file to easily deploy Elasticsearch cluster 7.x on Docker Swarm cluster.

# Make sure to tune your VM to run Elasticsearch cluster by executing on each server:
# sudo sysctl -w vm.max_map_count=262144
# add --default-ulimit memlock=-1 to Dockerd
# You can manage that by executing this command on each server consiting Swarm cluter:
#
# sudo "sed -i '/ExecStart=\/usr\/bin\/dockerd/ s/$/\ --default-ulimit memlock=-1/' /etc/systemd/system/docker.service.d/10-machine.conf"

version: "3.7"

x-default-logs: &default-opts
  logging:
    options:
      max-size: "1m"

x-master-opts: &master
  ES_JAVA_OPTS: "-Xmx2g -Xms2g"
  discovery.seed_resolver.timeout: 60s
  discovery.seed_hosts: master1,master2,master3
  cluster.initial_master_nodes: master1,master2,master3
  cluster.name: "es-cluster"
  network.host: 0.0.0.0
  node.master: "true"
  node.data: "false"
  node.ingest: "false"

x-data-opts: &data
  ES_JAVA_OPTS: "-Xmx2g -Xms2g"
  discovery.seed_resolver.timeout: 60s
  discovery.seed_hosts: master1,master2,master3
  cluster.initial_master_nodes: master1,master2,master3
  cluster.name: "es-cluster"
  network.host: 0.0.0.0
  node.master: "false"
  node.data: "true"
  node.ingest: "false"

services:
  master1:
    image: docker.elastic.co/elasticsearch/elasticsearch:7.4.2
    environment:
      <<: *master
      node.name: "master1"
    networks:
      esnet:
        aliases:
          - master1
    volumes:
      - esmaster1:/usr/share/elasticsearch/data
    deploy:
      placement:
        constraints: [node.labels.master1 ==  true]
      endpoint_mode: dnsrr
      mode: "replicated"
      replicas: 1
      resources:
        limits:
          memory: 4G

  master2:
    image: docker.elastic.co/elasticsearch/elasticsearch:7.4.2
    environment:
      <<: *master
      node.name: "master2"
    networks:
      esnet:
        aliases:
          - master2
    volumes:
      - esmaster2:/usr/share/elasticsearch/data
    deploy:
      placement:
        constraints: [node.labels.master2 ==  true]
      endpoint_mode: dnsrr
      mode: "replicated"
      replicas: 1
      resources:
        limits:
          memory: 4G

  master3:
    image: docker.elastic.co/elasticsearch/elasticsearch:7.4.2
    environment:
      <<: *master
      node.name: "master3"
    networks:
      esnet:
        aliases:
          - master3
    volumes:
      - esmaster3:/usr/share/elasticsearch/data
    deploy:
      placement:
        constraints: [node.labels.master3 ==  true]
      mode: "replicated"
      replicas: 1
      resources:
        limits:
          memory: 4G

  data1:
    image: docker.elastic.co/elasticsearch/elasticsearch:7.4.2
    environment:
      <<: *data
      node.name: "data1"
    healthcheck:
      test: curl -fs http://localhost:9200/_cat/health || exit 1
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 45s
    networks:
      esnet:
        aliases:
          - data1
    volumes:
      - esdata1:/usr/share/elasticsearch/data
    deploy:
      placement:
        constraints: [node.labels.data1 ==  true]
      endpoint_mode: dnsrr
      mode: "replicated"
      replicas: 1
      resources:
        limits:
          memory: 4G
  data2:
    image: docker.elastic.co/elasticsearch/elasticsearch:7.4.2
    environment:
      <<: *data
      node.name: "data2"
    healthcheck:
      test: curl -fs http://localhost:9200/_cat/health || exit 1
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 45s
    networks:
      esnet:
        aliases:
          - data2
    volumes:
      - esdata2:/usr/share/elasticsearch/data
    deploy:
      placement:
        constraints: [node.labels.data2 == true]
      endpoint_mode: dnsrr
      mode: "replicated"
      replicas: 1
      resources:
        limits:
          memory: 4G
  data3:
    image: docker.elastic.co/elasticsearch/elasticsearch:7.4.2
    environment:
      <<: *data
      node.name: "data3"
    networks:
      esnet:
        aliases:
          - data3
    volumes:
      - esdata3:/usr/share/elasticsearch/data
    deploy:
      placement:
        constraints: [node.labels.data3 == true]
      endpoint_mode: dnsrr
      mode: "replicated"
      replicas: 1
      resources:
        limits:
          memory: 4G
  data4:
    image: docker.elastic.co/elasticsearch/elasticsearch:7.4.2
    environment:
      <<: *data
      node.name: "data4"
    networks:
      esnet:
        aliases:
          - data4
    volumes:
      - esdata4:/usr/share/elasticsearch/data
    deploy:
      placement:
        constraints: [node.labels.data4 == true]
      endpoint_mode: dnsrr
      mode: "replicated"
      replicas: 1
      resources:
        limits:
          memory: 4G

networks:
  esnet:
    driver: overlay
    attachable: true
    name: esnet

volumes:
  esdata1:
  esdata2:
  esdata3:
  esdata4:
  esmaster1:
  esmaster2:
  esmaster3:

It would be nice to include doc on the node labels required