docker-compose_selenium_nutch_solr.yaml

# To execute this docker-compose yml file use `docker-compose -f <file_name> up`
# Add the `-d` flag at the end for detached execution
version: "2"
services:
  # selenium hub
  selenium-hub:
    image: selenium/hub:3.141.5-astatine
    container_name: selenium-hub
    ports:
      - "4444:4444"
    environment:
      - GRID_BROWSER_TIMEOUT=90
      - GRID_TIMEOUT=90
      - GRID_MAX_SESSION=100
    mem_limit: 2000000000
    restart: always

  # selenium node
#  chrome:
#    image: selenium/node-chrome:3.141.5-astatine
#    depends_on:
#      - selenium-hub
#    environment:
#      - HUB_HOST=selenium-hub
#      - HUB_PORT=4444
#      - START_XVFB=false   

  # selenium node
  firefox:
    image: selenium/node-firefox:3.141.5-astatine
    depends_on:
      - selenium-hub
    volumes:
      - /dev/shm:/dev/shm
    environment:
      - HUB_HOST=selenium-hub
      - HUB_PORT=4444
      - START_XVFB=false
    mem_limit: 2000000000
    restart: always


  # apache solr
  solr:
    image: "solr:7.3.1"
    container_name: "solr"
    ports:
      - "8983:8983"
    entrypoint:
      - docker-entrypoint.sh
      - solr-precreate
      - mycore
    depends_on:
      - selenium-hub
    mem_limit: 2000000000
    restart: always

  # apache nutch
  nutch:
    build: .
    container_name: "nutch"
    ports:
      - "5555:5555"
      - "8081:8081"
      - "8080:8080"
    depends_on:
      - solr
    environment:
      - JAVA_HOME=/usr/lib/jvm/java-8-openjdk-amd64
    volumes:
      - ./index-writers.xml:/root/conf/index-writers.xml
      - ./nutch-site.xml:/root/conf/nutch-site.xml
      - ./regex-urlfilter.txt:/root/conf/regex-urlfilter.txt
      - ./agents.txt:/root/conf/agents.txt
      - ./batch_urls:/root/conf/batch_urls
      - ./copy_conf_files.sh:/root/copy_conf_files.sh
        
    stdin_open: true
    mem_limit: 2000000000
    restart: always