Skip to main content

docker-compose-versions differences

 I have captured some of my learnings on Docker compose files and how they differ between versions. Docker compose is a tool used for defining and running multi-container Docker applications. I have used the famous multi-container voting application to illustrate the differences with compose versions.
Following are some questions that I have to tried to answer in this blog:
  • What is the difference between Compose versions 1, 2 and 3?
  • What is the difference between compose, stack and dab formats?
  • What are different ways to run compose files with different compose versions?
  • How does “docker stack deploy” really work?

Compose versions:

Following table captures the main differences between Compose versions:
FeatureCompose v1Compose v2Compose v3
Docker engine version1.9.1+1.10.0+1.13.0+
Compose versionCompose up to 1.6.xCompose 1.6.0+Compose 1.10.0+
Connect containersBridge network, need linksOverlay supported, networks can be used to connect containersOverlay supported, networks can be used to connect containers
VolumeNo named volumeNamed volume supportedNamed volume supported
Swarm capabilityOlder swarmOlder swarmSwarm mode
Deploy sectionNo deploy sectionNo deploy sectionDeploy section with replica present
There could be many reasons why someone would want to use older compose versions. It could either be because they are still using older Docker versions or they already have compose files that are currently in use. I have captured the famous votingapp in all 3 formats to capture the differences:

votingapp in compose v1:

Following is the votingapp in compose v1 format. This establishes container connectivity using links and can be deployed in standalone node or in cluster using old swarm mode.
vote:
    image: docker/example-voting-app-vote:latest
    environment:
      - "constraint:node==swarm-agent-753F9D8C000000"
    ports:
      - "5000:80"
    links:
      - redis

redis:
    image: redis:alpine
    environment:
      - "constraint:node==swarm-agent-753F9D8C000000"
    ports: ["6379"]

worker:
    image: docker/example-voting-app-worker:latest
    environment:
      - "constraint:node==swarm-agent-753F9D8C000000"
    links:
      - redis
      - db

db:
    image: postgres:9.4
    ports: ["5432"]
    environment:
      - "constraint:node==swarm-agent-753F9D8C000000"

result:
    image: tmadams333/example-voting-app-result:latest
    ports:
      - "5001:80"
    environment:
      - "constraint:node==swarm-agent-753F9D8C000000"
    links:
      - db
Following command can be used to deploy:
docker-compose up -d

votingapp in compose v2:

Following is the votingapp in compose v2 format. This establishes container connectivity using networks and can be deployed in standalone node or in cluster using old swarm mode. Containers connected through networks can be deployed in separate nodes and they can talk through overlay network here.
version: "2"

services:
  vote:
    image: docker/example-voting-app-vote:latest
    labels:
     - "com.example.description=Vote"
    ports:
      - "5000:80"
    networks:
      - front-tier
      - back-tier

  redis:
    image: redis:alpine
    ports: ["6379"]
    networks:
      - back-tier

  worker:
    image: docker/example-voting-app-worker:latest
    networks:
      - back-tier

  db:
    image: postgres:9.4
    ports: ["5432"]
    labels:
     - "com.example.description=Postgres Database"
    networks:
      - back-tier

  result:
    image: tmadams333/example-voting-app-result:latest
    ports:
      - "5001:80"
    networks:
      - front-tier
      - back-tier

networks:
  front-tier:
  back-tier:
Following command can be used to deploy:
docker-compose up -d

votingapp in compose v3:

Following is the votingapp in compose v3 format. This establishes container connectivity using networks and can be deployed in standalone node or in new swarm mode. This compose format has the deploy section that controls deployment parameters like number of replicas, constraints etc.
version: "3"
services:

  redis:
    image: redis:alpine
    ports:
      - "6379"
    networks:
      - frontend
    deploy:
      replicas: 2
      update_config:
        parallelism: 2
        delay: 10s
      restart_policy:
        condition: on-failure
  db:
    image: postgres:9.4
    volumes:
      - db-data:/var/lib/postgresql/data
    networks:
      - backend
    deploy:
      placement:
        constraints: [node.role == manager]
  vote:
    image: dockersamples/examplevotingapp_vote:before
    ports:
      - 5000:80
    networks:
      - frontend
    depends_on:
      - redis
    deploy:
      replicas: 2
      update_config:
        parallelism: 2
      restart_policy:
        condition: on-failure
  result:
    image: dockersamples/examplevotingapp_result:before
    ports:
      - 5001:80
    networks:
      - backend
    depends_on:
      - db
    deploy:
      replicas: 1
      update_config:
        parallelism: 2
        delay: 10s
      restart_policy:
        condition: on-failure

  worker:
    image: dockersamples/examplevotingapp_worker
    networks:
      - frontend
      - backend
    deploy:
      mode: replicated
      replicas: 1
      labels: [APP=VOTING]
      restart_policy:
        condition: on-failure
        delay: 10s
        max_attempts: 3
        window: 120s
      placement:
        constraints: [node.role == manager]

networks:
  frontend:
  backend:

volumes:
  db-data:
There are 2 ways to deploy an application in compose v3 format. The first is the traditional option using docker-compose.
docker-compose up -d
The above option ignores the parameters under deploy section.
The second preferred option is to use “docker stack” approach as shown below. With this, the Docker services gets directly deployed in the Swarm mode cluster.
docker stack deploy --compose-file  vote
In “docker stack deploy”, compose file parsing is done by Docker engine itself compared to docker-compose binary handling compose file parsing. There are few options still missing in “docker stack deploy”. 1 of them is environment variables which has been addressed recently and should be available in 17.04 Docker version. Another is related to “extends” feature and is being discussed here. The general goal is get Docker engine have all features of docker-compose so that “docker stack deploy” would become the universal way to deploy compose files.

Difference between compose, stack and dab file formats:

Stack file is same as compose file with additional deploy section. Stack file is part of Compose v3 format. If we are using new Swarm mode, it is better to move to stack file format which is the same as compose v3.
Distributed application bundle(Dab) was introduced in Docker 1.12 as an experimental feature. dab file specifies container image with digest rather than defining container image with tags. dab file is auto-generated from docker-compose file. Many options of compose is not supported in dab. “docker stack” introduced in Docker 1.13 has the capabilities of compose and dab combined. My understanding is “docker stack” will make dab format obsolete. I checked with few folks in Docker team and this seems to be the case though it is not fully confirmed.
To generate dab from compose file, we can use the following command:
docker-compose bundle -o 
To deploy the bundle, we can use the docker stack command:
docker stack deploy --bundle-file  
Following is the votingapp in dab format:
{
  "Services": {
    "db": {
      "Image": "postgres@sha256:c681c72ff62b637b9ad3806a946f14450f74ee0888ca70ec1472a4e0552054c2", 
      "Networks": [
        "backend"
      ]
    }, 
    "redis": {
      "Image": "redis@sha256:95f0c9434f37db0a4f9563379184ff4895685d2381e2d81bd10dfcbcadfc095f", 
      "Networks": [
        "frontend"
      ], 
      "Ports": [
        {
          "Port": 6379, 
          "Protocol": "tcp"
        }
      ]
    }, 
    "result": {
      "Image": "dockersamples/examplevotingapp_result@sha256:83b568996e930c292a6ae5187fda84dd6568a19d97cdb933720be15c757b7463", 
      "Networks": [
        "backend"
      ], 
      "Ports": [
        {
          "Port": 80, 
          "Protocol": "tcp"
        }
      ]
    }, 
    "visualizer": {
      "Image": "dockersamples/visualizer@sha256:f924ad66c8e94b10baeaf7bdb9cd491ef4e982a1d048a56a17e02bf5945401e5", 
      "Networks": [
        "default"
      ], 
      "Ports": [
        {
          "Port": 8080, 
          "Protocol": "tcp"
        }
      ]
    }, 
    "vote": {
      "Image": "dockersamples/examplevotingapp_vote@sha256:8e64b18b2c87de902f2b72321c89b4af4e2b942d76d0b772532ff27ec4c6ebf6", 
      "Networks": [
        "frontend"
      ], 
      "Ports": [
        {
          "Port": 80, 
          "Protocol": "tcp"
        }
      ]
    }, 
    "worker": {
      "Image": "dockersamples/examplevotingapp_worker@sha256:3e4dd0f59c15f432280a2c0679c4fc5a2ee5a797023c8ef0d3baf7b1385e9fed", 
      "Networks": [
        "frontend", 
        "backend"
      ]
    }
  }, 
  "Version": "0.1"
}

How does “docker stack deploy” really work?

I have prepared the flowchart below based on referring the notes from this issue. The flowchart below shows how the image specified in the compose file resolves to what gets deployed finally when “docker stack deploy” is used.
compose2
“docker stack deploy” is idempotent. If the stack is already deployed, “docker stack deploy” will restart only those services which has the digest or tag that is updated. In the above flowchart, the restart applies to the 2 cases where the latest image is pulled from registry.

Popular posts from this blog

How to read or extract text data from passport using python utility.

Hi ,  Lets get start with some utility which can be really helpful in extracting the text data from passport documents which can be images, pdf.  So instead of jumping to code directly lets understand the MRZ, & how it works basically. MRZ Parser :                 A machine-readable passport (MRP) is a machine-readable travel document (MRTD) with the data on the identity page encoded in optical character recognition format Most travel passports worldwide are MRPs.  It can have 2 lines or 3 lines of machine-readable data. This method allows to process MRZ written in accordance with ICAO Document 9303 (endorsed by the International Organization for Standardization and the International Electrotechnical Commission as ISO/IEC 7501-1)). Some applications will need to be able to scan such data of someway, so one of the easiest methods is to recognize it from an image file. I 'll show you how to retrieve the MRZ information from a picture of a passport using the PassportE

How to generate class diagrams pictures in a Django/Open-edX project from console

A class diagram in the Unified Modeling Language ( UML ) is a type of static structure diagram that describes the structure of a system by showing the system’s classes, their attributes, operations (or methods), and the relationships among objects. https://github.com/django-extensions/django-extensions Step 1:   Install django extensions Command:  pip install django-extensions Step 2:  Add to installed apps INSTALLED_APPS = ( ... 'django_extensions' , ... ) Step 3:  Install diagrams generators You have to choose between two diagram generators: Graphviz or Dotplus before using the command or you will get: python manage.py graph_models -a -o myapp_models.png Note:  I prefer to use   pydotplus   as it easier to install than Graphviz and its dependencies so we use   pip install pydotplus . Command:  pip install pydotplus Step 4:  Generate diagrams Now we have everything installed and ready to generate diagrams using the comm

Python questions and answers part 3

Q1).What is Python? Ans1:   Python  is a high-level, interpreted, interactive and object-oriented scripting language. Python is designed to be highly readable. It uses English keywords frequently where as other languages use punctuation, and it h as fewer syntactical constructions than other languages. Q2).Name some of the features of Python. Ans2:  Following are some of the salient features of  python It supports functional and structured programming methods as well as OOP. It can be used as a scripting language or can be compiled to byte-code for building large applications. It provides very high-level dynamic data types and supports dynamic type checking. It supports automatic garbage collection. It can be easily integrated with C, C++, COM, ActiveX, CORBA, and Java. Q3).Do you have any personal projects? Really? Ans3: This shows that you are willing to do more than the bare minimum in terms of keeping your skillset up to date. If you work on personal projects and