Automating Deployments Using Docker Stack
A docker-compose alternative for swarm cluster
We'll cover the following
In this lesson, we will automate all the steps we took to create and deploy our app using services.
Clone the project using
git clone https://github.com/venky8283/Docker.git
Typegit checkout fc0a9af7f7fb2c8d61e4777befdbf451568cf578to get the exact code used in this lesson.
We did the same thing while working with docker-compose. In docker-compose, we created a docker-compose file and automated all the steps in it.
We will modify the same docker-compose file to run with docker-swarm because Compose and Docker stack are similar on some level.
Let’s list all the steps we need to automate as follows:
-
Pull the Flask_app v3.0 image from Docker Hub
-
Increase the number of replicas of the Flask app to three
-
Create a database service
-
Establish communication between services using the network
-
Create a visualizer service to monitor the swarm
Docker stack implementation#
Docker stack is a bundle of services. When you want to deploy a set of services to a machine, instead of creating each service separately, Docker stack allows users to deploy the full-stack of the services at once.
Let’s see changes in our docker-compose file below.
-
Line 1:It is good practice to lock the exact version of Compose we are using. -
Line 7:We removed the build keyword and added an image location for the app. ‘Build’ is not supported in the swarm. -
Line 17-21:We added a deploy keyword to inform Docker swarm to maintain three instances of our app service. Also, we mentioned in which network our service will be located in. -
Line 41-45:We defined the deployment of database containers along with the network information. -
Line 47-65:We Defined the visualizer service. We have defined the image to be pulled from Docker Hub, ports mapping, mounted a Docker socket to sync events and the network to be used by the service. -
Line 71-73:We told the Docker swarm to create a network named app so that it can be used by the services.
Docker service also provides in-depth configuration for deployment at CPU level, but we will not make our docker-compose file more complex.
Since our docker-compose file is ready, let’s deploy the services on the swarm manager node below.
Type docker stack deploy <stack name> --compose-file=<location of docker-compose.yml file>.
＄ docker stack deploy login_app --compose-file=docker-compose.yml
Creating network login_app_app
Creating service login_app_web
Creating service login_app_database
Creating service login_app_viz
This will deploy a stack of all services on swarm manager with just one command.
Now, verify the services using docker service ls.
＄ docker service ls
ID NAME MODE REPLICAS IMAGE PORTS
iv1nkq62i6fx login_app_database replicated 2/2 mysql/mysql-server:5.7
yo7o89dr5s9b login_app_viz replicated 1/1 dockersamples/visualizer:latest *:8080->8080/tcp
rpuagx2yj9zg login_app_web replicated 3/3 venky8283/flask_app:3.0 *:5000->5000/tcp
Docker stack commands#
Docker stack has few commands to manage stack on the swarm node. Type docker stack --help to list them.
＄ docker stack --help
Usage: docker stack [OPTIONS] COMMAND
Manage Docker stacks
Options:
--orchestrator string Orchestrator to use (swarm|kubernetes|all)
Commands:
deploy Deploy a new stack or update an existing stack
ls List stacks
ps List the tasks in the stack
rm Remove one or more stacks
services List the services in the stack
Run 'docker stack COMMAND --help' for more information on a command.
-
deploy: We used this one to deploy our stack. We can update the docker-compose file and redeploy stack -
ls: This provides a short overview of stack
docker stack ls
NAME SERVICES ORCHESTRATOR
login_app 3 Swarm
ps: Lists down all the containers created by the stack
＄ docker stack ps login_app
ID NAME IMAGE NODE DESIRED STATE CURRENT STATE ERROR PORTS
zn1c6ro9f1ld login_app_viz.1 dockersamples/visualizer:latest docker-desktop Running Running 9 minutes ago
sjh64hceb9jh login_app_database.1 mysql/mysql-server:5.7 docker-desktop Running Running 9 minutes ago
ew7zdtczb2su login_app_web.1 venky8283/flask_app:3.0 docker-desktop Running Running 10 minutes ago
luo5j6vjljhv login_app_database.2 mysql/mysql-server:5.7 docker-desktop Running Running 9 minutes ago
ug3v9p7fy7a9 login_app_web.2 venky8283/flask_app:3.0 docker-desktop Running Running 10 minutes ago
9v3q0hfmirxc login_app_web.3 venky8283/flask_app:3.0 docker-desktop Running Running 10 minutes ago
services: Lists all the services in the stack
＄ docker stack services login_app
ID NAME MODE REPLICAS IMAGE PORTS
iv1nkq62i6fx login_app_database replicated 2/2 mysql/mysql-server:5.7
rpuagx2yj9zg login_app_web replicated 3/3 venky8283/flask_app:3.0 *:5000->5000/tcp
yo7o89dr5s9b login_app_viz replicated 1/1 dockersamples/visualizer:latest *:8080->8080/tcp
rm: Removes stack from the swarm nodes, including network
＄ docker stack rm login_app
Removing service login_app_database
Removing service login_app_viz
Removing service login_app_web
Removing network login_app_app
I guess that’s enough to get started with stack. You can always explore more as long as you understand the fundamentals.
In the next lesson, we will review all the commands we have used in this section and look at a cheat sheet with us in case we forget some of them.
