Now that the app is up and running, let’s try it out.
We will vote for cats and dogs, view the results, and monitor the manager and worker nodes, containers and services on a visualizer.
<MANAGER-IP:>5000 in a web browser to view the voting page from a user perspective.
Click on either cats or dogs to vote.
Now, go to
<MANAGER-IP:>5001 in a web browser to view the voting results tally, as one might do in the role of poll coordinator. The tally is shown by percentage in the current configuration of the app.
<MANAGER-IP:>8080 to get a visual map of how the application is
This surfaces some configuration and characteristics from docker-stack.yml, and you can see those strategies in action here. For example:
We have two nodes running: a
manager and a
The manager node is running the PostgreSQL container, as configured by setting
[node.role == manager] as a constraint in the deploy key for the
db service. This service must be constrained to run on the manager in order to work properly.
[node.role == manager]as a constraint in the deploy key for the
visualizerservice. This service must be constrained to run on the manager in order to work properly. If you remove the constraint, and it ends up on a worker, the web page display will be blank.
Two of the services are replicated:
vote(represented in the visulizer by
result(represented in the visulizer by
Both of these services are configured as
replicas: 2 under the
deploy key. In the current state of this app, one of each is running on a manager and on a worker. However, since neither are explicitly constrained to either node in
docker-stack.yml, all or some of these services could be running on either node, depending on workload and re-balancing choices we’ve left to the swarm orchestration.
In the next steps, we’ll customize the app and redploy it.