The stopped Docker containers are still available for troubleshooting. You can create an image of them and run them as new containers to inspect the log files and execute commands in them.
View the standard output of the failed container
docker logs MY_CONTAINER_ID
Run a failing container with a Bash terminal
If a container exists with an error within a few seconds, it can be beneficial to start a terminal window in it to view the log files and execute commands. We will override the entry point of the container to start a Bash terminal.
Create an image of the stopped container
docker commit MY_STOPPED_CONTAINER_ID MY_NEW_IMAGE NAME
Run the saved image as a new container and start a Bash terminal instead of the original entry point
docker run -it --entrypoint bash MY_NEW_IMAGE_NAME
service ( one or multiple instances of the same task, like multiple copies of the same web API )
stack ( one or multiple services that belong together, like a front end web application, middle tier, and database server launch scripted in a .yml file )
The difference between the service and the stack is like docker run vs. docker compose, but in a Docker Swarm cluster.
Docker Swarm Services
Global services will run on every available node once.
The Manager distributes the given number of tasks ( containers and commands to run ) of the replicated services on the nodes based on the desired scale number, that can be one. Once a task is assigned to a node it cannot be moved, it will run on that node until stops or fails.
Docker Swarm Networking
Uses the host’s network stack without any namespace separation, and sharing all of the host’s interfaces.
Docker-managed Linux bridge on the Docker host. By default, all containers created on the same bridge can talk to each other.
An overlay network that may span over multiple Docker hosts. Uses the gossip protocol to communicate between hosts.
The container’s own network stack and namespace, without any interfaces. It stays isolated from every other network, and even its own host’s network.
Establishes connections between container interfaces and parent host interfaces. They can be used to assign IP addresses that are routable on physical networks to containers.
Docker Swarm Load Balancing
Internal load balancing
Internal load balancing is enabled by default. When a container contacts another container in the same Docker Swarm, the internal load balancer routes the request.
External ingress load balancing
To enable the external ingress load balancing, publish the port of the service with the –publish flag. Every node in the cluster starts to listen on the published port to answer incoming requests. If the service does not run a container on the node that received the request, the Routing Mesh will route the request to the node that runs the container on the Ingress Network.
Create a service with an image in a private registry
These instructions will pass the login token from your local client to the Docker Swarm nodes, so those are able to log into the registry and pull the image.
# Save the Docker Registry password in the PASSWORD environment variable
# Log into the Docker Registry
echo $PASSWORD | docker login -u [user] registry.my_registry.com --password-stdin
# Create the service
docker service create \
--name my_service \
The format of the ~/.aws/config file is the following. Make sure to add the word “profile” within the square brackets for every profile you specified in the credentials file, except for the “default” one!!!
[default] region = us-east-1 output = json
[profile my-account] region = us-east-1 output = json