joebibe/vagrant-docker-swarm repository overview

⁠Docker Swarm

Sample 3 node Vagrant setup for getting started with Docker swarm

⁠Installation

Install Vagrant Binary from here⁠
Download debian/stretch64 box from vagrant using the command vagrant box add debian/stretch64
Clone this repo

git clone https://github.com/joebibe/vagrant-docker-swarm.git  
cd vagrant-docker-swarm  

⁠Initialize VMs

Run vagrant up

• This will set up 3 VMs - manager, worker-1 & worker-2 all provisioned with docker
• manager node also has Flask app image joebibe/vagrant-docker-swarm⁠ installed on it.
  Which can be run as docker run -p 8000:8000 --name dsv joebibe/vagrant-docker-swarm

To ssh into the manager node run vagrant ssh manager
To ssh into worker node run vagrant ssh worker-1 or vagrant ssh worker-2

⁠Initialize Swarm manager

• ssh into manager node
• run docker swarm init --advertise-addr <manager-IP-address>:2377
  IP address is of the Manager Virtual Machine. Can be found / setup in Vagrantfile⁠
  Port 2377 is default for swarm
  This will generate a command with a token value docker swarm join --token <token> <manager-IP-address>:2377
• ssh into the worker nodes and execute the command generated above
• run docker node ls on the manager node to see the list of nodes and their roles(worker/manager)
• run docker network create -d overlay <network-name>
  creates an overlay network with the name specified which each of the container can join
  random-network will be the overlay network that our application containers live on
  -d tells which driver to use. Currently overlay driver is being used
• run docker service create --name <service-name> --network <network-name> <image-name>
  Creates the service with the given service name on the network specified. This will basically run containers from the image name specified on the nodes connected to the swarm.

⁠Scale up and down

• Scale up : docker service update --replicas 7 <service-name>
  Replicates the container 7 times and balances into all available nodes based on a round-robin or any such algorithms. This will launch arbitrary number of containers in each node (including manager)
  Check which container is running on which node by running docker service ps <service-name>
  Check how many containers are running on the current node by running docker ps or docker container ls

• Scale down : docker service update --replicas 4 <service-name> This command will destroy some of the containers on some of the nodes. Check which containers got destroyed by using the commands specified above.

⁠Manage worker nodes

• Drain : Bring a node down for management purpose
  docker node update --availability drain <worker-node-id>
  can be found by running docker node ls
  This will destroy the containers on the specified worker node and will relaunch them on available nodes

• Active : Bring up a node drained for management purpose
  docker node update --availability active <worker-node-id>
  This will activate the drained node and will deploy further scaled up containers on this as well

⁠Helper commands

• docker service ls
  Lists the services created
• docker network ls
  Lists the networks created and available by default
• docker node ls
  Lists all the nodes in swarm
• docker service ps <service-name>
  Lists all the containers running in all swarm nodes.