Docker Swarm Example Blueprint

The blueprints in this project provide orchestration for starting, healing, and scaling a Docker Swarm cluster on Openstack. There are 3 blueprints, with slightly different use cases:

• swarm-local-blueprint.yaml : a cfy local blueprint that orchestrates setup and teardown of the cluster without a manager
• swarm-openstack-blueprint.yaml : an Openstack blueprint that orchestrates setup and teardown of the cluster with a manager
• swarm-scale-blueprint.yaml : an Openstack bluieprint that orchestrates setup, teardown, autohealing, and autoscaling of the cluster

Prerequisites

These blueprints have only been tested against an Ubuntu 14.04 image with 2GB of RAM. The image used must be pre-installed with Docker 1.12. Any image used should have passwordless ssh, and passwordless sudo with requiretty false or commented out in sudoers. Also required is an Openstack cloud environment. The blueprints were tested on Openstack Kilo.

Cloudify Version

These blueprints were tested on Cloudify 3.4.0.

Operation

swarm-local-blueprint.yaml instructions

• Clone the repo to a system that access to the target Openstack cloud
• Edit the inputs/local.yaml file as follows:
• image : the image id on Openstack of the Ubuntu 14.04 image
• flavor : the Openstack flavor id
• ssh_user : the ssh user to log into the instances (probably ubuntu)
• ssh_keyfile : the path to the key used to ssh to the instance
• ssh_keyname : the name of the key in openstack
• run cfy local install-plugins -p swarm-local-blueprint.yaml
• run cfy local execute -w install --task-retries 10

This will create the Swarm cluster. The manager node is assigned a public ip. You can see it by running cfy local outputs.

To tear down the cluster, run cfy local execute -w uninstall --task-retries 10.

swarm-openstack-blueprint.yaml instructions

• Start a Cloudify 3.4.0 manager.
• Edit the inputs/openstack.yaml file to add image, flavor, and user name (probably ubuntu).
• run cfy blueprints upload -b swarm -p swarm-openstack-blueprint.yaml
• run cfy deployments create -b swarm -d swarm -i input/openstack.yaml
• run cfy executions start -d swarm -w install

This will create the Swarm cluster. The manager node is assigned a public ip. You can see it by running cfy deployments outputs -d swarm.

To tear down the cluster, run cfy executions start -d swarm -w uninstall

swarm-scale-blueprint.yaml instructions

• Start a Cloudify 3.4.0 manager.
• Edit the inputs/openstack.yaml file to add image, flavor, and user name (probably ubuntu).
• run cfy blueprints upload -b swarm -p swarm-openstack-blueprint.yaml
• run cfy deployments create -b swarm -d swarm -i input/openstack.yaml
• run cfy executions start -d swarm -w install

This will create the Swarm cluster. The manager node is assigned a public ip. You can see it by running cfy deployments outputs -d swarm.

To see autohealing in action, go to the Openstack Horizon dashboard and terminate the worker. Then go to the Cloudify UI deployments tab. See the heal workflow begin and restore the missing node.

To see autoscaling in action:

• ssh to the Cloudify manager: cfy ssh
• ssh to the swarm manager: sudo ssh -i /root/.ssh/agent_key.pem ubuntu@<manager-ip>
• run the built in load generator container: sudo docker service create --constraint 'node.role == worker' --restart-condition none stress /start.sh
• Then go to the Cloudify UI deployments tab. See the scale workflow begin and grow the cluster.

In a few minutes, the cluster will scale down to it's original size (one worker) due to the scale down policy in the blueprint.

To tear down the cluster, run cfy executions start -d swarm -w uninstall