whyrg / compose-quickstart

asd-docker-cluster

For OSX

This collection of scripts and configs would quickly deploy a docker cluster on your OSX machine. The docker host cluster will be managed by docker-machine and created using docker-swarm. The container cluster and stack will be managed by docker-compose.

##Requirements Install docker-toolbox: https://docs.docker.com/engine/installation/mac/

##Usage Create your docker swarm environment with one of the following scripts on OSX:

• create-swarm.sh: Vmware swarm using consul
• create-vbox-swarm.sh: VirtualBox swarm using consul
• create-vbox-swarm-etcd.sh: VirtualBox swarm using etcd

The above scripts will create a docker cluster using docker-machine and docker-swarm. Swarm will be managed via either consul or etcd depending on which one is used. An overlay network named prod will be created as well. Docker client settings are also updated to point toward the SWARM endpoint.

In case you accidentally close the terminal, you can get the docker client settings back via: eval $(docker-machine env --swarm swarm-0)

##Start interlock instance Obsolete. See next section. Manual Method Run directly from dockerhub image:

docker run --name interlock --net prod -e AEROSPIKE_NETWORK_NAME=prod --rm  -v /var/lib/boot2docker:/etc/docker  aerospike/interlock --swarm-url=$DOCKER_HOST --swarm-tls-ca-cert=/etc/docker/ca.pem --swarm-tls-cert=/etc/docker/server.pem --swarm-tls-key=/etc/docker/server-key.pem --debug -p aerospike start

Old way Build and run the interlock container

eval $(docker-machine env --swarm swarm-0)
git clone [e-mail]:aerospike/interlock.git
cd interlock
make build-container && docker build -t interlock .
docker run --name interlock --net prod -e AEROSPIKE_NETWORK_NAME=prod --rm  -v /var/lib/boot2docker:/etc/docker  interlock --swarm-url=$DOCKER_HOST --swarm-tls-ca-cert=/etc/docker/ca.pem --swarm-tls-cert=/etc/docker/server.pem --swarm-tls-key=/etc/docker/server-key.pem --debug -p aerospike start

##Let Compose set everything up

eval $(docker-machine env --swarm swarm-0)
docker-compose --x-networking --x-network-driver overlay -p prod up -d
docker-compose --x-networking --x-network-driver overlay -p prod scale aerospike=SCALE

##Confirm cluster:

docker run --rm --net prod aerospike/aerospike-tools asadm -e i -h prod_aerospike_1

Data Persistance

Host Data Persistence Data can be persisted to the host that's running the Docker daemon.

aerospike:
  image: aerospike/aerospike-server
  volumes:
    - /host/data/dir:/opt/aerospike/data
   ...

In the above example, each docker host would have the directory /host/data/dir created if not already exists. Aerospike will then write to this volume which is mounted under /opt/aerospike/data.

Data-only Container A data-only container is another option to persist data. We can define a data-only container that exposes volumes to be used by Aerospike.

data:
  image:aerospike/aerospike-server
  entrypoint: tail -f /dev/null
  volumes:
    - /opt/aerospike/data
aerospike:
  image:aerospike/aerospike-server
  volumes_from:
    - data
  ....

Note that the volumes_from parameter replaces the corresponding volume parameter from the aerospike service definition. The aerospike-server image is chosen for the data-only container to minimize disk usage and the entrypoint overridden to trick docker-compose into keeping the data-only container. Scaling should be done with the data-only container first.

A data-only container would give you flexibility that host-based persistence would not, as there is no longer a requirement for identical host hardware configurations.