zmarcantel/docker-cassandra

Datastax Community image running on top of a Ubuntu-based Docker image

docker-cassandra

Standlone/Clustered Datastax Community running on top of a Ubuntu-based Docker image

Running as standalone

Docker

Just start the image normally

docker run -d zmarcantel/cassandra

The necessary ports are exposed, so you can interact with the docker image as you normally do.

If you need help, see Usage Guide.

Vagrant

Simply use the builtin docker provisioner to pull and start the image.

config.vm.provision "docker" do |d|
    d.pull_images "zmarcantel/cassandra"
    d.run "cass", image: "zmarcantel/cassandra"
end

Running as a cluster

To get the images to talk to each other, simply use the docker link and name options.

We can link up to 10 containers.

You are able to create as many containers as you'd like, but the init script only considers up to the first 10 as seeds.

Truly, you only need one other seed and for simplicity, that's how this example works. The Cassandra cluster will gossip its way into being fully networked rather quickly. There is no reason a machine cass9999 can't just reuse a seed from cass[0-9].

Docker

Start a first image, and then link them all up.

Start the first image

docker run -d -name cass0 zmarcantel/cassandra

Link the cluster

docker run -d -name cass1 -link cass0:cass0 zmarcantel/cassandra
docker run -d -name cass2 -link cass0:cass0 -link cass1:cass1 zmarcantel/cassandra

Vagrant

We'll do the same as above, but using the docker provisioner.

config.vm.provision "docker" do |d|
    d.run "seed", auto_assign_name: false,
      args: "--name cass0",
      image: "zmarcantel/cassandra"

    d.run "first", auto_assign_name: false,
      args: "--name cass1 --link cass0:cass0",
      image: "zmarcantel/cassandra"

    d.run "second", auto_assign_name: false,
      args: "--name cass2 --link cass0:cass0 --link cass1:cass1",
      image: "zmarcantel/cassandra"
end

Usage Guide

Basic Docker One-Liners

Get image ID for docker-cassandra on your machine

docker images | grep "zmarcantel/cassandra" | sed -e "s/\s\+/ /g" | cut -d' ' -f3

This is a very convenient export: the image ID

export CASSDOCK_ID=`docker images | grep "zmarcantel/cassandra" | sed -e "s/\s\+/ /g" | cut -d' ' -f3`

List the IPs of containers running docker-cassandra

docker inspect $CASSDOCK_ID | grep IPAddress | sed 's/"IPAddress": "/ /g' | sed 's/",//g' | sed 's/ //g'

Get the above in a comma-separated format

docker inspect $CASSDOCK_ID | grep IPAddress | sed 's/"IPAddress": "/ /g' | sed 's/",//g' | sed 's/ //g' \
sed -e :a -e N -e 's/\n/,/' -e ta

Or with spaces if you prefer

docker inspect $CASSDOCK_ID | grep IPAddress | sed 's/"IPAddress": "/ /g' | sed 's/",//g' | sed 's/ //g' \
sed -e :a -e N -e 's/\n/ /' -e ta

Talking to Cassandra

Every docker image is given an IP to communicate with.

Use the above commands to get the IP of your docker images. Or manually look them up using docker ps and docker inspect.

The Following IPs are open (default Cassandra ports):

7199 7000 7001 9160 9042

Standalone instance

Every language and client has a different syntax, but simply:

    cassandra.connect("172.17.0.2")

Obviously modify to fit the method of connecting for your library.

Cluster

The IPs of the cluster are determined using [one-liners] and your application input channels.

I provide a way to get the info, but how it gets into your application is up to you.

The below IPs are examples to give a gist, but you must get your relevant IPs using the above commands.

    cassandra.connect(["172.17.0.2", "172.17.0.3", "172.17.0.4"])