build vm

vagrant up
curl http://127.0.0.1:8098/ping     # should return 'OK' if everything worked

connect to vm and run scripts

vagrant ssh

# load data
python /vagrant/src/main.py load

# run queries
python /vagrant/src/main.py query

Create three nodes and make sure nodename and ip address is configured in riak.conf. The First node can be started as normal (riak start). For additional nodes you have to join them to any existing node.

Assume first nodes nodename is riak@198.192.96.216

1. Connect to each additional node and run: riak-admin cluster join riak@198.192.96.216
2. Check cluster plan: riak-admin cluster plan
3. Commit changes to build cluster: riak-admin cluster commit

Create a main project directory

mkdir ~/ccd-riak
cd ~/ccd-riak

Clone this repo for each node (you must name the folder using the format nodeX)

git clone https://github.com/meyersj/ccd-riak.git node1
git clone https://github.com/meyersj/ccd-riak.git node2
git clone https://github.com/meyersj/ccd-riak.git node3

Clone again for the manager

git clone https://github.com/meyersj/ccd-riak.git manager

Edit manager/bin/manager.sh

• export DO_API_TOKEN= must be set to valid digital ocean API token
• export DO_SSH_KEY= must be set to private ssh key that will be used for accessing the nodes

Build nodes

• Execute: manager/bin/manager.sh build
• This will run a vagrant up --provider=digital_ocean in each of repositories created above
• In each nodeX you will find a build.out file.
• When FINISHED is printed at the end of build.out that node has been built succesfully
• For each node you want to get the IP address
• In manager/bin/manager.sh set MAIN_NODE variable to the IP address of any node and set ALL_NODES to each additional nodes IP address

After setting MAIN_NODE and ALL_NODES create the cluster

• Execute: manager/bin/manager.sh cluster
• This will ssh onto each node in ALL_NODES and join it to MAIN_NODE
• Then changes are committed to create a cluster

Finally load the data

• Execute: manager/bin/manager.sh load
• The loopdata csv has been split into three files and loaded into S3
• Each node will download its section of data and begin looping over it inserting the records using put requests