Example of simplified DevOps using Terraform

This repo is meant to be an example of how to easily set up an auto-scaling cluster of CoreOS machines running Couchbase and Sync Gateway in Docker containers behind Nginx on AWS using Terraform.

You may want to check out the README at couchbase-cluster-go for more background info.

Usage

Do a search across all the files in the repo for "XXX" in order to place proper values in them. You'll want to grab a new etcd discovery URL over at https://discovery.etcd.io/new and place it into cloud-config.yaml ...the other replacements should be pretty self-explanatory however. You'd likely also want to use an updated HVM AMI to avoid immediate self-updates if you just want to quickly kick the tires.

To spin up your cluster, make sure you cd into this project's directory, then run terraform apply

Check on the status of your machines over at https://console.aws.amazon.com/ec2/v2/home, and once they're up log in (making sure to put in the proper path to your .pem file, and a proper IP for a machine in your newly spun up cluster):

ssh -o StrictHostKeyChecking=no -i /XXX_PATH_TO_AWS_PEM_FILE_XXX/aws.pem -A core@XXX_AWS_IP_XXX

Once you've SSHed in, run the following (making sure to use your own Sync Gateway config file):

etcdctl set /couchbase.com/enable-code-refresh true && \
sudo docker run --net=host tleyden5iwx/couchbase-cluster-go update-wrapper couchbase-fleet launch-cbs \
  --version latest \
  --num-nodes 3 \
  --userpass "user:passw0rd" \
&& \
sudo docker run --net=host tleyden5iwx/couchbase-cluster-go update-wrapper sync-gw-cluster launch-sgw \
  --launch-nginx \
  --num-nodes=3 \
  --config-url=http://XXX_YOUR_SYNC_GATEWAY_CONFIG_FILE/config.json \
  --create-bucket todos \
  --create-bucket-size 200 \
  --create-bucket-replicas 1

Here's a helpful command to sanity check the state of things in your cluster:

docker ps -a && echo "" && fleetctl list-unit-files && echo "" && fleetctl list-units && echo "" && etcdctl ls / --recursive

Avoid unexpected AWS bills

IMPORTANT: to completely destroy the cluster, run terraform destroy -force, otherwise you might get an unexpectedly more expensive AWS bill.