Pool Controller

Install infrakit

$ curl -sSL https://docker.github.io/infrakit/install | sh
Building for mac darwin / amd64
Building infrakit GOOS=darwin GOARCH=amd64, version=f9310606.m, revision=f9310606c5e1ae6afa2d7d46ffbc110351da1e67, buildtags=builtin providers

2. Add this playbook

infrakit playbook add pool https://raw.githubusercontent.com/docker/infrakit/master/docs/controller/pool/playbook.yml

2. Verify playbook added

infrakit use pool

3. Start up server

infrakit use pool start --accept-defaults

4. View objects in the pool controller. In another window

watch -d infrakit local mystack/pool describe -o

5. In another shell, commit the config yml:

infrakit use pool workers.yml | infrakit local mystack/pool commit -y -

You will be prompted to answer the size of the collection and parallelism:

$ infrakit use pool workers.yml | infrakit local mystack/pool commit -y -
How many? [5]: 10
How many at a time? [1]: 3

In the window where you are watching the objects, you should see resources being provisioned.

6. Scaling up / down Simply commit the same config and change the size / parallelism:

$ infrakit use pool workers.yml | infrakit local mystack/pool commit -y -
How many? [5]: 4
How many at a time? [1]: 2

The undesired instances will be marked UNMATCHED and then eventually terminated, using the batch size you specified.

You can verify the instances:

$ infrakit local az1/compute describe