This cookbook is used to bootstrap core services on a machine during the cloudinit initialization. By the core services here we specifically mean consul and swarm.
| Attribute | Description | Default value |
|---|---|---|
| machine.bootstrap_containers | The list of containers to bootstrap. Note that bootstrap will happen in the specified order. | %w[consul swarm] |
| machine.containers | Specifies containers configuration. The key is the container name (will be used as image if image is not provided), the value contains different container options such as image, volumes, exposed_ports etc. | see: attributes/containers.rb |
| machine.node_class | Specifies consul agent node_mata.class variable used to distinguish nodes. | "node" |
| machine.node_tags | A list of tags added to the dummy node_meta service. | [] |
At the moment cookbook generates consul agent node_meta containing class and instance_id which are set to machine.node_class and AWS instance_id. This metadata can be used to distinguish nodes when consul agent is running. There's another way to distinguish nodes while consul agent is stopped is to query the service catalog (when leave_on_terminate is set false). Specifically we use a dummy service named node_meta. This service contains the following tags: class:{{node_class}} and instance_id:{{AWS instance_id}} together with the list of tags from the machine.node_tags attribute.
| Attribute | Description | Default value |
|---|---|---|
| data_dir | Docker data directory. | "/var/lib/docker" |
| port | Specifies the port docker listens on. | 2375 |
| labels | Specifies the list of labels of the docker engine. | [] |
| listens | Specifies a list of endpoints the docker engine actually listens on (tcp://#{bind_address}:#{port} is automatically added). |
['unix:///var/run/docker.sock'] |
| bind_address | Specifies the bind address of the docker engine. | "0.0.0.0" |
These attributes define consul agent.json configuration. See attributes/options.rb.
Set the value to true to passthrough docker binary and socket as volumes into a container.
Author:: ActionML ([email protected])