Socker

A friendly chat with your Docker daemons.

What is this?

Socker is an experiment. I want to make interactions with Docker more accessible to a less technical crowd.

In this case I’m building a web-chat like interface for commanding daemons.

How does it work?

Socker has 2 components - the server and the agent.

You need only one instance of the server; The server mediates between the users and the agents and all communications are over WebSockets.

The agent can be installed anywhere but should have network access to both the Docker daemon and the server.

One serve can handle any number of agents.

First phase of implementation

For the POC phase all client input is relayed as-is to the agent.

Goals of future phases

Create a DSL over common daemon commands.
Stream daemon responses.

Using Socker

Building

In your GOPATH build the server and the agent binaries using:

go get github.com/noamt/socker/server

and

go get github.com/noamt/socker/server

Running the server

You can run the server using the following command:

./server

The server will by default bind to localhost:8080. You can modify this by using the address switch:

./server -address foo-bar:1337

Running the agent

You can run the agent using the following command:

./agent

The agent will by default identify itself as local and will try to communciate with a server at localhost:8080. You can modify these using the following switches:

./agent -id node1 -server foo-bar:1337

Accessing the Web UI

Simply browse to localhost:8080 or the address specified to the server.

Supported commands

Socker supports two types of commands - collective and specific.

Collective commands are directed at all agents, prefixed with @all and are followed by the directives. For example, to have all agents report in use:

@all report

Specific commands are directed at one agent, is prefixed with @AGENT_NAME and is followed by the directives. For example, to view what containers are running on a specific agent use:

@local ps -a