AWS Deployer

Should be considered very beta at the moment.

Provides a method to easily deploy/update code from a git repo to all servers within an AWS Autoscaling group. Similar to how AWS CodeDeploy works without the requirement of the CodeDeploy agent on the actual server.

Setup

In order to have things working properly a few different setup items need to be handled. The primary reason for this is to ensure the AWS resources are properly setup for secure access.

• A directory at ~/.aws. If you have the aws-cli tools installed this directory should already be created.
• In the ~/.aws directory, a profile file called "credentials" should be created.
• Add a new profile item with your access information.

[my-profile]
aws_access_key_id = key
aws_secret_access_key = secret

• Using the profile name from the credentials file, copy the pem to a file in the directory called my-profile.pem.

Sample Usage

var Deploy = require('aws-deployer')(config);

var opts = {
    "parallel_groups": "availability_zone",
    "auto_scaling_groups": "My-ASG",
    "command" : "sudo su; pwd;"
};

Deploy.run(opts, function (err, groups) {
    console.log(arguments);
});

Config Options

{
	aws : 'The standard configuration options for the aws-sdk.  This is not required',
	pem : 'String value of the AWS pem that is required to connect to the servers. This is not required',
	aws_profile : 'name of the profile to use.'
}

NOTE: The AWS IAM User will need to have EC2ReadOnly access.

##Public Functions

run(opts)

Runs a deployment.

Parameters:

• opts (object) - the actual payload of the run.
• opts.parallel_groups (string) - the method the deployment should be grouped.
• opts.auto_scaling_groups (string) - the name of the autoscaling group that the deployment should process on.
• opts.command (string) - the actual commands that should be run for the deployment.

##Events Emitted

Deployer will emit some helpful events:

ready

When the instance has been fully setup and ready for use.

Deploy.on("ready", function(config){ });

start-deploy

When the instance has been fully setup and ready for use.

data

• deployment_id (string): A uuid value that is assigned to this deployment.
• params (object): The parameters that were passed into run().
• groups (array): The actual groups that will be processed in parallel.
• grouping (string): How the groups were segmented. This is reflected based on what is passed into run().

Deploy.on("start-deploy", function(data){ });

start-deploy-to-group

When the deployment has been started for a specific group.

data

• deployment_id (string): A uuid value that is assigned to this deployment.
• group_key (string): The key of the group that is being started.

Deploy.on("start-deploy-to-group", function(data){ });

complete-deploy-to-group

When the deployment has been completed for a specific group.

data

• deployment_id (string): A uuid value that was assigned to this deployment.
• group_key (string): The key of the group that completed..

Deploy.on("complete-deploy-to-group", function(data){ });

end-deploy

When the deployment process has fully completed.

data

• deployment_id (string): A uuid value for the deployment that just completed.

Deploy.on("end-deploy", function(data){ });

start-deploy-to-instance

When the instance has been fully setup and ready for use.

data

• deployment_id (string): A uuid value for the deployment.
• instance (object): The details of the instance that is being deployed to.

Deploy.on("start-deploy-to-instance", function(data){ });

command-complete

When the instance has been fully setup and ready for use.

data

• deployment_id (string): A uuid value for the deployment.
• instance (object): The details of the instance that just completed.
• command (string): The command that was just run.
• response (string): The response that was returned based on the command that was run.

Deploy.on("command-complete", function(data){ });

complete-deploy-to-instance

When the deployment has been completed on a specific instance.

data

• deployment_id (string): A uuid value for the deployment.
• instance (object): The details of the instance that just completed.

Deploy.on("complete-deploy-to-instance", function(data){ });

error

If an error is thrown.

Deploy.on("error", function(error){ });

Sample Usage

var config = {
    aws_profile : 'my-profile'
    };

var Deploy = require('aws-deployer')(config);

var opts = {
    "parallel_groups": "availability_zone",
    "auto_scaling_groups": "My-AutoScaling-Group-Name",
    "command" : "sudo su;cd /data/application;git reset --hard;git pull;npm install;pm2 reload all;pm2 jlist;"
};

Deploy.run(opts, function (err, ASGroups) {
    console.log(arguments);
});

Deploy.on("ready", function(data){
    console.dir("..-->ready---->", data)
});

Deploy.on("start-deploy", function(data){
    console.log(".....Start Deploy---->", data)
});

Deploy.on("start-deploy-to-group", function(data){
    console.log("..........start-deploy-to-group---->", data)
});

Deploy.on("complete-deploy-to-group", function(data){
    console.log("...............complete-deploy-to-group---->", data)
});

Deploy.on("end-deploy", function(data){
    console.log("end-deploy---->", data)
});

Deploy.on("start-deploy-to-instance", function(data){
    console.log("..............start-deploy-to-instance---->", data)
});

Deploy.on("command-complete", function(data){

    if(data.command === 'pm2 jlist'){
        var response_parts = data.response.split('\n');
        var json = JSON.parse(response_parts[1]);

        var revision = json[0].pm2_env.versioning.revision

        console.log('..............command-complete---->",', revision);

    }

    //console.log("............command-complete---->", data)
});

Deploy.on("complete-deploy-to-instance", function(data){
    console.log("-complete-deploy-to-instance---->", data)
});

Deploy.on("error", function(err){
    console.log("-ERROR---->", err)
});