prokvk/node-stack

node-stack

this stack provides all necessary tools for developing Node APIs. It is built on Docker and several Node modules, including main module knapp which is designed to cover all basic functionality of API.

This documentation will refer to coffee files only. There is a tool prepared for compiling coffee to JavaScript (described below) and from the server point of view it doesn't matter whether CoffeeScript or JavaScript is used.

Prerequisites

• GIT
• Docker
• ansible

Installation

Create a directory for node-stack and cd into it. Then:

1. run git clone https://github.com/prokvk/node-stack.git .
2. configure .install_conf.yml
3. run ./install

This command will install node-stack in current directory.

Install script will pull 2 repositories with docker images and build them. These repositories are:

• ns-docker-nginx
• ns-node-server

After you've ran installation script you need to do the following:

1. add symlink to $NODESTACK_PATH/nodestack to one of bin folders in your $PATH
2. "source" $NODESTACK_PATH/.env in your .bashrc / .bash_profile
3. "source" $NODESTACK_PATH/ns_bash_completion in your .bashrc / .bash_profile

Points 1 and 3 are optional. Obviously without adding symlink you will have to use full path in your commands. Examples in this README will assume that nodestack command is available. Also as name suggests, ns_bash_completion provides bash completion which won't be available without point 3.

Test your installation

Run command install-test

All you should see is Install test successful

How does it work

• NGINX container is used as a proxy for both API and DOC.
• Node server container is used for running server with API, installing modules, compiling coffee, running tests...
• knapp module has prepared tools for performing automated tests and generating static documentation.

Every time you run the server (with nodestack runs) NGINX config files are generated and NGINX is started/restarted before node server container starts.

nsconf.yml and local.nsconf.yml config files are included in every project. By default, nsconf.yml has its local_conf var set to local.nsconf.yml. That means it will use this file to run locally. local.nsconf.yml is also in .gitignore. These config files hold all configuration for given project (not API itself, which has its own configs).

The whole stack is controlled by nodestack binary. ns_bash_completion provides bash completion for easy use. Available commands are:

• nodestack create-project - copy (and setup) files from new project template
• nodestack init-project - install modules based on nsconf.yml config file
• nodestack runs - run server
• nodestack stops - stop server
• nodestack gendoc - generate documentation
• nodestack atest - run automated tests
• nodestack install-test - validate nodestack installation
• nodestack npmi - npm install
• nodestack gcompile - grunt task for compiling CoffeeScript files into JS
• nodestack printsshkey - print nodestack public SSH key
• nodestack deploy - deploy project to remote
• nodestack pull - pull & setup an existing nodestack project
• nodestack s3s_cu - s3 secrets - create AWS IAM user
• nodestack s3s_cuak - s3 secrets - create AWS IAM user access keys (AWS security credentials will be saved in NODESTACK_PATH/ansible dir)
• nodestack s3s_cb - s3 secrets - create private encrypted S3 bucket
• nodestack s3s_ugba - s3 secrets - grant bucket access to specified user
• nodestack s3s_full - s3 secrets - whole s3 secrets flow (cu, cuak, cb, ugba)

IMPORTANT: All commands need to be run from project root, that is $NODESTACK_PROJECTS_PATH/project-folder/, which contains nsconf.yml and local.nsconf.yml files created with nodestack create-project command. Exceptions are:

• nodestack stops - can be run from anywhere
• nodestack install-test - can be run from anywhere
• nodestack s3s_ commands - can be run from anywhere

Init new project

1. Create new folder for your project inside $NODESTACK_PROJECTS_PATH and cd into it
2. run nodestack create-project
3. open and setup nsconf.yml and local.nsconf.yml config files (all options are documented in these files)
4. run nodestack init-project
5. in app/config folder, setup config.cson and .env files
6. in app/index.coffee edit the object passed to the init function (this is documented in knapp)

Now your project is all setup. You only need to work with routes.coffee and whatever libs you create should go into app/lib folder.

Recommended project filesystem structure

|——app
   |——config
      |——.env
      |——config.cson
   |——lib
      |——lib.js
   |——log
   |——node_modules
   |——index.coffee
   |——package.json
   |——routes.coffee
|——doc
|——.gitignore
|——local.nsconf.yml
|——nsconf.yml
|——supervisor_conf

.gitignore is setup to exclude app/config/.env (among other things)

nsconf.yml and local.nsconf.yml config files need to be in the project root

app contains the API

Develop your API

knapp documentation explains creating routes, setting their meta (inSchema, outSchema and testRequest) and working with requests.

Run API

Run command nodestack runs. This will setup and start/restart NGINX and start node server container. It will run on foregroung and it will output URLs for API and DOC. Sample:

NS running, links:
DOC: http://localhost:8080/ns/template-project/doc
API: http://localhost:8080/ns/template-project

Stop API

Good old CTRL+C will do :)

Run automated tests

Routes by default don't get included in automated tests. They get included when testRequest is defined in route meta object. testRequest contains data that will be sent to given endpoint in automated tests. It should honor the inputSchema, if it doesn't a "schema integrity error" will be thrown.

To run automated tests use command nodestack atest

Generate documententation

Run nodestack gendoc

Adding Node modules

If you just want to run npm install then use nodestack npmi

If you want to install specific module (npm install --save MODULE_NAME) then use nodestack npmi MODULE_NAME

Use JavaScript

To compile coffee files in app folder use nodestack gcompile

Deploy to remote

Prerequisites

• SSH access to remote
• gzip (local machine)
• gunzip (remote)
• supervisor (remote)
• docker permissions for your SSH user (remote)
• supervisor permissions for your SSH user (remote)

gzip and gunzip are used to pack / unpack docker image which is uploaded to remote during deploy (if necessary).

Prepare

In order for deploy to remote server, you need to setup a block under remotes block in projects nsconf.yml file. A sample dev block is included in template nsconf.yml file for every project. All the values are described there. You can have as many remote blocks as you wish.

You should go through deploy setup with NS maintainer, who will provide you with all information.

Deploy

After you have everything set up, you can deploy to this remote with nodestack deploy REMOTENAME.

Use s3 secrets commands

nodestack s3s_ commands use ansible role. You can use them to easily manage AWS if you choose to store your secrets in S3.

All s3s commands require that you either specify aws_credentials_path with your command or use --default-aws-credentials|-dac. Examples:

aws_credentials_path=/path/to/your/credentials/file uname=YOUR_USER_NAME nodestack s3s_cu

or

uname=YOUR_USER_NAME nodestack s3s_cu --default-aws-credentials

or

uname=YOUR_USER_NAME nodestack s3s_cu -dac

Notice uname parameter used in those samples. All s3s commands require some parameters, complete list is:

• s3s_cu - uname
• s3s_cuak - uname
• s3s_cb - bname, region
• s3s_ugba - uname, bname
• s3s_full - uname, bname, region

If you don't specify these required variables the command will fail with message about missing required variable.

Pull existing nodestack project

Run nodestack pull REPOSITORY_URL. This will:

• pull a repository
• npm install dependencies
• create necessary folders
• copy template Gruntfile

Update node-stack

To update NS files from repository run $NODESTACK_PATH/update. This will NOT update your .env file, so if there are changes you need to apply them manually.