A customizable tool for deploying your Ember app to Amazon's S3. Customize the deploy by running your own scripts within the process (beforeBuild, afterBuild, beforeDeploy, afterDeploy)
npm install ember-cli-s3-sync --save-dev
ember generate config-s3
This addon uses aws-sdk
for communicating with Amazon S3. You can provide authentication credentials in the following ways (listed in order of precedence):
ember deploy:s3 --aws-secret=my-secret --aws-key=my-cool-key
- shared credentials file at ~/.aws/credentials file.
- these shell environment variables:
AWS_ACCESS_KEY_ID
,AWS_SECRET_ACCESS_KEY
- deploy/config.js
{
...
options: {
accessKeyId: "mycoolkey",
secretAccessKey: "secretsarecool"
}
...
}
note if key & secret aren't found at any of the checks above then you will be prompted for credentials -blocking the deploy (keep this in mind if using with automated/continuous deployment systems).
ember deploy:s3 --environment=production --aws-key=12345 --aws-secret=asdfasdf --aws-bucket=buckets-o-fun
- this builds a production version of your app and deploys all files in the
/dist
directory to the S3 bucket "buckets-o-fun"
ember deploy:s3
- this will build development version of your app and prompt you for
awsKey
,awsSecret
, andawsBucket
possible cli arguments:
environment
(optional. uses app's default. Passed into deployl/config.js)output-path
(optional. uses app's default/dist
)aws-key
(required. will prompt if not found)aws-secret
(required. will prompt if not found)aws-bucket
(required. will prompt if not provided as cli arg or found in deploy/config.js)aws-region
(optional. will be verified and updated if necessary during deploy process)skip-build
(optional. will skip the build, deploying whatever is in/dist
)prepend-path
(optional. will upload assets to the 'subdirectory' path defined here)
notes: camelCase args are okay but they'll be converted to their dasherized version.
Generate a config file with ember generate config-s3
(creates a file at your-app/deploy/config.js)
The environment
is passed into config file, which returns an object containing deploy configuration options.
And here are the pieces to deploy/config.js:
{ // deploy/config.js
...
environment: 'development', // not used, but good practice to name the config incase you have several
promptCredFile: false, // prompts whether or not to use AWS cred file, if one is found.
verbose: false, // turns on AWS-SDK verbose flag. May be useful for troubleshooting.
...
}
{ // deploy/config.js
...
options: {
region: 'us-east-1',
maxRetries: 3,
sslEnabled: true,
params: {
Bucket: 'my-bucket' // yes that's a capital B
},
...
}
If you want the deploy process to prompt a user for additional options to be merged in for instantiating the S3 Object:
Uses the inquirer node module.
{ // deploy/config.js
...
additionalOptions: [
{
type: 'input',
name: 'maxRetries',
'default': 2,
message: 'Please enter a maximum number of retries to attempt when uploading a file',
validate: function(value) {
if ('number' !== typeof value) {
return false;
}
return value;
}
}
]
...
}
notes: when a name
(e.g., maxRetries
) is in both additionalOptions
and the options
hash,
then the value defined in the options
hash takes precedence and the user will not be prompted.
You can run scripts throughout the deploy process. These scripts must exit their process for the deploy to continue running.
{ // deploy/config.js
...
beforeBuild: [
{
command: 'curl http://my-site.nyc?new_build=start', // base command to run
includeOptions: ['someOption', 'anotherOption'], // options to include as cli-args for base command
fail: false // whether a non 0 exit code should halt the deploy process
}
],
afterBuild: [
...
],
beforeDeploy: [
...
],
afterDeploy: [
...
],
...
}
providing default cli-arguments to run with your custom scripts:
Running: ember deploy:s3 --compressed --head=false --header="Pragma: no-cache"
{ // deploy/config.js
...
beforeDeploy: [
{
command: 'curl http://httpbin.com/headers',
includeOptions: [
'compressed',
'beh',
{ header: 'X-Host: mysite.com' },
{ header: 'X-Update: 1' },
{ head: true }
],
fail: false
}
],
...
}
will run the following command, waiting for it to exit before deploying assets to S3 (beforeDeploy
hook):
curl http://httpbin.com/headers --compressed --header "X-Host: mysite.com" --header "X-Update: 1" --header "Pragma: no-cache"
explaination:
--compressed
was passed with
ember deploy:s3
and so it was included
--beh
was not passed with
ember deploy:s3
and so it was ignored
--header "X-Host: mysite.com"
and--header "X-Update: 1"
were defined as defaults so they were included
--header "Pragma: no-cache"
was passed with
ember deploy:s3
and included because there exists aheader
key inincludeOptions
Array. It did not overwrite any defaults since there were multiple defaults.
--head
was passed as
false
withember deploy:s3
and so it overwrote the default
Example config for deploying YUI docs with ember-cli-yuidoc:
Running: ember deploy:s3 --environment documentation --output-path docs --prepend-path docs --skip-build
var documentationConfig = { // deploy/config.js
...
params: {
Bucket: 'bucket-o-docs'
},
beforeDeploy: [
{
command: 'ember ember-cli-yuidoc',
includeOptions: [],
fail: true
}
],
...
};
module.exports = function(env) {
return (env === 'documentation') ? documentationConfig : 'development';
};
--environment
gets passed into deploy/config.js so that it can determine which config object to return--output-path
is the local folder (relative to project dir) where assets are located--prepend-path
is a way to specify a 'subdirectory', inside s3 bucket, to upload assets--skip-build
says to skip building the actual app (ember build
)
notes: beforeBuild
and afterBuild
are not run if you use --skip-build
flag.
- better test coverage
- write documentation for each function
- write documentation describing flow, configurable options, general how to use
- ability to save config file
- ability to generate
config-s3.js
for deploy configuration - ability to specify optional params in
config-s3.json
to be prompted for - ability to sync individual files to s3 bucket
- ability to do a dryrun
- ability to skip build task and just deploy a specified directory
- support gzipped files
- ability to set meta data (headers) for files, such as
Expires
- update s3 with file's ContentMD5