s3uploader is a node module to upload files to S3 with support for automatic retry and exponential backoff.

It is based off of intimidate which had several downfalls. The biggest two being the inability to specify headers when uploading a file (meaning that all files are uploaded using the default private ACL), and an improper mime-type method lookup that would fail when uploading a temp file (like from a file upload form).

It uses the excellent knox library to handle the heavy lifting.

When you need those uploads to back off, use intimidate™. - The Readme

npm install s3uploader

Upload a local file:

var Intimidate = require('intimidate')
// Create a client
var client = new Intimidate({
  key: 'some-aws-key',
  secret: 'some-aws-secret',
  bucket: 'lobsters',
  maxRetries: 5
})

client.upload('path/to/a/file.xml', 'destination/path/on/s3.xml', {'x-amz-acl': 'public-read'}, function(err, res) {
  if (err) {
    console.log('oh noes, all uploads failed! last error was', err)
  }
  else {
    console.log('yahoo, upload succeeded!')
  }
})

The constructor takes any opts that can be passed to knox's createClient function. Here are some important ones.

• key - S3 api key. Required.
• secret - S3 api secret. Required.
• bucket - S3 bucket to upload to. Required.
• region - S3 region to upload to. Defaults to 'us-west-2'
• maxRetries - the number of times to retry before failing. Defaults to 3.
• backoffInterval a multiplier used to calculate exponential backoff. Larger numbers result in much larger backoff times after each failure. Defaults to 51.
• defaultHeaders an object with some default headers so you don't have to specify them for every call

Example:

var Intimidate = require('intimidate')
var s3Uploader = new Intimidate({
  key: 'love',
  secret: 'a sneaky secret',
  bucket: 'kicked',
  maxRetries: 4,
  backoffInterval: 20
})

Upload a file at sourcePath with automatic retries and exponential backoff

Params:

• sourcePath location of the file to upload on the fs
• destination path in s3 to upload file to
• headers (optional) object containing headers to send to S3
• cb function(err, res) called when upload is done or has failed too many times. err is the last error, and res is the reponse object if the request succeeded

Example:

client.upload('a_car.zip', 'uploaded_cars/car.zip', {'x-amz-acl': 'public-read'}, function(err, res) {
  if (err) {
    console.log('Dang, guess you can\'t upload a car.', err)
  }
  else {
    console.log('I uploaded a car.')
  }
})

Upload a buffer

Params:

• buffer buffer to upload
• headers HTTP headers to set on request. 'Content-Length' will default to buffer.length, and 'Content-Type' will default to 'application/octet-stream' if not provided.
• destination path on S3 to put file
• cb function(err, res) called when request completes or fails too many times

Example:

var data = new Buffer('Shall I compare thee to a summer/'s day?')
var headers = {
  'Content-Type': 'application/text',
  'Content-Length': data.length
}

client.uploadBuffer(data, headers, 'poem_idea.txt', function(err, res) {
  if (err) {
    console.log('error uplaoding my sweet poem idea', err)
  }
  else {
    console.log('my poem idea is successfully archived to s3')
  }
})

Upload an array of files. The callback will be called when they all upload successfully, or when at least one of the uploads has failed.

Params:

• files Array of {src: 'some_path.file', dest: 'some_uploaded_path.file'} file object to be uploaded.
• headers (optional) object containing headers to send to S3
• cb function(err, res) that will be called when upload is complete or one of the files has failed to upload.

Example:

var files = [{src: 'hurp.txt', dest: 'durp.txt'}, {src: 'foo.txt', dest: 'foo.txt'}]
client.uploadFiles(files, {'x-amz-acl': 'public-read'}, function(err, res) {
  if (err) {
    console.error('error uploading one file', err)
  }
  else {
    console.log('hooray, successfully uploaded all files')
  }
})