I like CarrierWave. That being said, I don't like tying up app instances waiting for images to process.

This gem addresses that by offloading processing or storage/processing to a background task. We currently support Delayed Job, Resque, Sidekiq, SuckerPunch, Girl Friday, Qu, and Queue Classic.

There are currently two offerings for backgrounding upload tasks which are as follows;

# This stores the original file with no processing/versioning.
# It will upload the original file to s3.
# This was developed to use where you do not have control over the cache location such as Heroku.

Backgrounder::ORM::Base::process_in_background

# This does nothing to the file after it is cached which makes it super fast.
# It requires a column in the database which stores the cache location set by carrierwave so the background job can access it.
# The drawback to using this method is the need for a central location to store the cached files.
# Heroku may deploy workers on separate servers from where your dyno cached the files.
#
# IMPORTANT: Only use this method if you have full control over your tmp storage directory.

Backgrounder::ORM::Base::store_in_background

These instructions assume you have previously set up CarrierWave and your queuing lib of choice.

In Rails, add the following your Gemfile:

gem 'carrierwave_backgrounder'

Run the generator which will create an initializer in config/initializers.

  rails g carrierwave_backgrounder:install

You can pass additional configuration options to Girl Friday and Sidekiq:

CarrierWave::Backgrounder.configure do |c|
  c.backend :girl_friday, queue: :awesome_queue, size: 3, store: GirlFriday::Store::Redis
end

In your CarrierWave uploader file:

class AvatarUploader < CarrierWave::Uploader::Base
  include ::CarrierWave::Backgrounder::Delay

  #etc...
end

In your model:

mount_uploader :avatar, AvatarUploader
process_in_background :avatar

Optionally you can add a column to the database which will be set to true when the background processing is started and to false when the background processing is complete.

add_column :users, :avatar_processing, :boolean, null: false, default: false

In your model:

mount_uploader :avatar, AvatarUploader
store_in_background :avatar

Add a column to the model you want to background which will store the temp file location:

add_column :users, :avatar_tmp, :string

If you need to process/store the upload immediately:

@user.process_<column>_upload = true

This must be set before you assign an upload:

# In a controller
@user = User.new
@user.process_avatar_upload = true
@user.attributes = params[:user]

To override the worker in cases where additional methods need to be called or you have app specific requirements, pass the worker class as the second argument:

process_in_background :avatar, MyParanoidWorker

Then create a worker that subclasses carrierwave_backgrounder's worker:

class MyParanoidWorker < ::CarrierWave::Workers::ProcessAsset
  # ...or subclass CarrierWave::Workers::StoreAsset if you're using store_in_background

  def error(job, exception)
    report_job_failure  # or whatever
  end

  # other hooks you might care about
end

Use overriden worker that inherits from ActiveJob::Base and includes relevant worker mixin:

class MyActiveJobWorker < ActiveJob::Base
  include ::CarrierWave::Workers::ProcessAssetMixin
  # ... or include ::CarrierWave::Workers::StoreAssetMixin

  after_perform do
    # your code here
  end

  # Sometimes job gets performed before the file is uploaded and ready.
  # You can define how to handle that case by overriding `when_not_ready` method
  # (by default it does nothing)
  def when_not_ready
    retry_job
  end
end

Don't forget to set active_job as a backend in the config:

CarrierWave::Backgrounder.configure do |c|
  c.backend :active_job, queue: :carrierwave
end

We use the after_commit hook when using active_record. This creates a problem when testing with Rspec because after_commit never gets fired if you're using trasactional fixtures. One solution to the problem is to use the TestAfterCommit gem. There are various other solutions in which case google is your friend.

The workers fetch the document with the mounted uploader using the model class name and id. Uploads on embedded documents cannot be obtained this way. If the position of the document in the root document structure is known, a workaround is to override the embedded models find method like this:

class SomeRootDocument
  include Mongoid::Document

  embeds_many :embedded_documents
end

class EmbeddedDocument
  include Mongoid::Document

  embedded_in :some_root_document

  mount_uploader :image, ImageUploader
  process_in_background :image

  def self.find(id)
    bson_id = Moped::BSON::ObjectId.from_string(id) # needed for Mongoid 3

    root = SomeRootDocument.where('embedded_documents._id' => bson_id).first
    root.embedded_documents.find(id)
  end
end

Copyright (c) 2011 Larry Sprock

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.