Paperclip, S3, Heroku: Missing Image - ruby-on-rails

Long time viewer, first time asker. I've searched for this topic but don't believe I've found the answer.
I have a Post model that has an image. I'm using the Paperclip gem, saving to Amazon S3, and hosting on Heroku.
The file upload form works fine, because I can see that images are sent to my S3 bucket.
The issue is that, the images don't actually show in production.
Here's my model:
class Post < ActiveRecord::Base
attr_accessor :image_file_name, :image_content_type, :image_file_size, :image_updated_at
belongs_to :user
has_many :reviews
has_attached_file :image, styles: { medium: "700x500#", small: "350x250>" }
validates_attachment_content_type :image, content_type: /\Aimage\/.*\Z/
And here's my config/production.rb:
# Required for Paperclip / AWS
config.paperclip_defaults = {
:storage => :s3,
:s3_credentials => {
:bucket => ENV['S3_BUCKET_NAME'],
:access_key_id => ENV['AWS_ACCESS_KEY_ID'],
:secret_access_key => ENV['AWS_SECRET_ACCESS_KEY']
}
}
And here's my show.html.haml file:
.clearfix
.post_image_description
= image_tag #post.image.url(:medium)
.description= simple_format(#post.address)
.description= simple_format(#post.description)
Shouldn't the #post.image.url be enough? What may I be missing to properly route to the image?
This is what I see when I pull Heroku Logs:
2015-06-23T15:38:26.181383+00:00 app[web.1]: ActionController::RoutingError (No route matches [GET] "/images/medium/missing.png"):
For reference, here's my repository for the project: https://github.com/lucasvocos/pitstop
Please let me know if there is anything else to provide in the question, too. As this is my first time asking. Thanks everyone.

I had a similar problem that the upload was working but the display was showing a broken link. I checked the source for the link and it was not pointing to the url shown for the image on S3. I had to add the host name to my paperclip config
config.paperclip_defaults = {
:storage => :s3,
:s3_host_name => 's3-us-west-2.amazonaws.com',
:s3_credentials => {
:bucket => ENV['S3_BUCKET_NAME'],
:access_key_id => ENV['AWS_ACCESS_KEY_ID'],
:secret_access_key => ENV['AWS_SECRET_ACCESS_KEY']
}
}

You need to set heroku environment variable for s3 bucket variable:
heroku config:set S3_BUCKET_NAME='Your Bucket Name'
heroku config:set AWS_ACCESS_KEY_ID='Your AWS ID'
heroku config:set AWS_SECRET_ACCESS_KEY='Your AWS Secrete Key'

You have to commit the missing.png in \public\images\medium.(make sure \public\images\medium\missing.png exist, then commit). This images is the default until you upload some valid image.
However, also is recommended to define the url in model, something like this:
has_attached_file :image, styles: {
medium: "700x500#",
small: "350x250>" }, :url => '/:class/:attachment/:id/:style_:basename.:extension'

Related

AWS::S3::Errors::NoSuchKey No Such Key

I have a Rails 4 app with Paperclip running on Heroku. I recently started encountering a problem that is preventing me from updating any models with attached images. For example, any time I make a change through my site's CMS system to my Company model, the changes are not saved. I've tried changing only the image, the image and various other attributes, and only non-image attributes. It's as if the #company.save is never called in my controller's update action.
I found this issue report and tried the various solutions suggested, but nothing worked.
Here is my Company model:
class Company < ActiveRecord::Base
has_attached_file :logo, :styles => { :medium => "300x300>", :thumb => "100x100>" }
validates_attachment_content_type :logo, :content_type => /\Aimage\/.*\Z/
end
In my logs, I get the following error:
[AWS S3 404 0.022711 0 retries] head_object(:bucket_name=>"myApp",:key=>"companies/profile_photos/000/000/118/original/2015-09-05_19.05.34.jpg") AWS::S3::Errors::NoSuchKey No Such Key
I verified that my S3 credentials are set using heroku config and that they are valid.
I'm not certain, but I suspect that this may have something to do with my recent addition of the following buildpacks:
=== myApp Buildpack URLs
1. https://github.com/bobbus/image-optim-buildpack.git
2. https://github.com/heroku/heroku-buildpack-ruby.git
I've not been able to reproduce the problem on localhost.
Update
I've since opened up an issue report on this problem.
Here is my config.paperclip_defaults block in production.rd
config.paperclip_defaults = {
:storage => :s3,
:s3_protocol => :https,
:s3_host_name => 's3.amazonaws.com',
:s3_credentials => {
:bucket => ENV['S3_BUCKET_NAME'],
:access_key_id => ENV['AWS_ACCESS_KEY_ID'],
:secret_access_key => ENV['AWS_SECRET_ACCESS_KEY']
}
}
Here is my aws.yml:
development:
access_key_id: <%= ENV['AWS_ACCESS_KEY_ID'] %>
secret_access_key: <%= ENV['AWS_SECRET_ACCESS_KEY'] %>
production:
access_key_id: <%= ENV['AWS_ACCESS_KEY_ID'] %>
secret_access_key: <%= ENV['AWS_SECRET_ACCESS_KEY'] %>
Documentation from the GitHub paperclip wiki
Assuming you are using dotenv-rails...
In development.rb and production.rb files:
config.paperclip_defaults = {
storage: :s3,
s3_host_name: ENV['S3_HOST_NAME'],
bucket: ENV['S3_BUCKET_NAME']
}
Make sure to set these variables in your Heroku environment:
$ heroku config:set S3_HOST_NAME=your-s3-hostname-here S3_BUCKET_NAME=your-production-bucket-name-here
Paperclip-with-Amazon-S3
Documentation from rubydoc.info
This states you should have a yaml file with these details specified or write a method on your model.
rubydoc.info Paperclip::Storage::S3
No Such Key is S3's way of saying "404 Not Found."
I don't believe it has any alternate meanings.
The request was authorized and syntactically valid, but there's no file in the bucket at the specified path.
You may want to inspect the contents of your bucket from the AWS console.

Paperclip and S3 storage seed

I've setup my Rails 4 app to use S3 storage over Paperclip gem. Upload of files is working perfectly through create method, images are show using = image_tag(#listing.avatar.url)...
The problem I have is when I use seed_dump and create the seed.db; reset the database and seed the data again the paths to S3 where the images are not correct anymore?!
application.rb
config.paperclip_defaults = {
:storage => :s3,
:s3_credentials => {
:bucket => ENV['FOG_DIRECTORY'],
:access_key_id => ENV['AWS_ACCESS_KEY_ID'],
:secret_access_key => ENV['AWS_SECRET_ACCESS_KEY']
}
}
listing.rb
has_attached_file :avatar
validates_attachment_content_type :avatar, :content_type => /\Aimage\/.*\Z/
Long story short: how do you dump your data and seed it back again without losing the avatar URL?

Rails 4, Paperclip, Amazon S3 Config Amazon Path

I'm trying to configure the endpoint which is returned from paperclip when my object is successfully uploaded to Amazon's S3 service. The upload and everything is working correctly, but the URL that is being returned is incorrect for displaying the upload.
Right now, the url that is being returned is http://s3.amazonaws.com/path/to/my/items (as seen in the picture below).
Instead of s3.amazonaws.com, I would like the root to be specific to the bucket's location (e.g. s3-us-west-1.amazonaws.com/path/to/my/items)
Where should I try and configure a different url path (from s3.amazonaws.com to something else)? I've tried to add a url with the above path into my configuration file like:
#Paperclip Amazon S3
config.paperclip_defaults = {
:storage => :s3,
:url => "https://s3-us-west-1.amazonaws.com/",
:s3_credentials => {
:bucket => ENV['S3_BUCKET_NAME'],
:access_key_id => ENV['AWS_ACCESS_KEY_ID'],
:secret_access_key => ENV['AWS_SECRET_ACCESS_KEY']
}
}
Which did not appear to have any effect. Please advise on where I should be setting this option!
Thanks in advance!
If you're going to use S3, we've found that you have to include the S3 credentials in your actual model (not just the config files). Here's what we do:
Model
#Image Upload
Paperclip.options[:command_path] = 'C:\RailsInstaller\ImageMagick'
has_attached_file :image,
:styles => { :medium => "x300", :thumb => "x100" },
:default_url => "****",
:storage => :s3,
:bucket => '****',
:s3_credentials => S3_CREDENTIALS,
:url => "/:image/:id/:style/:basename.:extension",
:path => ":image/:id/:style/:basename.:extension"
config/application.rb
# Paperclip (for Amazon) (we use EU servers)
config.paperclip_defaults = {
:storage => :s3,
:s3_host_name => 's3-eu-west-1.amazonaws.com'
}
config/s3.yml
#Amazon AWS Config
development:
access_key_id: **********
secret_access_key: **************
bucket: ****
production:
access_key_id: ***********
secret_access_key: ***********
bucket: ****
Hope this helps?
I also had the same problem when migrating to Spree 2.2 and am still not sure how to solve it the correct way. It seems like Paperclip should have been updating the path from the configuration, but it isn't.
Lacking a better solution, I've overridden the Spree::Image class like this:
1 Spree::Image.class_eval do
2 has_attached_file :attachment,
3 styles: { mini: '48x48>', small: '100x100>', product: '240x240>', large: '600x600>' },
4 default_style: :product,
5 url: '/spree/products/:id/:style/:basename.:extension',
6 path: 'products/:id/:style/:basename.:extension',
7 convert_options: { all: '-strip -auto-orient -colorspace sRGB' }ยท
8 end
After some experimentation I have found that setting :s3_host_name globally suffices. I ended up with the same problem because I was setting :s3_region, which was being used by Paperclip (post-4.3.1, with aws-sdk 2) for storing attachments, but not when generating the URLs.
This may also be of interest to readers who end up on this problem: https://github.com/thoughtbot/paperclip/wiki/Restricting-Access-to-Objects-Stored-on-Amazon-S3

paperclip working in development but not working in production?

I'm pretty new to rails and seem to be having an issue with the paperclip gem. I installed the gem and it works well in development (localhost:3000) but when I'm running it on the heroku server, for some reason it does not want to attach files, and the app breaks (error 500 page).
Here is the process i ran... I pushed my file to heroku, heroku ran rake db:migrate (to add paperclip migrations), and then I ran heroku restart (to restart the app with new migrations). This did not seem to help.
Here is the code that I have for paperclip:
user.rb model:
has_attached_file :avatar,
:styles => {:small => "70x70>"},
:url => "/users/:attachment/:id/:style/:basename.:extension",
:path => ":rails_root/public/users/:attachment/:id/:style/:basename.:extension"
validates_attachment_size :avatar, :less_than => 500.kilobytes
validates_attachment_content_type :avatar, :content_type => ['image/jpeg', 'image/png']
edit_form.html.haml view:
= form_for (#user || User.new), :html => { :multipart => true } do |f|
...
.profile_picture.text_field
= image_tag current_profile.avatar.url(:small)
%br
= f.file_field :avatar
Again, for some reason it runs great in development, but breaks down in production. Any pointers would be greatly appreciated... I just cant seem to figure this out and it's pretty frustrating. Thank you so much for your time and any help!
In your model.
has_attached_file :picture,
:styles => {:large => "275x450>"},
:storage => :s3,
:s3_credentials => "#{RAILS_ROOT}/config/s3.yml",
:path => "appname/:attachment/:style/:id.:extension"
In s3.yml in your config dir:
development:
bucket: bucketname
access_key_id: key
secret_access_key: key
production:
bucket: bucketname
access_key_id: key
secret_access_key: key
Then go signup for a bucket at Amazon S3: http://aws.amazon.com/s3/
You could be having a few problems. However, the first is that you can not write to the file system on Heroku. You will have to implement a different storage mechanism such as s3. You can read about this limitation here: http://devcenter.heroku.com/articles/read-only-filesystem

AWS::S3::MissingAccessKey in Paperclip but I've defined both

I'm on Heroku, and this is a portfolio thing which I'm putting up on github for potential employers to look at, so obviously I don't want to stick my keys in a S3.yml file. I've exported S3_KEY and S3_SECRET to my environment both on Heroku and my machine and ruby can access them. But when I try and upload, it gives me the following error:
AWS::S3::MissingAccessKey in Portfolio itemsController#update
You did not provide both required access keys. Please provide the access_key_id and the secret_access_key.
The trace is irrelevant except for my controller line #, which works fine until I try and upload a file. Here's what I have:
class Asset < ActiveRecord::Base
attr_accessible :image, :image_file_name, :image_content_type, :image_file_size, :portfolio_item_id, :order
has_attached_file :image,
:styles => {
:thumb => "100x100#",
:small => "300x300",
:large => "600x600>"
},
:storage => :s3,
:s3_credentials => {
:access_key_id => ENV["S3_KEY"],
:secret_access_key => ENV["S3_SECRET"]
},
:bucket => "bucketybucket",
:path => "portfolio"
end
Anyone know what's going on here? How am I constructing this hash wrong?
Oh, and I've followed this thread, no dice: Paperclip and Amazon S3 Issue
same problem...
seems like that ENV const doesn't load before loading the module. solve by using file argument
like this
:s3_credentials => Rails.root.join('config/amazon_s3.yml')
and in amazon_s3.yml
access_key_id: 'your_key'
secret_access_key: 'your_sec_key'
bucket: 'somebucket'
furthermore, you can set environment variable by using heroku config:add command, which is describe in Heroku DevCenter
The problem is because the Enviroment variable in heroku is different that the enviroment variable in your system, so it may happen that the application works just in one enviroment

Resources