What storage method in uploader about carrierwave-aws HOT 9 CLOSED

CarrierWave does a dynamic lookup of the storage engine inside of the configuration block. If you say you want the storage to be :fog then it will try and use CarrierWave::Storage::Fog. If you want to use aws instead you need to configure it with storage = :aws. Here is the relevant part from the README:

CarrierWave.configure do |config|
  config.storage    = :aws
  config.aws_bucket = ENV.fetch('S3_BUCKET_NAME')
  config.aws_acl    = :public_read
  config.asset_host = 'http://example.com'
  config.aws_authenticated_url_expiration = 60 * 60 * 24 * 365

  config.aws_credentials = {
    access_key_id:     ENV.fetch('AWS_ACCESS_KEY_ID'),
    secret_access_key: ENV.fetch('AWS_SECRET_ACCESS_KEY')
  }
end

from carrierwave-aws.

CarrierWave does a dynamic lookup of the storage engine inside of
the configuration block. If you say you want the storage to be :fog
then it will try and use CarrierWave::Storage::Fog. If you want to
use aws instead you need to configure it with storage = :aws. Here
is the relevant part from the README:

CarrierWave.configure do |config|
config.storage = :aws
config.aws_bucket = ENV.fetch('S3_BUCKET_NAME')
config.aws_acl = :public_read
config.asset_host = 'http://example.com'
config.aws_authenticated_url_expiration = 60 * 60 * 24 * 365

config.aws_credentials = {
access_key_id: ENV.fetch('AWS_ACCESS_KEY_ID'),
secret_access_key: ENV.fetch('AWS_SECRET_ACCESS_KEY')
}
end

Yeah, I got that, but each uploader could have a storage method, with
an application having mixed methods. The carrierwave documentation
set it an uploader class, i.e.
class ImageUploader < CarrierWave::Uploader::Base
storage :file
and you can have N uploaders. With the initializer it's one-for-all.
By deduction, I'm inclined to guess that the uploader could state
storage :aws
and remove the config.storage line from the initializer? Or
am I waaay off base?

from carrierwave-aws.

No, you aren't off base. The storage deduction is being performed inside of CarrierWave's uploader in that case, but it uses the same exact lookup mechanism.

from carrierwave-aws.

Is there something else I can help with here?

from carrierwave-aws.

Is there something else I can help with here?

Thanks... & sorry if I let you hanging there!

My time zone's quite a bit ahead & lucidity has taken a dive. Will
give it a whirl tomorrow!

J

from carrierwave-aws.

Just to keep you informed, it is functioning in development mode. In production mode I keep hitting a SocketError getaddrinfo: nodename nor servname provided or not know. But that is not due to the SDK. I had earlier tried with fog with the same results. Sticking with this gem though when I understand what is causing the error... unless you can pick it out!
CarrierWave.configure do |config|
config.storage = :aws
config.aws_bucket = 'v4prove'
config.aws_acl = :public_read
config.asset_host = 'https://s3-eu-west-1.amazonaws.com/v4prove'
config.aws_authenticated_url_expiration = 60 * 60 * 24 * 365

config.aws_credentials = {
access_key_id: 'xxx',
secret_access_key: 'xxx'
}
end

from carrierwave-aws.

My guess is that the error is related to your asset_host, but I can't say for sure. I'm glad everything else is working for you though.

from carrierwave-aws.

HI Sorentwo, Got sidetracked for a few weeks on other matters.

I've resolved the above issues, but possibly in unexpected manners.
Minor points: mix 'n match methods means tracking everything down (I had on uploader amongst 7 showing as fog with the gem disabled. duh). I also had the syntax of calling https:// confused with carrierwave's host and endpoint syntassi.

Major point. config.aws_bucket = ENV.fetch('S3_BUCKET_NAME') was leading to the application complaining at start-up. /config/initializers/s3uploads.rb 22 in "fetch" gave me a clue and I had to amend all lines that used fetch as follows: config.aws_bucket = 'S3_BUCKET_NAME'

from carrierwave-aws.

Thanks for reporting back @dvodvo. This may help somebody else in the future.

from carrierwave-aws.