Wrong error management when destroying a compute site

[Nuanda]

I tried to destroy Cyfronet prod compute site on cloud-dev deployment (it should not be listed there since there is now a dedicated Atmosphere that manages that CS), and I couldn't remove it due to forbidden message ("You are not allowed to delete the image"). Which might or might not be ok.

The system, instead of showing me that flash error, went 500.

Part of the stack trace:

I, [2015-05-12T17:51:51.273601 #2746]  INFO -- : Completed 500 Internal Server Error in 129ms (ActiveRecord: 15.0ms)
F, [2015-05-12T17:51:51.451184 #2746] FATAL -- : 
Excon::Errors::Forbidden (Expected(204) <=> Actual(403 Forbidden)
excon.error.response
  :body          => "{\"forbidden\": {\"message\": \"You are not allowed to delete the image.\", \"code\": 403}}"
  :headers       => {
    "Content-Length"       => "83"
    "Content-Type"         => "application/json; charset=UTF-8"
    "Date"                 => "Tue, 12 May 2015 13:52:53 GMT"
    "X-Compute-Request-Id" => "req-c9ab8cbc-e54c-4b2a-9569-48c85a1a4127"
  }                                                                                                                                                                         
  :local_address => "10.184.0.3"                                                                                                                                            
  :local_port    => 50808                                                                                                                                                   
  :reason_phrase => "Forbidden"                                                                                                                                             
  :remote_ip     => "149.156.11.2"                                                                                                                                          
  :status        => 403
):
  vendor/bundle/ruby/2.1.0/gems/excon-0.45.1/lib/excon/middlewares/expects.rb:6:in `response_call'
  vendor/bundle/ruby/2.1.0/gems/excon-0.45.1/lib/excon/middlewares/response_parser.rb:8:in `response_call'
  vendor/bundle/ruby/2.1.0/gems/excon-0.45.1/lib/excon/connection.rb:372:in `response'
  vendor/bundle/ruby/2.1.0/gems/excon-0.45.1/lib/excon/connection.rb:236:in `request'
  vendor/bundle/ruby/2.1.0/gems/fog-core-1.30.0/lib/fog/core/connection.rb:81:in `request'
  vendor/bundle/ruby/2.1.0/gems/fog-1.29.0/lib/fog/openstack/compute.rb:351:in `request'
  vendor/bundle/ruby/2.1.0/gems/fog-1.29.0/lib/fog/openstack/requests/compute/delete_image.rb:6:in `delete_image'
  vendor/bundle/ruby/2.1.0/gems/fog-1.29.0/lib/fog/openstack/models/compute/image.rb:44:in `destroy'
  vendor/bundle/ruby/2.1.0/gems/fog-core-1.30.0/lib/fog/core/collection.rb:56:in `destroy'

[tbartynski]

@Nuanda cs delete feature is not well tested. We actually never used this on production. Is there a confirm dialog with a warning that all images and vms on this that are manged by Atmo will be deleted? Maybe we should rethink the destroy cascade?

[Nuanda]

This is rather about properly managing the error message as a flash notification for the user.

[Nuanda]

When this particular fog error is triggered? Can I somehow "simulate" it on my dev machine?

[Nuanda]

#136 is related to this.