NoMethodError - undefined method `cleanup' for #<CloudCrowd::AssetStore:0x...>

[hchoroomi]

Calling Job#destroy throws undefined method `cleanup' error. Seems autoload doesn't work for some reason.

[jashkenas]

crowd -c config/cloud_crowd/production -e production console
irb(main):001:0> Job.first.destroy
New RightAws::S3Interface using shared connections mode
Opening new HTTP connection to s3.amazonaws.com:80
=> #<CloudCrowd::Job id: 4254, status: 2, inputs: "[true]", action: "backup_database", 
options: "{}", outputs: "[true]", time: 54.545893, callback_url: nil, email: nil, created_at: 
"2010-06-03 05:00:03", updated_at: "2010-06-03 05:00:57">

Works fine for me. Can you provide a little more context to make it reproducible?

[hchoroomi]

It works in crowd console. But when you try to access it externally it cannot load asset stores which of course is because it's not initialized correctly.

irb(main):002:0> require 'cloud-crowd'
irb(main):003:0> ActiveRecord::Base.establish_connection(...)
irb(main):004:0> CloudCrowd::Job.first
=> #<CloudCrowd::Job id: 1...
irb(main):005:0> CloudCrowd::Job.first.destroy
=> NoMethodError: undefined method...

Is there a way to configure CloudCrowd to work in this way? Setting asset store directly?

[jashkenas]

I'm sure you can always go and load the piece you need manually ... but there's probably a simpler way we can accomplish your goal.

If you're trying to clean up old jobs automatically, you can use a "201 Created" response code from your callback_url to signal that the job has been completed successfully, and may be removed.

If your jobs are more fire-and-forget, then you can use crowd cleanup, in conjunction with Cron, if you like.

Both are documented at the bottom of this page:

http://wiki.github.com/documentcloud/cloud-crowd/the-job-api