Rake task copy between databases

Question

I have a Rails 4 app with a Mongoid database and want to introduce a sandbox environment for testing purpose. There is some data (two models), that I want to copy from the production db to the sandbox.

I would do this with a rake task, that is called by a cronjob. However, inside this rake task, I'm not sure how to establish two connections to the databases, and use the same model for different databases.

I was also thinking of doing it at the mongodb layer (like they do here How to copy a collection from one database to another in MongoDB), but one model consists of data, that only should be partially copied to the sandbox database. Therefore, I think I have to do it inside the Rails environment.

In this rake task, I can all my articles, but I don't know how to "push" them into the sandbox database:

namespace :sandbox do
  desc "Syncs production -> sandbox data"
    task(:sync => :environment) do |t, args|
      Article.all.each do |article|
        if my_model1.state == :active
          # here it should sync article to the sandbox models

          # and then, it should also sync all the comments to the sandbox models
          article.comments
        end
      end
    end
  end
end

score 1 · Accepted Answer · edited May 23 '17 at 12:28

I finally found the solution myself. The key are the mongoid sessions you can define inside mongoid.yml.

If you want to copy production data to sandbox, you can define a second session inside your sandbox environment (you start the task in sandbox env, and it gives you access to the production database):

sandbox:
  sessions:
    default:
      database: your_sandbox_db
      hosts:
        - localhost:27017
    production_sync:
      database: your_production_db
      hosts:
        - localhost:27017

You can access the production_sync session by defining a separate class for the model you want to sync like this:

class ArticleProduction
  include Mongoid::Document
  include Mongoid::Attributes::Dynamic
  store_in collection: "articles", session: "production_sync"
end

Now you can use this class to access the production database. If you want to copy all data from production to sandbox, do it like this:

production_articles = ArticleProduction.map(&:attributes)
production_articles.each do |attributes|
  Article.create(attributes)
end

I defined myself a rake task, that is called by a cronjob every day, and syncs data from production to sandbox.

One question remains: If you want to make sure, to not changing any data, some kind of read only mode would be useful (see here Make mongoid session read only).

Rake task copy between databases

1 Answers1