application design : background process

rubyonrails-talk
1

Hi im making a simple app which takes a users search term and scrapes several sites rss feeds for relevant results. The scrape can take upto 10seconds. I want to make this a background process so I can update the user with the scrapes status("currently scraping site x") and present results as they are scraped.

Im looking at backgroundrb and have read a guide in Advanced Rails Recipe however thats for an older version of backgroundrb and the api has changed. So im a little confused - im also new to processes and threads.

If my worker is responsible for scraping each site and providing updates do I have one worker which will be used by all the users currently online? I guess I then can use threads to process multiple search requests at times when the site is busy.

OR

do i create a new worker for each user? Would this also allow for concurrent processing?

2

I would use Workling + Starling, or the run_later plugin. You can find all of them at Github.

3

I've used them to send process to background, for a long running task, for example deliver a newsletter to thousands of users, which is a process which can take some time.