So you have a scenario where a single request comes in and you need to split the work up between multiple threads?
Also by handlers do you mean the view servers that couch manages for the design doc code? Those can only handle a single request at a time but couch manages a pool of them. Just trying to figure out what the scenario is. On Oct 1, 2011, at 7:14 AM, Behrad Zari <[email protected]> wrote: >> The new externals feature in Apache CouchDB 1.1.0 supports >concurrency, it >> uses http instead of stdin/stdout. >> >http://davispj.com/2010/09/26/new-couchdb-externals-api.html > Thanks Robert, I knew about the Http Proxy feature new in 1.1! > In our deployment our business logic is stored code (js/py) on CouchDB that > should be run at server-side, we have Apache Web server as our front end. So > that may not be the case for us to use CouchDB's http proxy for running > server-side code! > In our case I think solutions are to > 1) use mod-python handlers in Apache (and write business logic code in Python > => No code sharing through the whole system) > 2) ADD!!! NodeJS to our architecture as processes managed by CouchDB > os-daemons to which Apache routes requests > 3) Write fake CouchDB update handlers and use them as Stored-JS-code-Runners > (with couchapp code sharing through documents), and call these handlers from > mod-python to fulfill the business logic. > 4)?
