Welcome to CouchProxy

CouchProxy is a simple proxy server that distributes reads and writes to a cluster of Apache CouchDB servers so they appear to be a single huge database. Documents are stored and retrieved from a particular CouchDB instance, using consistent hashing of the document id. Map/reduce views are processed concurrently on each CouchDB instance and merged together by the proxy before returning the results to the client.

CouchProxy uses the Thin web server and EventMachine for asynchronous IO. There are no blocking IO calls in the server, which allows it to handle many clients and many backend CouchDB nodes concurrently, using very little processor and memory.

Usage

  1. gem install couchproxy

  2. couchproxy config

  3. Configure CouchDB nodes in couchproxy.yml

  4. couchproxy start

Dependencies

  • em-http-request >= 0.2.15

  • json >= 1.4.6

  • json-stream >= 0.1.0

  • thin >= 1.2.7

  • rbtree >= 0.3.0

  • ruby >= 1.9.1

Contact

Project contact: David Graham <[email protected]>

License

CouchProxy is released under the MIT license. Check the LICENSE file for details.