Skip to content
This repository

Nov 16, 2009

  1. Mathias Meyer

    Added explicit dependency on json gem. Closes #12

    roidrage authored

Aug 06, 2009

  1. Mathias Meyer

    Consistently use one version number all around.

    roidrage authored

Jun 26, 2009

  1. Mathias Meyer

    Merge commit 'ezmobius/master' into merge_ezmobius

    roidrage authored

Jun 25, 2009

  1. Mathias Meyer

    Rake default task is now to run specs.

    roidrage authored

Jun 22, 2009

  1. Raphaël Simon

    Bump up minor rev to make it easier to distinghish between vanilla an…

    …d RS nanite
    raphael authored

Mar 12, 2009

  1. ck

    Updated version to 0.4.0

    The version of Nanite is updated to reflect the significance of the recent changes:
    - support for Redis
    - addition of tags
    - nanite-agent stop 
    - nanite-mapper stop
    - writing pid file
    ck authored

Mar 07, 2009

  1. ck

    Removed dependency for Extlib, since it's not used.

    ck authored

Feb 16, 2009

  1. ck

    Updated Rakefile and gemspec and reflect version bump to 0.3.0.

    ck authored

Feb 14, 2009

  1. Distinct seperation of Mapper and Agent in terms of code and usage

    Ian Leitch authored

Jan 29, 2009

  1. James Tucker

    Added support for rdoc generation from Rakefile, and use hanna if ava…

    Signed-off-by: ezmobius <>
    raggi authored ezmobius committed

Jan 27, 2009

  1. adding nanite.gemspec so github builds a gem

    Ezra Zygmuntowicz authored

Jan 23, 2009

  1. Glenn Rempe

    Updated instructions for amqp installation. Require version >= 0.6.0.

    + Its preferable to install amqp gem from RubyForge
    + Minor update to README.rdoc related to --token
    Signed-off-by: ezmobius <>
    grempe authored ezmobius committed
  2. Dane Jensen

    experimental rake tasks to control rabbitmq.

    note that it requires that dtach is installed.
    Signed-off-by: ezmobius <>
    careo authored ezmobius committed

Jan 18, 2009

  1. Glenn Rempe

    Consolidate some example related code to the examples dir.

    - crew and rabbitconf should not be general binaries installed with the gem
    - the async rack example belongs in the examples dir
    - modified the Rakefile to no longer install crew and rabbitconf in the bin dir
    - modified README.rdoc to reference new location for rabbitconf
    - added some example related TODO's
    grempe authored

Jan 09, 2009

  1. Michael Klishin

    Add dependencies to gemspec

    michaelklishin authored
  2. Glenn Rempe

    Moving README so we can get some nice rdoc formatting on GitHub. Curr…

    …ently the README is broken in gem server and needs some love.
    grempe authored

Jan 05, 2009

  1. integrating halogrium's refactoring to make it actually work.

    Ezra Zygmuntowicz authored

Jan 02, 2009

  1. Glenn Rempe

    Added nanite-admin to gem binary list so now you can run 'nanite-admi…

    …n' from anywhere.
    grempe authored

Jan 01, 2009

  1. Glenn Rempe

    Cleaned up a few 'rake gem' warnings

    grempe authored
  2. Glenn Rempe

    De-Anonymize the generated gem.

    grempe authored

Oct 09, 2008

  1. slight changes

    Ezra Zygmuntowicz authored
  2. reworked mapper script. now called nanite-mapper.

    Ezra Zygmuntowicz authored

Sep 23, 2008

  1. commiting latest touch ups.

    Ezra Zygmuntowicz authored

Sep 05, 2008

  1. Added file broadcasting ability. You can now publish a file to a fanout

        exchange and have a nanite on each server in your cluster that subscribes
        to the named file broadcast 'channel'
        This allows for multi server depoyments with no clustered filesystem to
        try and maintain a replicated set of files. When an upload happens on one
        slice, it publishes the file to the exchange and all the subscribers write
        it out on their own filesystems in their Nanite.file_root
    Ezra Zygmuntowicz authored

Sep 01, 2008

  1. added a few Resource specific specs

    Ezra Zygmuntowicz authored

Aug 31, 2008

  1. big refactoring for security. Now agents listen on queues with crypto…

    names and they only share the hasehs queue names with the mapper. Any other
    agents in the system will only ever know their user name they used to  log into
    rabbitmq but will not know the name of the queue to send them ops on.
    Also when an agent makes a request for an op, he creates a temporary queue
    with a strong hashed name for the return results to come to. This way the
    worker agents still don't know the requestors true queue name. The return
    address queue is only temporary and is only used once and then remopved when
    all results for said op are collected.
    This system gives us better security then we had in XMPP since the mapper is
    the only thing that knows an agents true address  Temporary queues only take
    4bytes of ram in rabbitmq and are auto deleted after they are no longer in use.
    Ezra Zygmuntowicz authored

Aug 14, 2008

  1. updated libraries to do mapping in the mapper

    and reducing in the requesting client.
    can push 3000 ops to 30 agents in 12 seconds
    Ezra Zygmuntowicz authored

Aug 12, 2008

  1. Adding NEw Nanite Map/Reduce Agent Framework Code

    Ezra Zygmuntowicz authored
  2. deleting old code. this is the *new* nanite now!

    Ezra Zygmuntowicz authored

Mar 31, 2008

  1. Here be nanite. fear the self assembling builder of slices.

    Ezra Zygmuntowicz authored
Something went wrong with that request. Please try again.