Comment 3 for bug 1107064

Reviewed: https://review.openstack.org/22495
Committed: http://github.com/openstack/oslo-incubator/commit/6f9cef85353155538fe14e5025e2d0bab5cc63e4
Submitter: Jenkins
Branch: master

commit 6f9cef85353155538fe14e5025e2d0bab5cc63e4
Author: Mark McLoughlin <email address hidden>
Date: Wed Feb 20 23:08:50 2013 +0000

    Revert "Implement replay detection."

    This reverts Ib0260a0c62e3d312d2e3448a125bed64d861319e (commit a603678)

    The issue we're trying to fix here is bug #1107064 - when using mirrored
    queues with AMQP, acks can be lost while a master is failing over to a
    slace causing the new slave to re-send messages which had previously
    been acked.

    The "replay detection" code applies to more than just amqp and also has
    the appearance of a security measure (e.g. the use of the term 'nonce')
    when clearly it serves no security purpose until we actually have
    message signing.

    Revert the "replay detection" approach in favour of the more targetted
    amqp bugfix.

    Change-Id: I8b8d15835c8b4c85cd388f5df08b60ff4c74e38d