yahoo-eng-team team mailing list archive

[OpenStack Infra <1534886@xxxxxxxxxxxxxxxxxx>]

Reviewed:  https://review.openstack.org/268705
Committed: https://git.openstack.org/cgit/openstack/neutron-lbaas/commit/?id=7f3af5e414fc7c717b796e6c8dea918b7b08cc07
Submitter: Jenkins
Branch:    master

commit 7f3af5e414fc7c717b796e6c8dea918b7b08cc07
Author: Brandon Logan <brandon.logan@xxxxxxxxxxxxx>
Date:   Sat Jan 16 14:21:19 2016 -0600

    Fix timing bug that causes false negative tests
    
    All drivers call (or at least should) the helper method
    successful_completion method when they're driver action has
    completed succesfully.  Some tempest tests exposed an issue where
    the load balancer statuses would get updated before children were,
    causing an inconsistent behavior.  The load balancer should always
    have its status updated last.
    
    Change-Id: I4e422b7feaec5110ee856994631785e6e55a467e
    Closes-Bug: 1534886


** Changed in: neutron
       Status: In Progress => Fix Released

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to neutron.
https://bugs.launchpad.net/bugs/1534886

Title:
  neutron-lbaas drivers should not update child statuses before load
  balancer statuses

Status in neutron:
  Fix Released

Bug description:
  Some neutron jobs have run into an issue where a listener is created,
  and then once that listener gets created in the driver, the
  successful_completion method is called.  This method updates the load
  balancer's statuses first and then the listener's.  It should be the
  other way around and allows for a minor window where the statuses of
  the load balancer and children look inconsistent.

To manage notifications about this bug go to:
https://bugs.launchpad.net/neutron/+bug/1534886/+subscriptions