Alarm coordinator tests sporadically failed

Bug #1331453 reported by Mehdi Abaakouk
24
This bug affects 5 people
Affects Status Importance Assigned to Milestone
Ceilometer
Fix Released
Critical
Mehdi Abaakouk
oslo.messaging
Fix Released
Medium
Mehdi Abaakouk

Bug Description

Hi,

Sometime alarms coordinator tests sporadically failed,

This is due to the oslo.messaging listener that don't create the queues at creation time, but at the first bus polling.

The current workaround is to use eventlet.sleep() after the listener object creation and hope it polls the bus one time.

ie: http://git.openstack.org/cgit/openstack/ceilometer/tree/ceilometer/tests/alarm/test_rpc.py#n176

But oslo.messaging needs to be fixed to remove the workaround.

Cheers,

Mehdi Abaakouk (sileht)
Changed in ceilometer:
assignee: nobody → Mehdi Abaakouk (sileht)
status: New → Triaged
Changed in oslo.messaging:
assignee: nobody → Mehdi Abaakouk (sileht)
Revision history for this message
Mehdi Abaakouk (sileht) wrote :
description: updated
Mehdi Abaakouk (sileht)
Changed in oslo.messaging:
status: New → In Progress
Mehdi Abaakouk (sileht)
Changed in ceilometer:
importance: Undecided → Medium
Mark McLoughlin (markmc)
Changed in oslo.messaging:
importance: Undecided → Medium
Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix merged to oslo.messaging (master)

Reviewed: https://review.openstack.org/100889
Committed: https://git.openstack.org/cgit/openstack/oslo.messaging/commit/?id=e582da68f4c5edd3e4a3cdb62819c210c72dcfad
Submitter: Jenkins
Branch: master

commit e582da68f4c5edd3e4a3cdb62819c210c72dcfad
Author: Mehdi Abaakouk <email address hidden>
Date: Wed Jun 18 14:18:48 2014 +0200

    Ensures listener queues exist in fake driver

    The fanout queues of the fake driver are created at the first executor
    poll, but if we use eventlet executor and the fake driver, when the sender
    delivers a fanout message before the first poll, the message goes to the
    topic queue instead of the server fanout queue.

    The changes fixes that by ensuring the all queues exists when the
    listener is created.

    Closes bug #1331453

    Change-Id: I92e7c01dd87d634b741bbcaea92f48730fdd555e

Changed in oslo.messaging:
status: In Progress → Fix Committed
Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix proposed to oslo.messaging (stable/icehouse)

Fix proposed to branch: stable/icehouse
Review: https://review.openstack.org/107911

Eoghan Glynn (eglynn)
Changed in ceilometer:
importance: Medium → Critical
milestone: none → juno-3
status: Triaged → In Progress
Changed in oslo.messaging:
milestone: none → juno-2
status: Fix Committed → Fix Released
Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix merged to oslo.messaging (stable/icehouse)

Reviewed: https://review.openstack.org/107911
Committed: https://git.openstack.org/cgit/openstack/oslo.messaging/commit/?id=d822f6cbf0f41481fb934d8def228a1029eb7b44
Submitter: Jenkins
Branch: stable/icehouse

commit d822f6cbf0f41481fb934d8def228a1029eb7b44
Author: Mehdi Abaakouk <email address hidden>
Date: Wed Jun 18 14:18:48 2014 +0200

    Ensures listener queues exist in fake driver

    The fanout queues of the fake driver are created at the first executor
    poll, but if we use eventlet executor and the fake driver, when the sender
    delivers a fanout message before the first poll, the message goes to the
    topic queue instead of the server fanout queue.

    The changes fixes that by ensuring the all queues exists when the
    listener is created.

    Closes bug #1331453
    Related-Bug: #1334798

    Co-authored-by: Paul Michali <email address hidden>
    Change-Id: I92e7c01dd87d634b741bbcaea92f48730fdd555e
    (cherry picked from commit e582da68f4c5edd3e4a3cdb62819c210c72dcfad)
    (cherry picked from commit 0cfafac2468afe6b601a985495382aae3351b217)

tags: added: in-stable-icehouse
Revision history for this message
Eoghan Glynn (eglynn) wrote :

Apparently fixed by the combination of oslo-messaging fixes from Mehdi Abaakouk and Nejc Saje:

   https://review.openstack.org/100889

and:

  https://review.openstack.org/106993

Since olso-messaging 1.3.1 was finally released on Friday July 25th, the disappearance of this issue from the builds is very clear from the relevant log stash queries, e.g.:

  http://bit.ly/test-coordination-presence-failures

Changed in ceilometer:
status: In Progress → Fix Committed
Thierry Carrez (ttx)
Changed in ceilometer:
status: Fix Committed → Fix Released
Thierry Carrez (ttx)
Changed in ceilometer:
milestone: juno-3 → 2014.2
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Duplicates of this bug

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.