Comment 12 for bug 1647464

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix merged to nova (master)

Reviewed: https://review.openstack.org/407204
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=72d28ccd6e8ce8cf700891788f06578313b03c38
Submitter: Jenkins
Branch: master

commit 72d28ccd6e8ce8cf700891788f06578313b03c38
Author: Matt Riedemann <email address hidden>
Date: Mon Dec 5 16:24:05 2016 -0500

    Handle MarkerNotFound from cell0 database

    When listing instances in the cellv2 world we look them up
    from three locations:

    1. Build requests which exist before the instances are created
       in the cell database (after the scheduler picks a host to
       build the instance). Currently instances and build requests
       are both created before casting to conductor, but that's going
       away in Ocata with the support for multiple cellsv2 cells.
    2. The cell0 database for instances which failed to get scheduled
       to a compute host (and therefore a cell).
    3. The actual cell database that the instance lives in. Currently
       that's only a single traditional nova database, but could be one
       of multiple cellsv2 cells when we add that support in Ocata.

    If a marker is passed in when listing instances, if the instance
    lives in an actual cell database, we'll get a MarkerNotFound failure
    from cell0 because the instance doesn't exist in cell0, but we check
    cell0 before we check the cell database. This makes the instance
    listing short-circuit and fail with a 400 from the REST API.

    This patch simply handles the MarkerNotFound when listing instances
    from the cell0 database and ignores it so we can continue onto the
    cell database.

    Closes-Bug: #1647464

    Change-Id: I977497be262fb7f2333e32fb7313b29624323422