Comment 10 for bug 1816764

Revision history for this message
OpenStack Infra (hudson-openstack) wrote : Fix merged to stx-config (f/stein)

Reviewed: https://review.openstack.org/639397
Committed: https://git.openstack.org/cgit/openstack/stx-config/commit/?id=bf0aa2c78d2397073baa12c2efcf3bcf2cc9d84b
Submitter: Zuul
Branch: f/stein

commit 611a68a96ab915dc4e97d39dffa5c379bbffef3d
Author: Mingyuan Qi <email address hidden>
Date: Wed Jan 30 09:41:27 2019 +0800

    Allow user specified registries for config_controller

    Currently docker images were pulled from public registries during
    config_controller. For some users, the connection to the public
    docker registry may be slow such that installing the containerized
    services images may timeout or the system simply does not have
    access to the public internet.

    This change allows users to specify alternative public/private
    registries to replace k8s.gcr.io, gcr.io, quay.io and docker.io.
    Insecure registry is supported if all default registries were
    replaced by one unified registry. It lowers the complexity for
    those who build his own registry without internet access.

    Docker doesn't support ipv6 addr as registry name, instead
    hostname or domain name in ipv6 network is allowed.

    Test:
    AIO-SX/AIO-DX/Standard(2+2):
      Alternative public registry (ipv4/domain) with proxy
        - config_controller pass
      Private registry (ipv4/ipv6/domain) without internet
        - config_controller pass
      Default registry with/without proxy
        - config_controller pass

    Story: 2004711
    Task: 28742

    Change-Id: I4fee3f4e0637863b9b5ef4ef556082ac75f62a1d
    Signed-off-by: Mingyuan Qi <email address hidden>

commit cb4b30bf56195456ac6b8bd11abf7e23f90f81a4
Author: Angie Wang <email address hidden>
Date: Fri Feb 22 01:21:07 2019 -0500

    Solve the stx-openstack reapply issue on controller-1

    After stx-openstack applied, the stx-openstack reapply shouldn't
    trigger the charts reinstallation if there has no overrides changed
    for charts. However, the reinstallation happens after swacting active
    controller to controller-1 due to the generated images overrides on
    controller-1 are different from before. The images overrides generation
    requires walking through the stx-openstack charts stored under
    /scratch, but charts do not exist on controller-1's /scratch as it's
    an unreplicated filesystem. This causes the images overrides to differ
    between controller-1 and controller-0.

    This commit updates to walk through charts and get the images for
    charts during application-upload, then save the images list for each
    chart into the existing images file under aramda directory
    /opt/platform/armada. The images file would be used for retrieving
    the images for charts to generate images overrides.

    Closes-Bug: 1816173
    Change-Id: I4f00c3031decb063f8f126d0c837acd4dde56fc3
    Signed-off-by: Angie Wang <email address hidden>

commit a6934ac9d27e0357d0025018077441d989679409
Author: Bin Qian <email address hidden>
Date: Thu Feb 21 14:46:34 2019 -0500

    Boost sm process priority in VBox environment

    There is an instance that sm claimed its main thread ran sluggish
    as some critical timer run behind the scheuled timing.
    The issue could prevent the sm from scheduling services.
    As the result, the controller could fail to enable.

    The issue was found only on vbox labs on AIO-SX, the fix is to boost
    sm process priority to nice value -10 from current -2.

    Closes-Bug: 1816764
    Depends-On: https://review.openstack.org/638664
    Change-Id: Iafa17b1c47d65cc7394552ea1c8e7a78398e4869
    Signed-off-by: Bin Qian <email address hidden>