← Back to team overview

yahoo-eng-team team mailing list archive

[Bug 1938326] Re: Migration gets stuck at pre-migrating status if source compute node is down but maintenance enabled

 

Reviewed:  https://review.opendev.org/c/openstack/nova/+/802317
Committed: https://opendev.org/openstack/nova/commit/da6af3fb224d5033980537f8a677437c464c12b1
Submitter: "Zuul (22348)"
Branch:    master

commit da6af3fb224d5033980537f8a677437c464c12b1
Author: Lee Yarwood <lyarwood@xxxxxxxxxx>
Date:   Mon Jul 26 15:34:22 2021 +0100

    compute: Query the service group API within check_instance_host
    
    The check_instance_host decorator when called with the check_is_up kwarg
    is supposed to ensure that the compute service associated with a given
    instance is fully up. Previously it used the HostStatus returned by
    get_instance_host_status to determine this allowing either a state of UP
    or MAINTENANCE.
    
    However get_instance_host_status currently returns a HostStatus of
    MAINTENANCE if the service is simply disabled regardless of the
    underlying service being active or not. As a result of this and as set
    out in bug #1938326 the check_instance_host decorator will pass allowing
    calls to resize an instance to continue and later fail when RPC requests
    to the instance host eventually fail.
    
    To resolve this check_instance_host is now updated to only use the
    service group API to determine if the underlying service is up.
    
    Future work will look at updating the meaning of the MAINTENANCE
    HostStatus to mean disabled but up. As this will potentially
    change values returned in the API to admins doing an instance show this
    will need to be done behind a new microversion to ensure backward
    compatibility with the previous behaviour and meaning of the HostStatus.
    
    Closes-Bug: #1938326
    Change-Id: I45fd908b458e00b5e18848f624043ceb88b04825


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

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1938326

Title:
  Migration gets stuck at pre-migrating status if source compute node is
  down but maintenance enabled

Status in OpenStack Compute (nova):
  Fix Released

Bug description:
  Description
  ===========
  Currently nova rejects migration(resize) if the source compute node is down but not if the service has previously been disabled.

  Steps to reproduce
  ==================
  1. Create an instance
  2. Shutdown the compute node where the instance is started
  3. Enable maintenance of the nova-compute service on the source compute node
  4. Migrate the instance

  Expected result
  ===============
  Migration is rejected

  Actual result
  =============
  Migration is accepted but gets stuck in pre-migrating status

  Environment
  ===========
  1. Exact version of OpenStack you are running. See the following
    list for all releases: http://docs.openstack.org/releases/

     If this is from a distro please provide
         $ dpkg -l | grep nova
         or
         $ rpm -ql | grep nova
     If this is from git, please provide
         $ git log -1

  2. Which hypervisor did you use?
     (For example: Libvirt + KVM, Libvirt + XEN, Hyper-V, PowerKVM, ...)
     What's the version of that?

  2. Which storage type did you use?
     (For example: Ceph, LVM, GPFS, ...)
     What's the version of that?

  3. Which networking type did you use?
     (For example: nova-network, Neutron with OpenVSwitch, ...)

  Logs & Configs
  ==============

  https://bugzilla.redhat.com/show_bug.cgi?id=1985712#c0

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1938326/+subscriptions



References