Uploaded image for project: 'Red Hat OpenStack Services on OpenShift'
  1. Red Hat OpenStack Services on OpenShift
  2. OSPRH-12886

BZ#2185897 [TestOnly] live-migrated instance in MIGRATING status for a long time

XMLWordPrintable

    • False
    • False
    • No Docs Impact
    • openstack-neutron-18.6.1-17.1.20250408181013.85ff760.el9ost
    • None
    • Moderate

      Description of problem:
      This bug is similar to BZ2172873. That one was closed because the tobiko migration tests were not correct. The tests have been fixed, but test_7_live_migrate_server_with_host still fails sometimes.

      I have reproduced the issue manually, so I'll focus in the manual reproducer, which is simple.
      Apparently, the operation only fails when the live-migrated VM has a trunked port.

      A VM is created with a trunk port using the following commands:
      $ openstack port create --network heat_tempestconf_network parent-trunk-port
      $ openstack network trunk create --parent-port parent-trunk-port parent-trunk
      $ openstack network create heat_tempestconf_network-trunk
      $ openstack subnet create --network heat_tempestconf_network-trunk --subnet-range 10.111.222.0/24 heat_tempestconf_network-trunk-subnet
      $ openstack port create --network heat_tempestconf_network-trunk --mac-address fa:16:3e:5c:51:b8 subport-trunk-port # the mac corresponds with the parent-trunk-port mac
      $ openstack network trunk set --subport port=subport-trunk-port,segmentation-type=vlan,segmentation-id=55 parent-trunk
      $ openstack server create --flavor ubuntu --image tobiko.openstack.stacks._ubuntu.UbuntuImageFixture --port parent-trunk-port ubuntu0

      The VM is successfully created and its status is ACTIVE.

      Then, the following command is run to perform the live migration:
      $ openstack server migrate --live-migration ubuntu0

      It often fails the second time and sometimes fails the third time the live-migration operation is performed. I don't know if it is a coincidence or not, but the first live-migration never failed during my tests.
      The VM remains with status=MIGRATING and task_state=migrating for hours. Then, apparently nova cancels the migration and the VM status changes to ACTIVE, but the migration has not really occured (the VM's hypervisor is the previous one before the migration).

      I'm assigning this bug to component neutron because it can't be reproduced when there is no trunk port, so I assume there may be something wrong in neutron, but it could be a nova issue too.

      I will provide logs in a later comment.

      Version-Release number of selected component (if applicable):
      RHOS-17.1-RHEL-9-20230404.n.1

      How reproducible:
      100% (the migration command has to be executed two or three times on the VM)

              rodolfo_alonso Rodolfo Alonso
              jira-bugzilla-migration RH Bugzilla Integration
              Eduardo Olivares Toledo Eduardo Olivares Toledo
              rhos-dfg-networking-squad-neutron
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

                Created:
                Updated:
                Resolved: