yahoo-eng-team team mailing list archive

[Nan Zhang <zn.share@xxxxxxxxx>]

in cases:
compute01
    vcpu_pin_set=4-31
compute02
    vcpu_pin_set=8-31

boot instance01 on compute01, grep virsh xml:
virsh dumpxml instance-0000006b | grep cpu 
  <vcpu placement='static' cpuset='4-31'>2</vcpu>

live-migration instance01 to compute02, grep virsh xml:
virsh dumpxml instance-0000006b | grep cpu 
  <vcpu placement='static' cpuset='4-31'>2</vcpu>

after live-migration instance's cpuset is not changed by compute02
vcpu_pin_set config, I thin it is a bug.


** Changed in: nova
       Status: Invalid => Confirmed

-- 
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/1496135

Title:
  libvirt live-migration will not honor destination vcpu_pin_set config

Status in OpenStack Compute (nova):
  Confirmed

Bug description:
  Reporting this based on code inspection of the current master (commit:
  9f61d1eb642785734f19b5b23365f80f033c3d9a)

  When we attempt to live-migrate an instance onto a host that has a
  different vcpu_pin_set than the one that was on the source host, we
  may either break the policy set by the destination host or fail (as we
  will not recalculate the vcpu cpuset attribute to match that of the
  destination host, so we may end up with an invalid range).

  The first solution that jumps out is to make sure the XML is updated
  in
  https://github.com/openstack/nova/blob/6d68462c4f20a0b93a04828cb829e86b7680d8a4/nova/virt/libvirt/driver.py#L5422

  However that would mean passing over the requested info from the
  destination host.

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1496135/+subscriptions