[openstack-dev] [nova][scheduler] bug in handling of ISOLATE thread policy

Finucane, Stephen stephen.finucane at intel.com
Wed Jun 8 09:22:57 UTC 2016


On 08 Jun 10:11, Finucane, Stephen wrote:
> On 07 Jun 11:36, Chris Friesen wrote:
> > Hi,
> > 
> > The full details are available at
> > https://bugs.launchpad.net/nova/+bug/1590091 but the short version
> > is this:
> > 
> > 1) I'm running stable/mitaka in devstack.  I've got a small system
> > with 2 pCPUs, both marked as available for pinning.  They're two
> > cores of a single processor, no threads.
> > 
> > 2) I tried to boot an instance with two dedicated CPUs and a thread
> > policy of ISOLATE, but the NUMATopology filter fails my host.
> 
> This sounds like bug #1550317 [1], which has been fixed on master but
> clearly needs to be backported to stable/mitaka. I'll do this as soon
> as soon I figure out how to :)
> 
> Stephen
> 
> PS: Marked #1590091 as a duplicate, per above.
> 
> [1] https://bugs.launchpad.net/nova/+bug/1550317

Done (thanks to bauzas). Review is available here [1]. I'll also point
out the other bug fix that's available for the CPU thread policy
feature [2]. This fixes less obvious issues with the 'require' policy.

Stephen

[1] https://review.openstack.org/326944
[2] https://review.openstack.org/285232/



More information about the OpenStack-dev mailing list