Bug 1384850

Summary: [RFE] Puppet Validation for PMD with PCI matching the numa_node associated
Product: Red Hat OpenStack Reporter: Franck Baudin <fbaudin>
Component: openstack-tripleo-validationsAssignee: Florian Fuchs <flfuchs>
Status: CLOSED WONTFIX QA Contact: nlevinki <nlevinki>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: 10.0 (Newton)CC: apannu, atelang, beth.white, fbaudin, fherrman, gchamoul, jjoyce, jrist, jschluet, lsurette, slinaber, tvignaud, vchundur
Target Milestone: ---Keywords: FutureFeature, Triaged
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard: NeedsAllocation
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2018-06-12 09:59:52 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Franck Baudin 2016-10-14 09:25:44 UTC 
Description of problem:

When installing an OVS-DPDK compute node (NFV composable role), the list of CPU’s for PMD’s given by the operator shall be validated if it falls under the numa node associated with the PCI devices (DPDK nics), as the devices are connected to a given NUMA node and cross assignment between NICs and CPUs would kill the performances.
Comment 1 Jason E. Rist 2017-01-11 15:49:30 UTC 
Franck does this mean you want a new validation for tripleO/UI?
Comment 2 Franck Baudin 2017-01-20 17:38:44 UTC 
(In reply to Jason E. Rist from comment #1)
> Franck does this mean you want a new validation for tripleO/UI?

Ideally yes.
Comment 5 Vijay Chundury 2017-10-16 13:22:50 UTC 
This additional check as mentioned by Franck B is not required with the mistral derived parameters. As the workflow takes care of these checks.
Franck pls confirm with us if still in doubt and we can Closing this BZ.
Comment 7 Gaël Chamoulaud 2018-05-31 09:12:32 UTC 
(In reply to Vijay Chundury from comment #5)
> This additional check as mentioned by Franck B is not required with the
> mistral derived parameters. As the workflow takes care of these checks.
> Franck pls confirm with us if still in doubt and we can Closing this BZ.

Franck, could you please give your two cents on Vijay's comment ^^^^?
Comment 8 Franck Baudin 2018-06-12 09:59:52 UTC 
Vijay is right, now, from a UI perspective, we have a new track to help on this NUMA issue with lsurette. So closing this BZ.