Note: This bug is displayed in read-only format because the product is no longer active in Red Hat Bugzilla.

Bug 1430337

Summary: [QE][Docs][RFE][OSP 16.0.2] Verify PCI Passthrough Configuration Steps for GPUs
Product: Red Hat OpenStack Reporter: Nikhil Shetty <nshetty>
Component: documentationAssignee: Irina <igallagh>
Status: CLOSED CURRENTRELEASE QA Contact: RHOS Documentation Team <rhos-docs>
Severity: urgent Docs Contact:
Priority: urgent    
Version: 9.0 (Mitaka)CC: amodi, asoni, cchen, ealcaniz, egallen, fbaudin, ggrimaux, igallagh, jhakimra, jraju, lars, lyarwood, mariel, mburns, nshetty, owalsh, pablo.iranzo, rchincho, rhos-docs, rlondhe, sbauza, sclewis, sgordon, sputhenp, srevivo
Target Milestone: Upstream M2Keywords: Documentation, FutureFeature, TestOnly, Triaged
Target Release: 16.1 (Train on RHEL 8.2)   
Hardware: All   
OS: Linux   
Whiteboard: docs-accepted
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2020-02-28 16:38:20 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:
Embargoed:
Bug Depends On:    
Bug Blocks: 1572787, 1732835, 1928217    

Description Nikhil Shetty 2017-03-08 12:10:24 UTC
Description of problem:
We do not have Any documentation which provides Steps for Enabling PCI Pass through on Openstack


Expected results: We Need documentation on how to Create PCI Pass through on Our Environment.


Additional info: We have Customers who wish to enable PCI Pass through on their Oepnstack Env.

Please do help in this Regards

Comment 1 Stephen Gordon 2017-03-09 05:37:34 UTC
Additionally note that we should provide two examples:

1) PCI Passthrough of a network device.

2) PCI Passthrough of a GPU for which device assignment is supported/tested from a RHEL POV:

https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/Virtualization_Deployment_and_Administration_Guide/sect-device-GPU.html

Comment 2 Lucy Bopf 2017-03-09 07:03:38 UTC
Hi Nikhil,

Thanks for raising this request. To help us triage this bug, can you specify which version(s) your customers are using and requesting documentation for?


Kind Regards,

Lucy

Comment 3 Nikhil Shetty 2017-03-09 07:12:48 UTC
Hi Lucy,

Thanks for the acknowledgement, The customer is trying to USE GPU as PCI Passthrough for Redhat Openstack Version 9.


I am using the upstream link to check if it works. You could refer the same link

URL:- https://docs.openstack.org/admin-guide/compute-pci-passthrough.html


Regards
Nikhil Shetty.

Comment 4 Stephen Gordon 2017-03-09 13:13:20 UTC
(In reply to Nikhil Shetty from comment #3)
> Hi Lucy,
> 
> Thanks for the acknowledgement, The customer is trying to USE GPU as PCI
> Passthrough for Redhat Openstack Version 9.
> 
> 
> I am using the upstream link to check if it works. You could refer the same
> link
> 
> URL:- https://docs.openstack.org/admin-guide/compute-pci-passthrough.html
> 
> 
> Regards
> Nikhil Shetty.

Indeed, ideally for the GPU case though I think we can add value by going a little further and showing the user how to verify the presence of a supported card from inside the guest.

Comment 5 Nikhil Shetty 2017-03-14 11:16:26 UTC
Hi,

Could you please, update on Documentation Availability of the same.

Comment 14 Stephen Gordon 2017-06-08 14:19:55 UTC
*** Bug 1321584 has been marked as a duplicate of this bug. ***

Comment 16 Stephen Gordon 2017-06-23 13:21:04 UTC
(In reply to Stephen Gordon from comment #1)
> Additionally note that we should provide two examples:
> 
> 1) PCI Passthrough of a network device.
> 
> 2) PCI Passthrough of a GPU for which device assignment is supported/tested
> from a RHEL POV:
> 
> https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/
> html/Virtualization_Deployment_and_Administration_Guide/sect-device-GPU.html

https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linux/7/html/Virtualization_Deployment_and_Administration_Guide/chap-Guest_virtual_machine_device_configuration.html#sect-device-GPU

Comment 18 Bryan Yount 2017-08-26 00:49:18 UTC
Here is the updated upstream docs URL for PCI Passthrough:
https://docs.openstack.org/nova/pike/admin/pci-passthrough.html#enable-pci-passthrough-compute

Comment 23 Lars Kellogg-Stedman 2018-05-02 00:41:58 UTC
*** Bug 1573269 has been marked as a duplicate of this bug. ***

Comment 24 Lars Kellogg-Stedman 2018-05-02 00:44:23 UTC
I've just marked #1573269 as a duplicate of this, but I wanted to include the following from the description of that bug:

The necessary instructions for blacklisting kernel graphics drivers when using GPU passthrough can be found in our "Virtualization Deployment and Administration Guide" [3], but they are buried in the section on the vGPU feature when in fact that are required for any GPU passhthrough.

[3]: https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/virtualization_deployment_and_administration_guide/sect-device-gpu#sect-device-GPU-vGPU

The instructions that come earlier in that document that suggest setting pci-stub.ids do not seem to be effective in practice.

Comment 41 Erwan Gallen 2018-12-16 23:45:25 UTC
Here is a blog post about RHOSP13 NVIDIA Tesla GPUs PCI Passthrough:
https://egallen.com/openstack-nvidia-tesla-gpu-passthrough/

Comment 58 Irina 2020-02-28 16:38:20 UTC
New content - "Enabling PCI passthrough for a GPU device" - available on the Customer Portal:

https://access.redhat.com/documentation/en-us/red_hat_openstack_platform/16.0/html-single/instances_and_images_guide/index?#gpu-pci-passthru

Comment 59 Irina 2020-03-09 16:23:47 UTC
*** Bug 1687947 has been marked as a duplicate of this bug. ***