Asking for different vGPU types is racey

[Sylvain Bauza]

When testing on Victoria virtual GPUs, I wanted to have different types :

[devices]
enabled_vgpu_types = nvidia-320,nvidia-321

[vgpu_nvidia-320]
device_addresses = 0000:04:02.1,0000:04:02.2

[vgpu_nvidia-321]
device_addresses = 0000:04:02.3

Unfortunately, I saw that only the first type was used.
When restarting the nova-compute service, we got the log :
WARNING nova.virt.libvirt.driver [None req-a23d9cb4-6554-499c-9fcf-d7f9706535ef None None] The vGPU type 'nvidia-320' was listed in '[devices] enabled_vgpu_types' but no corresponding '[vgpu_nvidia-320]' group or '[vgpu_nvidia-320] device_addresses' option was defined. Only the first type 'nvidia-320' will be used.

It's due to the fact that we call _get_supported_vgpu_types() first when creating the libvirt implementation [1] while we only register the new CONF options by init_host() [2] which is called after.

[1] https://github.com/openstack/nova/blob/90777d790d7c268f50851ac3e5b4e02617f5ae1c/nova/virt/libvirt/driver.py#L418

[2] https://github.com/openstack/nova/blob/90777d7/nova/compute/manager.py#L1405

A simple fix would just be to make sure we have dynamic options within _get_supported_vgpu_types()

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.opendev.org/758470

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/758470
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=2bd8900d9b2600fba74341e249701051fb78eb37
Submitter: Zuul
Branch: master

commit 2bd8900d9b2600fba74341e249701051fb78eb37
Author: Sylvain Bauza <email address hidden>
Date: Thu Oct 15 19:19:38 2020 +0200

    Fix the vGPU dynamic options race

    As we lookup the existing dynamic options *before* creating them as
    _get_supported_vgpu_types() is called *before* compute init_host(),
    we need to make sure we call again the dynamic options registration
    within it.

    Change-Id: Ib9387c381d39fac389374c731b210815c6d4af03
    Closes-Bug: #1900006

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/nova 23.0.0.0rc1 release candidate.

[Sylvain Bauza (sylvain-bauza)]

Victoria backport candidate : https://review.opendev.org/c/openstack/nova/+/784907

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/nova 22.2.1 release.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: stable/ussuri
Review: https://review.opendev.org/c/openstack/nova/+/831524

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.opendev.org/c/openstack/nova/+/831524
Committed: https://opendev.org/openstack/nova/commit/2ddc2e6ab0259fdc1437998542fb5fd020dfef31
Submitter: "Zuul (22348)"
Branch: stable/ussuri

commit 2ddc2e6ab0259fdc1437998542fb5fd020dfef31
Author: Sylvain Bauza <email address hidden>
Date: Thu Oct 15 19:19:38 2020 +0200

    Fix the vGPU dynamic options race

    As we lookup the existing dynamic options *before* creating them as
    _get_supported_vgpu_types() is called *before* compute init_host(),
    we need to make sure we call again the dynamic options registration
    within it.

    Change-Id: Ib9387c381d39fac389374c731b210815c6d4af03
    Closes-Bug: #1900006
    (cherry picked from commit 2bd8900d9b2600fba74341e249701051fb78eb37)
    (cherry picked from commit c7d9d6d9dd25e21ec76ceea294cdf1690686a086)

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/nova ussuri-eol release.