Repository navigation
Conversation
The lspci example uses -DD, which prints neither the class code nor the [vendor:device] pair shown in the sample output; -nn does. The sample output also had a '.' where the PCI domain separator is ':'. The T4 entry sets externalResourceProvider: true, so its resourceName has to match the name the external device plugin advertises. The NVIDIA KubeVirt GPU device plugin upper-cases the name it derives from pci.ids (TU104GL_TESLA_T4), and resource names are case sensitive, so the mixed-case name in the example never matches. Use the upper-case name and say how to look up the advertised names on a node. Assisted-by: Claude <[email protected]> Signed-off-by: YoungjuKim <[email protected]>
|
[APPROVALNOTIFIER] This PR is NOT APPROVED This pull-request has been approved by: The full list of commands accepted by this bot can be found here. DetailsNeeds approval from an approver in each of these files:Approvers can indicate their approval by writing |
|
Hi @fjvbn2003. Thanks for your PR. I'm waiting for a kubevirt member to verify that this patch is reasonable to test. If it is, they should reply with Tip We noticed you've done this a few times! Consider joining the org to skip this step and gain Once the patch is verified, the new status will be reflected by the I understand the commands that are listed here. DetailsInstructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. |
Two small corrections in
docs/compute/host-devices.md.lspci example. The command is
lspci -DD, but the sample output shows the class code[0302]and the[10de:1eb8]vendor:device pair, which only-nnprints. The sample also has0000.65:00.0where the domainseparator is
:. On one of my nodes (pciutils 3.14.0):T4 resource name. The T4 entry in the
permittedHostDevicesexample setsexternalResourceProvider: true, sothe GPU is advertised by an external device plugin and
resourceName/deviceNamemust be the name that pluginuses. The NVIDIA KubeVirt GPU device plugin upper-cases the name it builds from
pci.ids(
strings.ToUpperinpkg/device_plugin/device_plugin.go), so the T4 is advertised asnvidia.com/TU104GL_TESLA_T4.Extended resource names are case sensitive, so a VMI copied from the example asks for
nvidia.com/TU104GL_Tesla_T4,which no node advertises. The same mismatch is in the plugin's own examples (NVIDIA/kubevirt-gpu-device-plugin#89,
fix proposed in NVIDIA/kubevirt-gpu-device-plugin#222).
KubeVirt's own device plugin (without
externalResourceProvider) registersresourceNameexactly as written, sothe other examples are fine. On my cluster
resourceName: "nvidia.com/GA106M_GeForce_RTX_3060_Mobile"withoutexternalResourceProvidershows up in the node's allocatable with that exact casing.Changes:
lspci -DD|grep NVIDIA→lspci -Dnn | grep NVIDIA, and0000.65:00.0→0000:65:00.0in the sample output.nvidia.com/TU104GL_Tesla_T4→nvidia.com/TU104GL_TESLA_T4in the KubeVirt CR and the VMI examples.plugin, and how to list the names on a node.
I used Claude (Anthropic) to help with the research and wording, including the checks on my nodes above. The commit carries
an
Assisted-bytrailer.