List view

GPU sharing on cuda compute capability >=7.5

bugIssue/PR to expose/discuss/fix a bugIssue/PR to expose/discuss/fix a bug

Status: Open (in progress).

NVIDIA/k8s-dra-driver-gpunumber 231

#231 In NVIDIA/k8s-dra-driver-gpu;

· guptaNswati opened on Jan 24, 2025

Draft: Add more logs to computedomain manager

debuggabilityissue/pr related to the ability to debug the systemissue/pr related to the ability to debug the system

Status: Open (in progress).

NVIDIA/k8s-dra-driver-gpunumber 285

#285 In NVIDIA/k8s-dra-driver-gpu;

· guptaNswati opened on Mar 12, 2025

CUDA_MPS_PINNED_DEVICE_MEM_LIMIT is not set when using MPS

bugIssue/PR to expose/discuss/fix a bugIssue/PR to expose/discuss/fix a bug

Status: Open.

#319 In NVIDIA/k8s-dra-driver-gpu;

· harche opened on Apr 16, 2025

Make container log verbosity configurable / dynamic

featureissue/PR that proposes a new feature or functionalityissue/PR that proposes a new feature or functionality

Status: Open.

#335 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on Apr 26, 2025

debuggability: allow for <code>compute-domain-daemon</code> container log inspection in case of startup failure

debuggabilityissue/pr related to the ability to debug the systemissue/pr related to the ability to debug the system

Status: Open.

#295 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on Mar 25, 2025

When the node GPU does not support setting timeslice, the plugin will crash directly.

bugIssue/PR to expose/discuss/fix a bugIssue/PR to expose/discuss/fix a bug

Status: Open.

#81 In NVIDIA/k8s-dra-driver-gpu;

· wawa0210 opened on Mar 5, 2024

ComputeDomain: add current state (health) and debug info to event stream/conditionals

debuggabilityissue/pr related to the ability to debug the systemissue/pr related to the ability to debug the system

Status: Open.

#350 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on May 12, 2025

ComputeDomain: explore exposing Prometheus metrics

debuggabilityissue/pr related to the ability to debug the systemissue/pr related to the ability to debug the system

Status: Open.

#352 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on May 12, 2025

ComputeDomain: support more than one domain per node, with subsets of GPUs

featureissue/PR that proposes a new feature or functionalityissue/PR that proposes a new feature or functionality

Status: Open.

#353 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on May 12, 2025

GPU sharing: fix time-slicing config for old devices (Tesla P4)

bugIssue/PR to expose/discuss/fix a bugIssue/PR to expose/discuss/fix a bug

Status: Open.

#363 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on May 14, 2025

GPUs: remove state associated with deleted ResourceClaims

robustnessissue/pr: edge cases & fault toleranceissue/pr: edge cases & fault tolerance

Status: Open.

#365 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on May 14, 2025

GPU sharing: support Dynamic MIG (using DRA partitionable devices)

featureissue/PR that proposes a new feature or functionalityissue/PR that proposes a new feature or functionality

Status: Open.

#361 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on May 14, 2025

GPU sharing: revisit MPS support (change semantics of config, and daemon control)

featureissue/PR that proposes a new feature or functionalityissue/PR that proposes a new feature or functionality

Status: Open.

#362 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on May 14, 2025

Productionize the HTTP request lifecycle of the webhook

robustnessissue/pr: edge cases & fault toleranceissue/pr: edge cases & fault tolerance

Status: Open.

#484 In NVIDIA/k8s-dra-driver-gpu;

· klueska opened on Aug 21, 2025

Reimplement release of workload pods in ComputeDomain in terms of BindingConditions

usabilityissue/pr related to UXissue/pr related to UX

Status: Open.

#653 In NVIDIA/k8s-dra-driver-gpu;

· klueska opened on Oct 8, 2025

ComputeDomain: do not require a priori node count (numNodes) for creation

featureissue/PR that proposes a new feature or functionalityissue/PR that proposes a new feature or functionality

Status: Open.

#364 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on May 14, 2025

Define new ComputeDomain API in <code>resource.nvidia.com/v1beta2</code>

APIissue/pr related to API (changes, specification, ...)issue/pr related to API (changes, specification, ...)

Status: Open.

#596 In NVIDIA/k8s-dra-driver-gpu;

· klueska opened on Sep 18, 2025

GPUs: take device offline when unhealthy (build logic in go-nvlib)

featureissue/PR that proposes a new feature or functionalityissue/PR that proposes a new feature or functionality

robustnessissue/pr: edge cases & fault toleranceissue/pr: edge cases & fault tolerance

Status: Open.

#360 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on May 14, 2025

DRA Admin Access integration

featureissue/PR that proposes a new feature or functionalityissue/PR that proposes a new feature or functionality

Status: Open.

#337 In NVIDIA/k8s-dra-driver-gpu;

· ritazh opened on Apr 29, 2025

Expose addressing mode as a device attribute

featureissue/PR that proposes a new feature or functionalityissue/PR that proposes a new feature or functionality

Status: Open.

#483 In NVIDIA/k8s-dra-driver-gpu;

· elezar opened on Aug 21, 2025

First-class the use case of multiple pods per node in a specific ComputeDomain

featureissue/PR that proposes a new feature or functionalityissue/PR that proposes a new feature or functionality

Status: Open.

#309 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on Mar 28, 2025

Improve checkpointing semantics and reliability across new versions

featureissue/PR that proposes a new feature or functionalityissue/PR that proposes a new feature or functionality

resource cleanupissue/pr related to resource cleanup / teardown / state reconciliationissue/pr related to resource cleanup / teardown / state reconciliation

Status: Open.

#507 In NVIDIA/k8s-dra-driver-gpu;

· klueska opened on Aug 28, 2025

Manage lifecycle of compute-domain-daemon's service account outside of helm

robustnessissue/pr: edge cases & fault toleranceissue/pr: edge cases & fault tolerance

security

Status: Open.

#586 In NVIDIA/k8s-dra-driver-gpu;

· klueska opened on Sep 17, 2025

wip: atomic updates to /etc/hosts

robustnessissue/pr: edge cases & fault toleranceissue/pr: edge cases & fault tolerance

Status: Open (in progress).

NVIDIA/k8s-dra-driver-gpunumber 537

#537 In NVIDIA/k8s-dra-driver-gpu;

· jgehrcke opened on Sep 4, 2025

Allow to modify procDriverNvidiaPath via ENV VAR or Flag

ci/testingissue/PR related to CI and/or testingissue/PR related to CI and/or testing

configissue/PR about user-facing configuration interfaceissue/PR about user-facing configuration interface

Status: Open.

#696 In NVIDIA/k8s-dra-driver-gpu;

· ArangoGutierrez opened on Oct 20, 2025