Issues

• [FEA]: Drain through the EvictionRequest API (KEP-4563) when the cluster serves it

  component/operatorneeds-triage
  Enhancement

  NVIDIA/nodewright#1039 · Filter by author ayuskauskas opened on Nov 13, 2026

• [BUG]: A merged service interrupt leaks into the interrupt Job of another node in the same pass

  component/operator
  Bug

  NVIDIA/nodewright#1024 · Filter by author 0z5a opened on Nov 12, 2026
  2

  

• CLI v0.5.x is built with Go 1.27.1 and x/net v0.58.0, affected by 10 Go advisories

  component/cli
  Bug

  NVIDIA/nodewright#1009 · Filter by author lockwobr opened on Nov 6, 2026
  

• [BUG]: operator: a node that never completes a higher-priority NodeWright holds a batch slot in every lower-priority NodeWright

  component/operator
  Bug

  NVIDIA/nodewright#1007 · Filter by author tdnvidia opened on Nov 6, 2026

• Remove the legacy skyhook.nvidia.com API group and the migration machinery in operator v0.22.0

  component/chart
  Epic

  NVIDIA/nodewright#986 · Filter by author lockwobr opened on Nov 5, 2026

• Phase 3 finalizer cleanup does not remove per-NodeWright Node conditions on any delete

  component/operatorgood first issue
  Bug

  NVIDIA/nodewright#983 · Filter by author Jbansal2 opened on Nov 5, 2026
  2

• [FEA]: Preflight stage and dynamic cordon/drain: run package checks before a change to decide what each node needs

  component/agentcomponent/clicomponent/operatordocneeds-triage
  Enhancement

  NVIDIA/nodewright#981 · Filter by author mlsorensen opened on Nov 5, 2026
  1

• [BUG]: A dependsOn cycle is admitted and silently blocks its NodeWright and every one ordered after it

  component/operator
  Bug

  NVIDIA/nodewright#980 · Filter by author ayuskauskas opened on Nov 5, 2026
  1

  

• [BUG]: A package version bump puts every selected node in the batch, ignoring interruptionBudget

  component/operator
  Bug

  NVIDIA/nodewright#979 · Filter by author ayuskauskas opened on Nov 5, 2026
  2

  

• Investigate whether a kubelet can mount older ConfigMap data than the config-hash stamp recorded at Job creation

  component/operatorneeds-triage
  Task

  NVIDIA/nodewright#975 · Filter by author nko-cpu opened on Nov 5, 2026
  

• test(e2e): assert the success TTL on completed Jobs in simple-nodewright

  component/testsgood first issue
  Task

  NVIDIA/nodewright#967 · Filter by author lockwobr opened on Nov 4, 2026
  1

  

• test(operator): convert the dal log-tail tests to Ginkgo

  component/operatorgood first issue
  Task

  NVIDIA/nodewright#965 · Filter by author lockwobr opened on Nov 4, 2026

• operator: keep the kustomize Role's webhook Secret name in sync with WEBHOOK_SECRET_NAME

  component/operatorgood first issue
  Enhancement

  NVIDIA/nodewright#945 · Filter by author lockwobr opened on Nov 2, 2026
  4

  

• test(operator): tighten the ImagePullFailed pod-order spec and fix the PullAlways comment

  component/operatorgood first issue
  Task

  NVIDIA/nodewright#931 · Filter by author lockwobr opened on Oct 30, 2026
  2

  

• [BUG]: operator: node state changes trigger repeated metadata ConfigMap updates

  component/operator
  Bug

  NVIDIA/nodewright#910 · Filter by author tdnvidia opened on Oct 28, 2026
  

• operator: repeated node-list messages flood container logs on large clusters

  component/operatorgood first issue
  Bug

  NVIDIA/nodewright#909 · Filter by author tdnvidia opened on Oct 28, 2026
  3

• ci: chart docs publishing waits on the NGC chart upload approval

  component/chartneeds-triage
  Task

  NVIDIA/nodewright#907 · Filter by author lockwobr opened on Oct 28, 2026

• operator: scope the Pod and ConfigMap caches to the operator namespace, like Jobs

  component/chartcomponent/operatorneeds-triage
  Enhancement

  NVIDIA/nodewright#906 · Filter by author ayuskauskas opened on Oct 27, 2026
  1

• [FEA]: operator: opt-in Node CREATE webhook to apply the runtime-required taint at registration

  component/chartcomponent/operatordocneeds-triage
  Enhancement

  NVIDIA/nodewright#901 · Filter by author lockwobr opened on Oct 27, 2026

• [FEA]: operator: node labels for NodeWright bootstrap completion and current convergence

  component/clicomponent/operatordocneeds-triage
  Enhancement

  NVIDIA/nodewright#900 · Filter by author lockwobr opened on Oct 27, 2026

• Terminal erroring node retains sticky batch priority and starves waiting nodes

  component/operator
  Bug

  NVIDIA/nodewright#890 · Filter by author gauricb03 opened on Oct 23, 2026
  4

  

• [FEA]: Package artifact cache: handle more digests in use than the cache holds

  component/chartcomponent/operator
  Enhancement

  NVIDIA/nodewright#886 · Filter by author ayuskauskas opened on Oct 22, 2026
  

• [FEA]: Package artifacts: kubelet-equivalent registry credentials through a node-side preflight stage

  component/agentcomponent/chartcomponent/operatordoc
  Enhancement

  NVIDIA/nodewright#871 · Filter by author ayuskauskas opened on Oct 20, 2026
  1

  

• operator: a node that never returns from an interrupt keeps its batch slot until its Node object is deleted

  component/operator
  Bug

  NVIDIA/nodewright#869 · Filter by author tdnvidia opened on Oct 19, 2026
  1

  

• [BUG]: Agent registry reads for package artifacts have no timeout

  component/agent
  Bug

  NVIDIA/nodewright#866 · Filter by author ayuskauskas opened on Oct 19, 2026