Skip to content

fix(operator): persist worker hash in DGD status - #13991

Draft
tmonty12 wants to merge 2 commits into
mainfrom
tmonty12/dyn-13495-status-worker-hash
Draft

fix(operator): persist worker hash in DGD status#13991
tmonty12 wants to merge 2 commits into
mainfrom
tmonty12/dyn-13495-status-worker-hash

Conversation

@tmonty12

Copy link
Copy Markdown
Contributor

Summary

  • Make status.currentWorkerHash the authoritative completed v2 worker hash, while retaining the v2 annotation as a downgrade-compatible mirror.
  • Backfill status from existing v2 annotations and restore an unambiguous legacy v1 DCD suffix when a replacement strips controller-owned annotations.
  • Update status on initial, unsupported-pathway, migration, and completed-rollout hash commits; preserve the target annotation while an in-progress rollout retries its central status write.
  • Expose the field in both served DGD API versions and CRD schemas.

Fixes #13495

Validation

  • Focused worker rollout controller tests passed.
  • API v1alpha1 and v1beta1 package tests passed.
  • Verified both DGD CRD versions expose status.currentWorkerHash as a string.

The unrestricted controller package suite was not run because this isolated worktree lacks the envtest etcd binary.

Signed-off-by: Thomas Montfort <tjmontfort12@gmail.com>
@github-actions github-actions Bot added fix deployment::k8s Relates to dynamo deployment in kubernetes labels Aug 28, 2026
Signed-off-by: Thomas Montfort <tjmontfort12@gmail.com>
@copy-pr-bot

copy-pr-bot Bot commented Aug 29, 2026

Copy link
Copy Markdown

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

deployment::k8s Relates to dynamo deployment in kubernetes fix size/L

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[BUG]: Operator treats a DGD with missing worker-hash annotations as a first deploy; the previous worker generation is never rolled or removed

1 participant