Highlights
- (alpha) Ray kubectl plugin
get,session,log,job submit
- (alpha) Kubernetes events: create Kubernetes events for important information about the interactions between KubeRay and the Kubernetes API server
- (alpha) Apache YuniKorn integration
Changelog
- [release] Update Ray image to 2.34.0 (#2303, @kevin85421)
- Revert "[release] Update Ray image to 2.34.0 (#2303)" (#2413, @kevin85421)
- Revert "[release] Update Ray image to 2.34.0 (#2303)" (#2413) (#2415, @kevin85421)
- [Build][kubectl-plugin] Add release script for kubectl plugin (#2407, @MortalHappiness)
- [Feat][kubectl-plugin] Add Long, Example, shell completion for kubectl ray log (#2405, @MortalHappiness)
- Support gang scheduling with Apache YuniKorn (#2396, @yangwwei)
- [Feat][Kubectl-Plugin]Implement kubectl ray job submit (#2394, @chiayi)
- Add 1K, 5K and 10K RayCluster/RayJob scalability test results (#2218, @andrewsykim)
- [Feat][kubectl-plugin] Add dynamic shell completion for kubectl ray session (#2390, @MortalHappiness)
- [Feature][RayJob]: Generate submitter and RayCluster creation/deletion events (#2389, @rueian)
- [RayJob] Add Failure Feedback (log and event) for Failed k8s Creation Task (#2306, @tinaxfwu)
- [Feat][Kubectl-Plugin] Implement kubectl session for RayJob and RayService (#2379, @MortalHappiness)
- [Feat][kubectl-plugin] Add instructions for static shell completion (#2384, @MortalHappiness)
- [Feat][RayJob] UserMode SubmissionMode (#2364, @MortalHappiness)
- [Feature] Add Kubernetes manifest validation in pre-commit. (#2380, @LeoLiao123)
- [Feature][RayCluster]: Generate GCS FT Redis Cleanup Job creation events (#2382, @rueian)
- [Chore][Minor] Add .gitignore to kubectl-plugin (#2383, @MortalHappiness)
- Remove default option for batch scheduler name (#2371, @yangwwei)
- RayCluster Headless Worker Service Should PublishNotReadyAddresses (#2375, @ryanaoleary)
- [CI][GitHub-Actions] Upgrade actions/upload-artifact to v4 (#2373, @MortalHappiness)
- add support for pipeline-parallel-size in vLLM example (#2370, @andrewsykim)
- Add kubectl ray cluster log command (#2296, @chiayi)
- [Chore] Fix lint errors caused by casting int to int32 (#2368, @kevin85421)
- [Feature][kubectl-plugin] Implement kubectl ray session (#2298, @MortalHappiness)
- Use longer exec probe timeouts for Head pods (#2353, @andrewsykim)
- Remove redundant log line that is failing golangci-lint (#2366, @andrewsykim)
- [Chore][Linter] Upgrade golangci-lint to 1.60.3 (#2362, @MortalHappiness)
- Add batch-scheduler option, deprecate enable-batch-scheduler option (#2300, @yangwwei)
- [Feature] Display reconcile failures as events (ServiceAccount) (#2290, @cchen777)
- [Feature][RayCluster]: Deprecate the RayCluster .Status.State field (#2288, @rueian)
- Don't print redundant time unit in the log message (#2335, @tczekajlo)
- [Refactor][sample-yaml-test] Create sampleyaml package and run tests in CI (#2312, @MortalHappiness)
- [Refactor] Fix CreatedWorkerPod for worker Pod deletion event and refactor logs (#2346, @kevin85421)
- raycluster_controller: generate events for failed pod creation (#2286, @MadhavJivrajani)
- [Refactor][kubectl-plugin] Rename filenames and variables based on kubectl repo (#2295, @MortalHappiness)