|
| 1 | +## v0.15.2 |
| 2 | + |
| 3 | +Changes since `v0.15.1`: |
| 4 | + |
| 5 | +## Changes by Kind |
| 6 | + |
| 7 | +### Feature |
| 8 | + |
| 9 | +- Ray: Support RayJob InTreeAutoscaling by using the ElasticJobsViaWorkloadSlices feature. (#8284, @hiboyang) |
| 10 | + |
| 11 | +### Bug or Regression |
| 12 | + |
| 13 | +- Kubeflow TrainJob v2: fix the bug to prevent duplicate pod template overrides when starting the Job is retried. (#8271, @j-skiba) |
| 14 | +- MultiKueue: Fixed status sync for CRD-based jobs (JobSet, Kubeflow, Ray, etc.) that was blocked while the local job was suspended. (#8344, @IrvingMg) |
| 15 | +- MultiKueue: fix the bug that for Pod integration the AdmissionCheck status would be kept Pending indefinitely, |
| 16 | + even when the Pods are already running. |
| 17 | + |
| 18 | + The analogous fix is also done for the batch/Job when the MultiKueueBatchJobWithManagedBy feature gate is disabled. (#8288, @IrvingMg) |
| 19 | +- Scheduling: fix a bug that evictions submitted by scheduler (preemptions and eviction due to TAS NodeHotSwap failing) |
| 20 | + could result in conflict in case of concurrent workload modification by another controller. |
| 21 | + This could lead to indefinite failing requests sent by scheduler in some scenarios when eviction is initiated by |
| 22 | + TAS NodeHotSwap. (#8313, @mbobrovskyi) |
| 23 | +- TAS NodeHotSwap: fixed the bug that allows workload to requeue by scheduler even if already deleted on TAS NodeHotSwap eviction. (#8310, @mbobrovskyi) |
| 24 | +- TAS: fix a performance bug that continues reconciles of TAS ResourceFlavor (and related ClusterQueues) |
| 25 | + were triggered by updates to Nodes' heartbeat times. (#8355, @PBundyra) |
| 26 | +- TAS: fixed performance issue due to unncessary (empty) request by TopologyUngater (#8333, @mbobrovskyi) |
| 27 | + |
| 28 | +### Other (Cleanup or Flake) |
| 29 | + |
| 30 | +- Improve error messages for validation errors regarding WorkloadPriorityClass changes in workloads. (#8352, @olekzabl) |
| 31 | +- MultiKueue: improve the MultiKueueCluster reconciler to skip attempting to reconcile and throw errors |
| 32 | + when the corresponding Secret or ClusterProfile objects don't exist. The reconcile will be triggered on |
| 33 | + creation of the objects. (#8290, @mszadkow) |
| 34 | + |
1 | 35 | ## v0.15.1 |
2 | 36 |
|
3 | 37 | Changes since `v0.15.0`: |
|
0 commit comments