sigs.k8s.io/kueue@v0.6.2/CHANGELOG/CHANGELOG-0.4.md (about) 1 ## v0.4.2 2 3 Changes since `v0.4.1`: 4 5 ### Bug or Regression 6 7 - Adjust resources (based on LimitRanges, PodOverhead and resource limits) on existing Workloads when a LocalQueue is created (#1197, @alculquicondor) 8 - Fix resuming of RayJob after preempted. (#1190, @kerthcet) 9 10 ## v0.4.1 11 12 Changes since `v0.4.0`: 13 14 ### Bug or Regression 15 16 - Fixed missing create verb for webhook (#1053, @stuton) 17 - Fixed scheduler to only allow one admission or preemption per cycle within a cohort that has ClusterQueues borrowing quota (#1029, @alculquicondor) 18 - Prevent workloads in ClusterQueue with StrictFIFO from blocking higher priority workloads in other ClusterQueues in the same cohort that require preemption (#1030, @alculquicondor) 19 20 ## v0.4.0 21 22 Changes since `v0.3.0`: 23 24 ### Features 25 26 - Add LimitRange based validation before admission. #613 27 - Move the workloads evicted due to pods ready timeout to the end of the queue. #689 28 - Manage number of Pods as part of ClusterQueue quota. #732 29 - Add a new `withinClusterQueue` preemption policy, `LowerOrNewerEqualPriority`. #710 30 - Consider preempted workloads admitted until the owner job becomes inactive. #692 31 - Add `flavorUsage` to `status` in `LocalQueue`. #737 32 33 ### Production Readiness 34 35 36 ### Bug fixes 37 38 - Fix a bug that updates a queue name in workloads with an empty value when using framework jobs that use batch/job internally, such as MPIJob. #713