All Products
Search

This Product

    Container Service for Kubernetes:kube-scheduler

    Document Center

    Container Service for Kubernetes:kube-scheduler

    Last Updated:Jul 15, 2026

    kube-scheduler is a control plane component. It assigns pods to suitable nodes in a cluster based on node resource usage and pod scheduling requirements.

    Introduction

    kube-scheduler

    kube-scheduler selects a valid node for each pod in the scheduling queue based on the resource request of the pod and the allocatable resources on the node. In addition, kube-scheduler can sort all valid nodes in a specific order and select a suitable node to host the pod. By default, kube-scheduler spreads pods across nodes based on pod requests. For more information, see kube-scheduler.

    Filter and score plugins

    The Kubernetes scheduling framework transforms the complex scheduling logic into plugins and enables flexible extension of scheduling through plugins. The filter plugins are used to filter out nodes that cannot run a specific pod during pod scheduling. The score plugins assign a score to each node that has passed the filtering phase based on algorithm. The score indicates how suitable a node is for running the pod.

    The following table lists the enabled filter and score plugins and their default weights in different versions of kube-scheduler.

    Default enabled plugins

    Component version

    Filter

    Score

    v1.30.1-aliyun.6.5.4.fcac2bdf

    • Plugins that are enabled by default in Kubernetes:

      Same as those listed in Filter plugins enabled by default in Kubernetes v1.30.1.

    • Plugins that are enabled by default in Container Service for Kubernetes (ACK):

      • NodeNUMAResource

      • topologymanager

      • EciPodTopologySpread

      • ipawarescheduling

      • BatchResourceFit

      • PreferredNode

      • gpushare

      • NetworkTopology

      • CapacityScheduling

      • elasticresource

      • resourcepolicy

      • gputopology

      • ECIBinderV1

      • loadawarescheduling

      • EciScheduling

    • Plugins that are enabled by default in Kubernetes:

      Same as those listed in Score plugins enabled by default in Kubernetes v1.30.1.

    • Plugins that are enabled by default in ACK and their default weights:

      • Name: NodeNUMAResource Default Weight: 1

      • Name: ipawarescheduling Default Weight: 1

      • Name: gpuNUMAJointAllocation Default Weight: 1

      • Name: PreferredNode Default Weight: 10000

      • Name: gpushare Default Weight: 20000

      • Name: gputopology Default Weight: 1

      • Name: numa Default Weight: 1

      • Name: EciScheduling Default Weight: 2

      • Name: NodeAffinity Default Weight: 2

      • Name: elasticresource Default Weight: 1000000

      • Name: resourcepolicy Default Weight: 1000000

      • Name: NodeBEResourceLeastAllocated Default Weight: 1

      • Name: loadawarescheduling Default Weight: 10

    v1.28.3-aliyun-6.5.2.7ff57682

    • Plugins that are enabled by default in Kubernetes:

      Same as those listed in Filter plugins enabled by default in Kubernetes v1.28.3.

    • Plugins that are enabled by default in ACK:

      • NodeNUMAResource

      • topologymanager

      • EciPodTopologySpread

      • ipawarescheduling

      • BatchResourceFit

      • PreferredNode

      • gpushare

      • NetworkTopology

      • CapacityScheduling

      • elasticresource

      • resourcepolicy

      • gputopology

      • ECIBinderV1

      • loadawarescheduling

      • EciScheduling

    • Plugins that are enabled by default in Kubernetes:

      Same as those listed in Score plugins enabled by default in Kubernetes v1.28.3.

    • Plugins that are enabled by default in ACK and their default weights:

      • Name: NodeNUMAResource Default Weight: 1

      • Name: ipawarescheduling Default Weight: 1

      • Name: gpuNUMAJointAllocation Default Weight: 1

      • Name: PreferredNode Default Weight: 10000

      • Name: gpushare Default Weight: 20000

      • Name: gputopology Default Weight: 1

      • Name: numa Default Weight: 1

      • Name: EciScheduling Default Weight: 2

      • Name: NodeAffinity Default Weight: 2

      • Name: elasticresource Default Weight: 1000000

      • Name: resourcepolicy Default Weight: 1000000

      • Name: NodeBEResourceLeastAllocated Default Weight: 1

      • Name: loadawarescheduling Default Weight: 10

    v1.26.3-aliyun-6.6.1.605b8a4f

    • Plugins that are enabled by default in Kubernetes:

      Same as those listed in Filter plugins enabled by default in Kubernetes v1.26.3.

    • Plugins that are enabled by default in ACK:

      • NodeNUMAResource

      • topologymanager

      • EciPodTopologySpread

      • ipawarescheduling

      • BatchResourceFit

      • PreferredNode

      • gpushare

      • NetworkTopology

      • CapacityScheduling

      • elasticresource

      • resourcepolicy

      • gputopology

      • ECIBinderV1

      • loadawarescheduling

      • EciScheduling

    • Plugins that are enabled by default in Kubernetes:

      Same as those listed in Score plugins enabled by default in Kubernetes v1.26.3.

    • Plugins that are enabled by default in ACK:

      • Name: NodeNUMAResource Default Weight: 1

      • Name: ipawarescheduling Default Weight: 1

      • Name: gpuNUMAJointAllocation Default Weight: 1

      • Name: PreferredNode Default Weight: 10000

      • Name: gpushare Default Weight: 20000

      • Name: gputopology Default Weight: 1

      • Name: numa Default Weight: 1

      • Name: EciScheduling Default Weight: 2

      • Name: NodeAffinity Default Weight: 2

      • Name: elasticresource Default Weight: 1000000

      • Name: resourcepolicy Default Weight: 1000000

      • Name: NodeBEResourceLeastAllocated Default Weight: 1

      • Name: loadawarescheduling Default Weight: 10

    Plugin feature introduction

    Plugin descriptions and references

    Plugin name

    Description

    Reference

    NodeNUMAResource

    Manages topology-aware CPU scheduling.

    Enable CPU topology-aware scheduling

    topologymanager

    Manages the allocation of non-uniform memory access (NUMA) resources for nodes.

    Enable topology-aware NUMA scheduling

    EciPodTopologySpread

    Enhances topology spread constraints in virtual node-based pod scheduling scenarios.

    Enable pod scheduling policies on virtual nodes

    ipawarescheduling

    Schedules pods based on idle vSwitch IP addresses.

    Scheduling FAQ

    BatchResourceFit

    Enables and manages colocation.

    Best practices for colocation

    PreferredNode

    Reserves nodes for node pools with node scaling enabled.

    Node scaling

    gpushare

    Manages GPU sharing.

    GPU topology-aware scheduling

    NetworkTopology

    Manages network topology-aware scheduling.

    Topology-aware scheduling

    CapacityScheduling

    Manages the capacity scheduling feature.

    Work with capacity scheduling

    elasticresource

    Manages Elastic Container Instance-based scheduling.

    Use elastic resources to implement Elastic Container Instance-based scheduling (discontinued)

    resourcepolicy

    Manages custom elastic resource scheduling.

    Custom priority scheduling for elastic resources

    gputopology

    Manages topology-aware GPU scheduling.

    GPU topology-aware scheduling

    ECIBinderV1

    Binds virtual nodes in Elastic Container Instance-based pod scheduling scenarios.

    Schedule pods to run on ECI

    loadawarescheduling

    Manages load-aware scheduling.

    Use load-aware scheduling

    EciScheduling

    Manages virtual node-based pod scheduling.

    Enable pod scheduling policies on virtual nodes

    Usage notes

    kube-scheduler is automatically installed in a Kubernetes cluster. It requires no additional configuration. Update kube-scheduler to the latest version at the earliest opportunity to use the latest features and fix bugs. To update kube-scheduler, log on to the ACK console, click the cluster that you want to manage, then click Components and Add-ons in the left navigation.

    Release notes

    Release notes for v1.36

    Version

    Release date

    Description

    v1.36.0-apsara.6.12.3.56930cec

    2026-07-14

    • New feature:

      • Optimized the preemption workflow with task-level node filtering and parallel recovery, improving preemption efficiency in multi-task concurrent scenarios.

    • Bug fixes:

      • Fixed an occasional scheduler crash caused by concurrent execution of Gang scheduling and preemption.

      • Fixed a memory leak caused by residual cache during preemption, and a scheduling deadlock caused by expired scheduling plans.

      • Fixed scheduling delays caused by blocked ACS pod event processing.

      • Fixed an issue where Reservations marked as unschedulable were incorrectly matched by pods.

      • Fixed a scheduling deadlock in WaitForFirstConsumer (WFFC) scenarios when inventory was insufficient to complete the advanced scheduling workflow.

    v1.36.0-apsara.6.12.2.5e33fe4e

    2026-07-03

    • New features:

      • Optimized the scheduling efficiency of the reservation feature.

      • Fully supported all previous features within ACK clusters of version 1.36.

    • Bug fixes:

      • Fixed an issue where pods matching a Reservation were incorrectly counted in InterPodAffinity and TopologySpread, preventing scheduling decision deviations.

      • Fixed an issue where Reservations caused InterPodAffinity and TopologySpread to become ineffective.

    Release notes for v1.35

    Version

    Release date

    Description

    v1.35.2-apsara.6.12.2.2241959a

    2026-07-03

    • New features:

      • Optimized the scheduling efficiency of the reservation feature.

      • Fully supported all previous features within ACK clusters of version 1.35.

    • Bug fixes:

      • Fixed an issue where pods matching a Reservation were incorrectly counted in InterPodAffinity and TopologySpread, preventing scheduling decision deviations.

    Release notes for v1.34

    Version

    Release date

    Description

    v1.34.0-apsara.6.11.14.80067fc3

    2026-06-30

    • New features:

      • Added equivalent-class scheduling capability.

      • Optimized ACS inventory-aware scheduling logic under tight inventory conditions; the new version prevents prolonged pod pending states during inventory shortages by utilizing randomized dispatch.

      • Supported scheduling for image cache disks.

      • Optimized ACS+ECS scheduling performance to prevent large volumes of ACS pods from impacting ECS pod scheduling.

      • Optimized scheduler performance and reduced redundant plugin calls.

      • Added support for the agent-sandbox compute class.

      • Optimized the pod queuing algorithm to ensure that newly submitted pods can be scheduled even when a large number of pods are pending.

    • Bug fixes:

      • Fixed an issue in CapacityScheduling preemption where guaranteed resources were unable to preempt non-guaranteed resources.

      • Fixed an issue where ACS pods could be repeatedly created.

      • Fixed an issue where matchLabelKeys led to infinite pod scheduling when no pods were matched.

      • Fixed multi-container GPU card allocation to ensure precise allocation based on GPURequestCount.

      • Fixed a memory leak issue associated with ResourcePolicy + ACS.

      • Fixed occasional scheduling crashes during preemption.

      • Fixed occasional scheduling crashes with ResourcePolicy.

      • Fixed occasional high scheduling latency for pods using image cache disks.

      • Fixed an issue where CapacityScheduling unreserve caused Usage values to become negative.

      • Fixed an EciProfile matching error.

    v1.34.0-apsara.6.11.9.4a833691

    2026-02-09

    Bug fix:

    • Fixed unexpected Selector matching rule configurations in ECI profiles.

    v1.34.0-apsara.6.11.8.a32868e8

    2026-01-05

    • New features:

      • Optimized sharing GPU scheduling efficiency.

      • Introduced several new metrics for serverless resource scheduling (including processing latency, timestamp tracking, and concurrency configuration) to enhance the observability of serverless workloads.

    • Bug fixes:

      • Fixed the logic for updating scheduling allocation annotations for GPUShare pods during the Reserve phase, ensuring the correct persistence of scheduling decisions.

      • Fixed an issue where NUMA IDs failed to be removed properly.

      • Fixed an issue where resource allocation became imbalanced because NUMA allocation results could not be rebuilt from pod annotations after scheduler restarts.

      • Fixed an issue where pod NominatedNodeName was not properly cleared under specific conditions (such as out of stock or concurrent preemption).

      • Fixed an issue where resources of pods with NominatedNodeName were not reserved by Quotas when Reservation was disabled.

      • Fixed an issue where Gang preemption failures did not fail atomically as a whole, ensuring consistent behavior of task groups and avoiding redundant invalid scheduling of multi-replica tasks.

      • Fixed an issue in NetworkTopology preemption scenarios where incorrect target objects were invoked for StateData and Filter, leading to preemption failures.

      • Fixed a scheduling failure for self-built virtual kubelets equipped with GPUs.

      • Fixed a scheduling failure when multi-container requests demanded whole GPU cards.

      • Optimized ElasticQuota min/max guarantee logic to ensure that Quotas exceeding the min threshold after scheduling are only allowed to preempt within themselves.

    v1.34.0-apsara.6.11.7.43cab345

    2025-12-08

    • New features:

      • Supported Expert Parallelism (EP) size scheduling in network topology-aware scheduling, which continuously places PyTorchJobs sequentially by index during scheduling.

    • Bug fixes:

      • Optimized autoscaling efficiency.

      • Optimized ACS instance provisioning so that the scheduler no longer updates the pod scheduled condition, preventing unnecessary node pool autoscaling.

      • Fixed an issue where the scheduler could not properly read the ACS GPU partition of already-scheduled pods in the cluster upon reboot.

    v1.34.0-apsara.6.11.6.3c0b732b

    2025-11-10

    • Bug fixes:

      • Fixed the memory leak in the remaining-IP-aware scheduling logic.

      • Fixed the accounting error that may occur if a CapacityScheduling quota was updated before a pod was bound.

    • New features:

      • Supported __IGNORE__RESOURCE__.

      • Supported declaring pods as temporarily unschedulable by using the alibabacloud.com/schedule-admission annotation.

      • Supported GPU sharing in ACS.

      • Optimized persistent volume claim (PVC)-related scheduling to improve performance when creating pods with PVs.

      • Fixed incorrect ScheduleCycle updates when a ResourcePolicy and Gang scheduling were used together.

    v1.34.0-apsara.6.11.5.3c117f21

    2025-10-23

    • Bug fixes:

      • Fixed the issue where scheduling was not correctly triggered by the alibabacloud.com/acs: "true" or alibabacloud.com/eci: "true" labels.

      • Fixed the scheduling failure that occurred when multiple containers within a pod requested nvidia.com/gpu resources.

      • Fixed the potential scheduler crash that may occur under high concurrency when using ACS compute resources.

    v1.34.0-apsara.6.11.3.ff6b62d8

    2025-09-17

    Supported all features from previous versions in ACK clusters running Kubernetes 1.34.

    Release notes for v1.33

    Version

    Release date

    Description

    v1.33.0-apsara.6.11.14.0ab35bf7

    2026-06-30

    • New features:

      • Optimized ACS inventory-aware scheduling logic under tight inventory conditions; the new version prevents prolonged pod pending states during inventory shortages by utilizing randomized dispatch.

      • Supported scheduling for image cache disks.

      • Optimized scheduler performance and reduced redundant plugin calls.

      • Added support for the agent-sandbox compute class.

      • Optimized the pod queuing algorithm to ensure that newly submitted pods can be scheduled even when a large number of pods are pending.

    • Bug fixes:

      • Fixed an issue in CapacityScheduling preemption where guaranteed resources were unable to preempt non-guaranteed resources.

      • Fixed an issue where ACS pods could be repeatedly created.

      • Fixed an issue where matchLabelKeys led to infinite pod scheduling when no pods were matched.

      • Improved scheduling stability for large-scale ACS clusters.

      • Fixed an issue where CapacityScheduling unreserve caused Usage values to become negative.

      • Fixed a memory leak in CapacityScheduling preemption.

      • Fixed an issue where podResourceDetails entries were never removed from the map.

      • Fixed an EciProfile matching error.

      • Fixed a scheduling failure that occurred when acs=true was used together with a WaitForFirstConsumer PVC.

    v1.33.0-apsara.6.11.9.7c8819c5

    2026-02-09

    Bug fix:

    • Fixed unexpected Selector matching rule configurations in ECI Profiles.

    v1.33.0-apsara.6.11.8.709bb6e6

    2026-01-05

    • New features:

      • Optimized sharing GPU scheduling efficiency.

      • Introduced several new metrics for serverless resource scheduling (including processing latency, timestamp tracking, and concurrency configuration) to enhance the observability of serverless workloads.

    • Bug fixes:

      • Fixed the logic for updating scheduling allocation annotations for GPUShare pods during the Reserve phase, ensuring the correct persistence of scheduling decisions.

      • Fixed an issue where NUMA IDs failed to be removed properly.

      • Fixed an issue where resource allocation became imbalanced because NUMA allocation results could not be rebuilt from pod annotations after scheduler restarts.

      • Fixed an issue where pod NominatedNodeName was not properly cleared under specific conditions (such as out of stock or concurrent preemption).

      • Fixed an issue where resources of pods with NominatedNodeName were not reserved by Quotas when Reservation was disabled.

      • Fixed an issue where Gang preemption failures did not fail atomically as a whole, ensuring consistent behavior of task groups and avoiding redundant invalid scheduling of multi-replica tasks.

      • Fixed an issue in NetworkTopology preemption scenarios where incorrect target objects were invoked for StateData and Filter, leading to preemption failures.

      • Fixed a scheduling failure for self-built virtual kubelets equipped with GPUs.

      • Fixed a scheduling failure when multi-container requests demanded whole GPU cards.

      • Optimized ElasticQuota min/max guarantee logic to ensure that Quotas exceeding the min threshold after scheduling are only allowed to preempt within themselves.

    v1.33.0-apsara.6.11.7.4a6779f8

    2025-12-05

    • New features:

      • Supported Expert Parallelism (EP) size scheduling in network topology-aware scheduling, which continuously places PyTorchJobs sequentially by index during scheduling.

    • Bug fixes:

      • Optimized autoscaling efficiency.

      • Optimized ACS pod scheduling so that the scheduler no longer updates the pod scheduled condition, preventing unnecessary node pool autoscaling.

      • Fixed an issue where the scheduler could not properly read the ACS GPU partition of already-scheduled pods in the cluster upon reboot.

      • Fixed an issue where the presence of SelectedNode on a PVC could interfere with pod scheduling.

    v1.33.0-apsara.6.11.6.2fce98cb

    2025-11-10

    • Bug fixes:

      • Fixed the memory leak in the remaining-IP-aware scheduling logic.

      • Fixed the accounting error that may occur if a CapacityScheduling quota was updated before a pod was bound.

    • New features:

      • Supported __IGNORE__RESOURCE__.

      • Supported declaring pods as temporarily unschedulable by using the alibabacloud.com/schedule-admission annotation.

      • Supported GPU sharing in ACS.

      • Optimized PVC-related scheduling to improve performance when creating pods with PVs.

      • Fixed incorrect ScheduleCycle updates when a ResourcePolicy and Gang scheduling were used together.

    v1.33.0-apsara.6.11.5.8dd6f5f4

    2025-10-23

    Bug fix:

    • Fixed the issue where scheduling was not correctly triggered by the alibabacloud.com/acs: "true" or alibabacloud.com/eci: "true" labels.

    v1.33.0-apsara.6.11.4.77470105

    2025-09-15

    Bug fixes:

    • Fixed the scheduling failure that occurred when multiple containers within a pod requested nvidia.com/gpu resources.

    • Fixed the potential scheduler crash that may occur under high concurrency when using ACS compute resources.

    v1.33.0-apsara.6.11.3.ed953a31

    2025-09-08

    • New features:

      • Added the alibabacloud.com/ignore-empty-resource annotation in ElasticQuotaTree to ignore resource limits that are not explicitly declared in a quota.

      • Supported declaring spread distribution in JobNetworkTopology using the constraint field.

    • Bug fix:

      • Fixed the potential scheduler crash that may occur when using PodTopologySpread.

    v1.33.0-aliyun.6.11.2.330dcea7

    2025-08-19

    • Improved the scheduling determinism of GOAT. Nodes with the node.cloudprovider.kubernetes.io/uninitialized or node.kubernetes.io/unschedulable taints are no longer incorrectly treated as not ready.

    • Fixed the ElasticQuotaTree fairness check issue where quotas with a null min value or an empty internal request were incorrectly considered unsatisfied.

    • Fixed a potential scheduler crash that may occur during the creation of ACS instances.

    • The scheduler error triggered when an InitContainer was defined with empty resource requests is fixed. (29d1951)

    v1.33.0-aliyun.6.11.1.382cd0a6

    2025-07-25

    Fixed an issue where Elastic Container Instance-based scheduling through ElasticResources (a deprecated feature) was not functioning correctly.

    v1.33.0-aliyun.6.11.0.87e9673b

    2025-07-18

    • Optimized the scheduling determinism of GOAT to prevent race conditions related to NodeReady status updates during pod scheduling.

    • Fixed the incorrect pod count issue in gang scheduling. The error occurred when a PodGroup Custom Resource (CR) was deleted and recreated while some of its pods were already scheduled.

    • Fixed issues in the ElasticQuota preemption strategy. Previously, a pod could be preempted by another pod of the same priority, and intra-quota preemption may occur even when resource usage was below the min guarantee.

    • Fixed the remaining-IP-aware scheduling issue where the scheduler was not correctly preventing pods from being scheduled to nodes with insufficient available IP addresses.

    • Fixed policies TimeoutOrExceedMax and ExceedMax in ResourcePolicy, which were broken in version 6.9.x.

    • Fixed the issue in ResourcePolicy that may occur when the MaxPod count was miscalculated after an auto scaling event.

    • Added a scheduling fairness check to ElasticQuotaTree. When a quota with pending pods has not met its resource guarantee (min), new pods will not be scheduled for quotas that have already met their guarantees. This feature can be enabled by using the StrictFairness parameter and is enabled by default when the preemption algorithm is set to None.

    • Introduced a ScheduleAdmission capability. Pods with the alibabacloud.com/schedule-admission annotation will not be scheduled by the scheduler.

    • Supported pods with the labels alibabacloud.com/eci=true, alibabacloud.com/acs=true, and eci=true in the scheduler. For these pods, only a limited set of plugins (such as VolumeBind, VolumeRestrictions, VolumeZone, and virtual node plugins) will be evaluated. If such a pod has no PVC-based volumes, all scheduling checks will be skipped, and the pod will be passed directly to the virtual node handler.

    • Added a security check for ResourcePolicy scheduling. A Unit will be skipped if it could potentially patch pod labels in a way that conflicts with the matchLabels selector of a ReplicaSet or StatefulSet.

    v1.33.0-aliyun.6.9.4.8b58e6b4

    2025-06-10

    • Fixed the issue that inter-pod affinity rules and the PodTopologySpread feature may fail to take effect during consecutive pod scheduling.

    • Fixed the scheduling issues that occasionally occur when ResourcePolicies are used.

    • Optimized the kube-scheduler behavior when interacting with auto-scaling node pools.

    • Fixed the issue of wrong pod counting when ResourcePolicies are used to configure priority-based resource scheduling.

    • Fixed cloud disk leaks that may occur when you use serverless computing power in combination with cloud disks mounted in WaitForFirstConsumer.

    v1.33.0-aliyun.6.9.2.09bce458

    2025-04-28

    Supported all features provided by earlier versions in kube-scheduler 1.33.

    Release notes for v1.32

    Version

    Release date

    Description

    v1.32.0-apsara.6.11.12.0b7f0b0d

    2026-06-05

    New features:

    • Added support for the agent sandbox compute class.

    Bug fixes:

    • Improved scheduling stability for large-scale ACS clusters.

    • Fixed a nil pod panic issue.

    • Fixed an incorrect EciProfile matching issue.

    • Fixed an issue with state.SkipAllChecksExceptVB.

    • Fixed a scheduling failure that occurred when acs=true was used together with a WaitForFirstConsumer PVC.

    v1.32.0-apsara.6.11.9.08a987c9

    2026-02-09

    Bug fix:

    • Fixed unexpected Selector matching rule configurations in ECI Profiles.

    v1.32.0-apsara.6.11.8.df9f2fa6

    2026-01-05

    • New features:

      • Optimized sharing GPU scheduling efficiency.

      • Introduced several new metrics for serverless resource scheduling (including processing latency, timestamp tracking, and concurrency configuration) to enhance the observability of serverless workloads.

    • Bug fixes:

      • Fixed the logic for updating scheduling allocation annotations for GPUShare pods during the Reserve phase, ensuring the correct persistence of scheduling decisions.

      • Fixed an issue where NUMA IDs failed to be removed properly.

      • Fixed an issue where resource allocation became imbalanced because NUMA allocation results could not be rebuilt from pod annotations after scheduler restarts.

      • Fixed an issue where pod NominatedNodeName was not properly cleared under specific conditions (such as out of stock or concurrent preemption).

      • Fixed an issue where resources of pods with NominatedNodeName were not reserved by Quotas when Reservation was disabled.

      • Fixed an issue where Gang preemption failures did not fail atomically as a whole, ensuring consistent behavior of task groups and avoiding redundant invalid scheduling of multi-replica tasks.

      • Fixed an issue in NetworkTopology preemption scenarios where incorrect target objects were invoked for StateData and Filter, leading to preemption failures.

      • Fixed a scheduling failure for self-built virtual kubelets equipped with GPUs.

      • Fixed a scheduling failure when multi-container requests demanded whole GPU cards.

      • Optimized ElasticQuota min/max guarantee logic to ensure that Quotas exceeding the min threshold after scheduling are only allowed to preempt within themselves.

    v1.32.0-apsara.6.11.7.4489ebf4

    2025-12-10

    Bug fixes:

    • Optimized autoscaling efficiency.

    • Stopped the scheduler from updating the pod scheduled condition when triggering ACS pod scheduling, preventing unnecessary node pool autoscaling.

    • Fixed an issue where the scheduler could not properly read the ACS GPU partition of already-scheduled pods in the cluster upon reboot.

    v1.32.0-apsara.6.11.6.03248691

    2025-11-10

    • Bug fixes:

      • Fixed the memory leak in the remaining-IP-aware scheduling logic.

      • Fixed the accounting error that may occur if a CapacityScheduling quota was updated before a pod was bound.

    • New features:

      • Supported __IGNORE__RESOURCE__.

      • Supported declaring pods as temporarily unschedulable by using the alibabacloud.com/schedule-admission annotation.

      • Supported GPU sharing in ACS.

      • Optimized PVC-related scheduling to improve performance when creating pods with PVs.

      • Fixed incorrect ScheduleCycle updates when a ResourcePolicy and Gang scheduling were used together.

    v1.32.0-apsara.6.11.5.c774d3c3

    2025-10-23

    Bug fix:

    • Fixed the issue where scheduling was not correctly triggered by the alibabacloud.com/acs: "true" or alibabacloud.com/eci: "true" labels.

    v1.32.0-apsara.6.11.4.4a4f4843

    2025-09-15

    Bug fixes:

    • Fixed the scheduling failure that occurred when multiple containers within a pod requested nvidia.com/gpu resources.

    • Fixed the potential scheduler crash that may occur under high concurrency when using ACS compute resources.

    v1.32.0-apsara.6.11.3.b651c575

    2025-09-12

    New features:

    • Added the alibabacloud.com/ignore-empty-resource annotation in ElasticQuotaTree to ignore resource limits that are not explicitly declared in a quota.

    • Supported declaring spread distribution in JobNetworkTopology using the constraint field.

    v1.32.0-aliyun.6.11.2.58302423

    2025-08-21

    • Improved the scheduling determinism of GOAT. Nodes with the node.cloudprovider.kubernetes.io/uninitialized or node.kubernetes.io/unschedulable taints are no longer incorrectly treated as not ready.

    • Fixed the ElasticQuotaTree fairness check issue where quotas with a null min value or an empty internal request were incorrectly considered unsatisfied.

    • Fixed a potential scheduler crash that may occur during the creation of ACS instances.

    v1.32.0-aliyun.6.11.1.ab632d8c

    2025-07-25

    Fixed an issue where Elastic Container Instance-based scheduling through ElasticResources (a deprecated feature) was not functioning correctly.

    v1.32.0-aliyun.6.11.0.0350a0e7

    2025-07-18

    • Optimized the scheduling determinism of GOAT to prevent race conditions related to NodeReady status updates during pod scheduling.

    • Fixed the incorrect pod count issue in gang scheduling. The error occurred when a PodGroup CR was deleted and recreated while some of its pods were already scheduled.

    • Fixed issues in the ElasticQuota preemption strategy. Previously, a pod could be preempted by another pod of the same priority, and intra-quota preemption may occur even when resource usage was below the min guarantee.

    • Fixed the remaining-IP-aware scheduling issue where the scheduler was not correctly preventing pods from being scheduled to nodes with insufficient available IP addresses.

    • Fixed policies TimeoutOrExceedMax and ExceedMax in ResourcePolicy, which were broken in version 6.9.x.

    • Fixed the issue in ResourcePolicy that may occur when the MaxPod count was miscalculated after an auto scaling event.

    • Added a scheduling fairness check to ElasticQuotaTree. When a quota with pending pods has not met its resource guarantee (min), new pods will not be scheduled for quotas that have already met their guarantees. This feature can be enabled by using the StrictFairness parameter and is enabled by default when the preemption algorithm is set to None.

    • Introduced a ScheduleAdmission capability. Pods with the alibabacloud.com/schedule-admission annotation will not be scheduled by the scheduler.

    • Supported pods with the labels alibabacloud.com/eci=true, alibabacloud.com/acs=true, and eci=true in the scheduler. For these pods, only a limited set of plugins (such as VolumeBind, VolumeRestrictions, VolumeZone, and virtual node plugins) will be evaluated. If such a pod has no PVC-based volumes, all scheduling checks will be skipped, and the pod will be passed directly to the virtual node handler.

    • Added a security check for ResourcePolicy scheduling. A Unit will be skipped if it could potentially patch pod labels in a way that conflicts with the matchLabels selector of a ReplicaSet or StatefulSet.

    1.32.0-aliyun.6.9.4.d5a8a355

    2025-06-04

    • Fixed the issue that inter-pod affinity rules and the PodTopologySpread feature may fail to take effect during consecutive pod scheduling.

    • Fixed the scheduling issues that occasionally occur when ResourcePolicies are used.

    • Fixed the issue with the abnormal behavior of the ElasticQuota preemption feature.

    v1.32.0-aliyun.6.9.3.515ac311

    2025-05-14

    • Optimized the behavior of the Scheduler when interacting with the elastic node pool.

    • Fixed the issue with the incorrect counting of ResourcePolicy pods in custom elastic resource priority scheduling.

    • Fixed the issue of potential disk leakage when WaitForFirstConsumer-type cloud disks are used simultaneously with serverless computing resources.

    v1.32.0-aliyun.6.9.2.09bce458

    2025-04-16

    • Fixed the issue with the abnormal behavior of the ElasticQuota preemption feature.

    • Supported the scheduling of pods to ACS GPU-HPN nodes in ACK clusters.

    v1.32.0-aliyun.6.8.6.bd13955d

    2025-04-02

    Fixed the issue that WaitForFirstConsumer-type cloud disks in ACK Serverless clusters failed to be provisioned through the CSI plugin.

    v1.32.0-aliyun.6.9.0.a1c7461b

    2025-02-28

    • Supported dynamic scheduling based on available IP addresses per node.

    • Added a plugin to perform resource validation before tasks are dequeued from Kube Queue.

    • Supported switching the preemption algorithm through component configuration.

    v1.32.0-aliyun.6.8.5.28a2aed7

    2025-02-19

    • Fixed the duplicate disk creation issue that may occur in elastic container instances or ACS instances.

    • Fixed the max parameter issue in priority-based resource scheduling when PodLabels are declared.

    v1.32.0-aliyun.6.8.4.2b585931

    2025-01-17

    Supported all features from earlier versions in Kubernetes v1.32.

    Release notes for v1.31

    Version

    Release date

    Description

    v1.31.0-apsara.6.11.9.899506ac

    2026-06-22

    • New features:

      • Added support for the agent-sandbox compute class.

    • Bug fixes:

      • Fixed an issue where matchLabelKeys led to infinite pod scheduling when no pods were matched.

      • Improved scheduling stability for large-scale ACS clusters.

      • Fixed an EciProfile matching error.

      • Fixed an issue with state.SkipAllChecksExceptVB.

      • Fixed a scheduling failure that occurred when acs=true was used together with a WaitForFirstConsumer PVC.

    v1.31.0-apsara.6.11.5.19ed320c

    2026-02-09

    Bug fix:

    • Fixed unexpected Selector matching rule configurations in ECI Profiles.

    v1.31.0-apsara.6.11.5.28c6b51a

    2025-10-20

    Bug fix:

    • Fixed the issue where scheduling was not correctly triggered by the alibabacloud.com/acs: "true" or alibabacloud.com/eci: "true" labels.

    v1.31.0-apsara.6.11.4.69d7e1fa

    2025-09-15

    Bug fixes:

    • Fixed the scheduling failure that occurred when multiple containers within a pod requested nvidia.com/gpu resources.

    • Fixed the potential scheduler crash that may occur under high concurrency when using ACS compute resources.

    v1.31.0-apsara.6.11.3.9b41ad4a

    2025-09-12

    • New features:

      • Added the alibabacloud.com/ignore-empty-resource annotation in ElasticQuotaTree to ignore resource limits that are not explicitly declared in a quota.

      • Supported declaring spread distribution in JobNetworkTopology using the constraint field.

      • Improved the scheduling determinism of GOAT. Nodes with the node.cloudprovider.kubernetes.io/uninitialized or node.kubernetes.io/unschedulable taints are no longer incorrectly treated as not ready.

    • Bug fixes:

      • Fixed the ElasticQuotaTree fairness check issue where quotas with a null min value or an empty internal request were incorrectly considered unsatisfied.

      • Fixed a potential scheduler crash that may occur during the creation of ACS instances.

    v1.31.0-aliyun.6.11.1.c9ed2f40

    2025-07-25

    Fixed an issue where Elastic Container Instance-based scheduling through ElasticResources (a deprecated feature) was not functioning correctly.

    v1.31.0-aliyun.6.11.0.ea1f0f94

    2025-07-18

    • Optimized the scheduling determinism of GOAT to prevent race conditions related to NodeReady status updates during pod scheduling.

    • Fixed the incorrect pod count issue in gang scheduling. The error occurred when a PodGroup CR was deleted and recreated while some of its pods were already scheduled.

    • Fixed issues in the ElasticQuota preemption strategy. Previously, a pod could be preempted by another pod of the same priority, and intra-quota preemption may occur even when resource usage was below the min guarantee.

    • Fixed the remaining-IP-aware scheduling issue where the scheduler was not correctly preventing pods from being scheduled to nodes with insufficient available IP addresses.

    • Fixed policies TimeoutOrExceedMax and ExceedMax in ResourcePolicy, which were broken in version 6.9.x.

    • Fixed the issue in ResourcePolicy that may occur when the MaxPod count was miscalculated after an auto scaling event.

    • Added a scheduling fairness check to ElasticQuotaTree. When a quota with pending pods has not met its resource guarantee (min), new pods will not be scheduled for quotas that have already met their guarantees. This feature can be enabled by using the StrictFairness parameter and is enabled by default when the preemption algorithm is set to None.

    • Introduced a ScheduleAdmission capability. Pods with the alibabacloud.com/schedule-admission annotation will not be scheduled by the scheduler.

    • Supported pods with the labels alibabacloud.com/eci=true, alibabacloud.com/acs=true, and eci=true in the scheduler. For these pods, only a limited set of plugins (such as VolumeBind, VolumeRestrictions, VolumeZone, and virtual node plugins) will be evaluated. If such a pod has no PVC-based volumes, all scheduling checks will be skipped, and the pod will be passed directly to the virtual node handler.

    • Added a security check for ResourcePolicy scheduling. A Unit will be skipped if it could potentially patch pod labels in a way that conflicts with the matchLabels selector of a ReplicaSet or StatefulSet.

    v1.31.0-aliyun.6.9.4.c8e540e8

    2025-06-04

    • Fixed the issue that inter-pod affinity rules and the PodTopologySpread feature may fail to take effect during consecutive pod scheduling.

    • Fixed the scheduling issues that occasionally occur when ResourcePolicies are used.

    • Fixed the issue with the abnormal behavior of the ElasticQuota preemption feature.

    v1.31.0-aliyun.6.9.3.051bb0e8

    2025-05-14

    • Optimized the behavior of the Scheduler when interacting with the elastic node pool.

    • Fixed the issue with the incorrect counting of ResourcePolicy pods in custom elastic resource priority scheduling.

    • Fixed the issue of potential disk leakage when WaitForFirstConsumer-type cloud disks are used simultaneously with serverless computing resources.

    v1.31.0-aliyun.6.8.6.520f223d

    2025-04-02

    Fixed the issue that WaitForFirstConsumer-type cloud disks in ACK Serverless clusters failed to be provisioned through the CSI plugin.

    v1.31.0-aliyun.6.9.0.8287816e

    2025-02-28

    • Supported dynamic scheduling based on available IP addresses per node.

    • Added a plugin to perform resource validation before tasks are dequeued from Kube Queue.

    • Supported switching the preemption algorithm through component configuration.

    v1.31.0-aliyun.6.8.5.2c6ea085

    2025-02-19

    • Fixed the duplicate disk creation issue that may occur in elastic container instances or ACS instances.

    • Fixed the max parameter issue in priority-based resource scheduling when PodLabels are declared.

    v1.31.0-aliyun.6.8.4.8f585f26

    2025-01-02

    • Custom priority-based resource scheduling

      • Supported GPU-enabled resources in ACS.

      • Fixed the leak issue of elastic container instances that may occur when you use a PVC in an ACK Serverless cluster.

    • Capacity scheduling

      • Fixed the usage errors in ElasticQuotaTree that may occur during ACS resource adjustment.

    v1.31.0-aliyun.6.8.3.eeb86afc

    2024-12-16

    Cutom priority-based resource scheduling:

    • Added units that support multiple ACS types.

    v1.31.0-aliyun.6.8.2.eeb86afc

    2024-12-05

    Cutom priority-based resource scheduling:

    • Supported configuring PodAnnotations in the units field of the configuration file.

    v1.31.0-aliyun.6.8.1.116b8e1f

    2024-12-02

    • Improved the performance of network topology-aware scheduling.

    • Fixed the issue that Elastic Container Instance-based pods may be scheduled back to ECS nodes.

    • Updated workload-aware scheduling to no longer restrict DaemonSet pods during scheduling.

    v1.31.0-aliyun.6.7.1.1943173f

    2024-11-06

    • Custom priority-based resource scheduling

      • Supported configuring the maximum number of pods that triggers scaling. This prevents excessive node scaling by limiting the number of Pods that initiate the scaling process.

      • Deprecated resource: elastic in Unit. Instead, k8s.aliyun.com/resource-policy-wait-for-ecs-scaling in PodLabels is used.

    • Topology-aware CPU scheduling

      • Fixed the following issue: Exceptions may occur when the ECS instance type changes.

    v1.31.0-aliyun.6.7.0.740ba623

    2024-11-04

    • Capacity scheduling

      • Fixed the issue that elastic quota preemption is executed without creating an ElasticQuotaTree.

    • Custom priority-based resource scheduling

      • Added support for the ACS type.

    v1.31.0-aliyun.6.6.1.5bd14ab0

    2024-10-22

    • Fixed the Invalid Score error that occasionally occurs in PodTopologySpread.

    • Optimized event notifications for Coscheduling. The number of Coscheduling failures is included.

    • Optimized notifications for virtual node scheduling. Warning events are no longer sent during the virtual node scheduling process.

    • Network topology-aware scheduling

      • Fixed the following issue: Pods cannot be scheduled after preemption.

    • NUMA topology-aware scheduling

      • Fixed the following issue: NUMA topology-aware scheduling does not take effect.

    v1.31.0-aliyun.6.6.0.ba473715

    2024-09-13

    Supported all features provided by earlier versions in kube-scheduler V1.31.

    Release notes for v1.30

    Version

    Release date

    Description

    v1.30.3-apsara.6.11.13.0b02858b

    2026-06-22

    • New features:

      • Added support for the agent-sandbox compute class.

    • Bug fixes:

      • Fixed an issue where matchLabelKeys led to infinite pod scheduling when no pods were matched.

      • Improved scheduling stability for large-scale ACS clusters.

      • Fixed an EciProfile matching error.

      • Fixed an issue with state.SkipAllChecksExceptVB.

      • Fixed a scheduling failure that occurred when acs=true was used together with a WaitForFirstConsumer PVC.

    v1.30.3-apsara.6.11.9.1b8b27f8

    2026-02-24

    New features:

    • Added support for requesting GPU topology-aware scheduling using aliyun.com/gpu-mem.

    v1.30.3-apsara.6.11.8.c61519ca

    2026-02-09

    Bug fix:

    • Fixed unexpected Selector matching rule configurations in ECI Profiles.

    v1.30.3-apsara.6.11.7.3cfed0f9

    2025-12-10

    Bug fixes:

    • Optimized autoscaling efficiency.

    • Updated the scheduler to no longer update the pod scheduled condition when triggering ACS pod scheduling, preventing unnecessary node pool autoscaling.

    • Fixed an issue where the scheduler could not properly read the ACS GPU partition of already-scheduled pods in the cluster upon reboot.

    v1.30.3-apsara.6.11.6.a298df6b

    2025-11-10

    • New features:

      • Supported __IGNORE__RESOURCE__.

      • Supported declaring pods as temporarily unschedulable by using the alibabacloud.com/schedule-admission annotation.

      • Supported GPU sharing in ACS.

      • Optimized persistent volume claim (PVC)-related scheduling to improve performance when creating pods with PVs.

      • Fixed incorrect ScheduleCycle updates when a ResourcePolicy and Gang scheduling were used together.

      • Added the alibabacloud.com/ignore-empty-resource annotation in ElasticQuotaTree to ignore resource limits that are not explicitly declared in a quota.

      • Improved the scheduling determinism of GOAT. Nodes with the node.cloudprovider.kubernetes.io/uninitialized or node.kubernetes.io/unschedulable taints are no longer incorrectly treated as not ready.

    • Bug fixes:

      • Fixed the memory leak in the remaining-IP-aware scheduling logic.

      • Fixed the accounting error that may occur if a CapacityScheduling quota was updated before a pod was bound.

      • Fixed the ElasticQuotaTree fairness check issue where quotas with a null min value or an empty internal request were incorrectly considered unsatisfied.

    v1.30.3-apsara.6.11.3.bc707580

    2025-10-23

    Bug fix:

    • Fixed the issue where scheduling was not correctly triggered by the alibabacloud.com/acs: "true" or alibabacloud.com/eci: "true" labels.

    v1.30.3-apsara.6.11.2.463d59c9

    2025-09-15

    Bug fixes:

    • Fixed the scheduling failure that occurred when multiple containers within a pod requested nvidia.com/gpu resources.

    • Fixed the potential scheduler crash that may occur under high concurrency when using ACS compute resources.

    v1.30.3-aliyun.6.11.1.c005a0b0

    2025-07-25

    Fixed an issue where Elastic Container Instance-based scheduling through ElasticResources (a deprecated feature) was not functioning correctly.

    v1.30.3-aliyun.6.11.0.84cdcafb

    2025-07-18

    • Optimized the scheduling determinism of GOAT to prevent race conditions related to NodeReady status updates during pod scheduling.

    • Fixed the incorrect pod count issue in gang scheduling. The error occurred when a PodGroup CR was deleted and recreated while some of its pods were already scheduled.

    • Fixed issues in the ElasticQuota preemption strategy. Previously, a pod could be preempted by another pod of the same priority, and intra-quota preemption may occur even when resource usage was below the min guarantee.

    • Fixed the remaining-IP-aware scheduling issue where the scheduler was not correctly preventing pods from being scheduled to nodes with insufficient available IP addresses.

    • Fixed policies TimeoutOrExceedMax and ExceedMax in ResourcePolicy, which were broken in version 6.9.x.

    • Fixed the issue in ResourcePolicy that may occur when the MaxPod count was miscalculated after an auto scaling event.

    • Added a scheduling fairness check to ElasticQuotaTree. When a quota with pending pods has not met its resource guarantee (min), new pods will not be scheduled for quotas that have already met their guarantees. This feature can be enabled by using the StrictFairness parameter and is enabled by default when the preemption algorithm is set to None.

    • Introduced a ScheduleAdmission capability. Pods with the alibabacloud.com/schedule-admission annotation will not be scheduled by the scheduler.

    • Supported pods with the labels alibabacloud.com/eci=true, alibabacloud.com/acs=true, and eci=true in the scheduler. For these pods, only a limited set of plugins (such as VolumeBind, VolumeRestrictions, VolumeZone, and virtual node plugins) will be evaluated. If such a pod has no PVC-based volumes, all scheduling checks will be skipped, and the pod will be passed directly to the virtual node handler.

    • Added a security check for ResourcePolicy scheduling. A Unit will be skipped if it could potentially patch pod labels in a way that conflicts with the matchLabels selector of a ReplicaSet or StatefulSet.

    v1.30.3-aliyun.6.9.4.818b6506

    2025-06-04

    • Fixed the issue that inter-pod affinity rules and the PodTopologySpread feature may fail to take effect during consecutive pod scheduling.

    • Fixed the scheduling issues that occasionally occur when ResourcePolicies are used.

    • Fixed the issue with the abnormal behavior of the ElasticQuota preemption feature.

    v1.30.3-aliyun.6.9.3.ce7e2faf

    2025-05-14

    • Optimized the behavior of the Scheduler when interacting with the elastic node pool.

    • Fixed the issue with the incorrect counting of ResourcePolicy pods in custom elastic resource priority scheduling.

    • Fixed the issue of potential disk leakage when WaitForFirstConsumer-type cloud disks are used simultaneously with serverless computing resources.

    v1.30.3-aliyun.6.8.6.40d5fdf4

    2025-04-02

    Fixed the issue that WaitForFirstConsumer-type cloud disks in ACK Serverless clusters failed to be provisioned through the CSI plugin.

    v1.30.3-aliyun.6.9.0.f08e56a7

    2025-02-28

    • Supported dynamic scheduling based on available IP addresses per node.

    • Added a plugin to perform resource validation before tasks are dequeued from Kube Queue.

    • Supported switching the preemption algorithm through component configuration.

    v1.30.3-aliyun.6.8.5.af20249c

    2025-02-19

    • Fixed the duplicate disk creation issue that may occur in elastic container instances or ACS instances.

    • Fixed the max parameter issue in priority-based resource scheduling when PodLabels are declared.

    v1.30.3-aliyun.6.8.4.946f90e8

    2025-01-02

    • Custom priority-based resource scheduling

      • Supported GPU-enabled resources in ACS.

      • Fixed the leak issue of elastic container instances that may occur when using a persistent volume claim (PVC) in an ACK Serverless cluster.

    • Capacity scheduling

      • Fixed the usage errors in ElasticQuotaTree that may occur during ACS resource adjustment.

    v1.30.3-aliyun.6.8.3.697ce9b5

    2024-12-16

    Cutom priority-based resource scheduling:

    • Added units that support multiple ACS types.

    v1.30.3-aliyun.6.8.2.a5fa5dbd

    2024-12-05

    Custom priority-based resource scheduling

    • Supported configuring PodAnnotations in the units field of the configuration file.

    v1.30.3-aliyun.6.8.1.6dc0fd75

    2024-12-02

    • Improved the performance of network topology-aware scheduling.

    • Fixed the issue that Elastic Container Instance-based pods may be scheduled back to Elastic Compute Service (ECS) nodes.

    • Optimized workload-aware scheduling to no longer restrict DaemonSet pods during scheduling.

    v1.30.3-aliyun.6.7.1.d992180a

    2024-11-06

    • Custom priority-based resource scheduling

      • Supported configuring the maximum number of pods that can trigger scaling. This prevents excessive node scaling by limiting the number of Pods that initiate the scaling process.

      • Deprecated resource: elastic in Unit. Use k8s.aliyun.com/resource-policy-wait-for-ecs-scaling in PodLabels instead.

    • Topology-aware CPU scheduling

      • Fixed the following issue: Exceptions may occur when the Elastic Compute Service (ECS) instance type changes.

    v1.30.3-aliyun.6.7.0.da474ec5

    2024-11-04

    • Capacity scheduling

      • Fixed the issue that elastic quota preemption is executed without creating an ElasticQuotaTree.

    • Custom priority-based resource scheduling

      • Added support for the ACS type.

    v1.30.3-aliyun.6.6.4.b8940a30

    2024-10-22

    • Fixed the Invalid Score error that occasionally occurs in PodTopologySpread.

    v1.30.3-aliyun.6.6.3.994ade8a

    2024-10-18

    • Optimized event notifications for Coscheduling. The number of Coscheduling failures is included.

    • Optimized notifications for virtual node scheduling. Warning events are no longer sent during the virtual node scheduling process.

    v1.30.3-aliyun.6.6.2.0be67202

    2024-09-23

    • Network topology-aware scheduling

      • Fixed the following issue: Pods cannot be scheduled after preemption.

    • NUMA topology-aware scheduling

      • Fixed the following issue: NUMA topology-aware scheduling does not take effect.

    v1.30.3-aliyun.6.6.1.d98352c6

    2024-09-11

    • Supported scheduling preemptible instances in Network topology-aware scheduling.

    • SlurmOperator

      • Supported hybrid scheduling of ACK clusters with Slurm clusters.

    • Coscheduling

      • Supported the latest community edition of the custom resource definition (CRD).

    v1.30.3-aliyun.6.5.6.fe7bc1d5

    2024-08-20

    Fixed the following issue induced by v1.30.1-aliyun.6.5.1.5dad3be8: PodAffinity/PodAntiAffinity scheduling errors.

    v1.30.3-aliyun.6.5.5.8b10ee7c

    2024-08-01

    • Rebased to the open source version 1.30.3.

    v1.30.1-aliyun.6.5.5.fcac2bdf

    2024-08-01

    • CapacityScheduling

      • Fixed the following issue: The quota may be incorrectly calculated when Coscheduling and CapacityScheduling are used at the same time.

    • GPUShare

      • Fixed the following issue: Incorrect calculation of remaining resources during computing power scheduling on nodes.

    • Custom priority-based resource scheduling

      • Optimized the scale-out activity when ResourcePolicy and ClusterAutoscaler are used at the same time. Nodes are not added if all units reach their maximum number of pods.

    v1.30.1-aliyun.6.5.4.fcac2bdf

    2024-07-22

    • Coscheduling

      • Fixed the following issue: Incorrect quota statistics when you use elastic container instances.

    • Fixed the xxx is in cache, so can't be assumed error that occasionally occurs.

    v1.30.1-aliyun.6.5.3.9adaeb31

    2024-07-10

    Fixed the following issue induced by v1.30.1-aliyun.6.5.1.5dad3be8: Pods remain pending for a long period of time.

    v1.30.1-aliyun.6.5.1.5dad3be8

    2024-06-27

    • Coscheduling

      • Optimized Coscheduling to accelerate the scheduling.

    • Supported Pod scheduling in sequence.

    • Enhanced scheduling performance by declaring the equivalence class.

    • Optimized scheduler plugin performance by PreEnqueue.

    v1.30.1-aliyun.6.4.7.6643d15f

    2024-05-31

    • Supported all features in earlier versions in ACK clusters that are installed with kube-scheduler V1.30.

    Release notes for v1.28

    Version

    Release date

    Description

    v1.28.12-apsara.6.11.10.9bfc64e1

    2026-06-25

    • New features:

      • Added support for the agent-sandbox compute class.

    • Bug fixes:

      • Fixed an issue where matchLabelKeys led to infinite pod scheduling when no pods were matched.

      • Improved scheduling stability for large-scale ACS clusters.

      • Fixed an EciProfile matching error.

      • Fixed an issue with state.SkipAllChecksExceptVB.

      • Fixed a scheduling failure that occurred when acs=true was used together with a WaitForFirstConsumer PVC.

    v1.28.12-apsara-6.11.6.04aa9f94

    2026-02-09

    Bug fix:

    • Fixed unexpected Selector matching rule configurations in ECI Profiles.

    v1.28.12-apsara-6.11.5.db9be0f5

    2025-10-20

    Bug fix:

    • Fixed the issue where scheduling was not correctly triggered by the alibabacloud.com/acs: "true" or alibabacloud.com/eci: "true" labels.

    v1.28.12-apsara-6.11.4.a48c5b6c

    2025-09-15

    Bug fixes:

    • Fixed the scheduling failure that occurred when multiple containers within a pod requested nvidia.com/gpu resources.

    • Fixed the potential scheduler crash that may occur under high concurrency when using ACS compute resources.

    v1.28.12-apsara-6.11.3.1a06b13e

    2025-09-09

    New features:

    • Added the alibabacloud.com/ignore-empty-resource annotation in ElasticQuotaTree to ignore resource limits that are not explicitly declared in a quota.

    v1.28.12-aliyun-6.11.1.f23c663c

    2025-07-25

    Fixed an issue where Elastic Container Instance-based scheduling through ElasticResources (a deprecated feature) was not functioning correctly.

    v1.28.12-aliyun-6.11.0.4003ef92

    2025-07-18

    • Optimized the scheduling determinism of GOAT to prevent race conditions related to NodeReady status updates during pod scheduling.

    • Fixed the incorrect pod count issue in gang scheduling. The error occurred when a PodGroup CR was deleted and recreated while some of its pods were already scheduled.

    • Fixed issues in the ElasticQuota preemption strategy. Previously, a pod could be preempted by another pod of the same priority, and intra-quota preemption may occur even when resource usage was below the min guarantee.

    • Fixed the remaining-IP-aware scheduling issue where the scheduler was not correctly preventing pods from being scheduled to nodes with insufficient available IP addresses.

    • Fixed policies TimeoutOrExceedMax and ExceedMax in ResourcePolicy, which were broken in version 6.9.x.

    • Fixed the issue in ResourcePolicy that may occur when the MaxPod count was miscalculated after an auto scaling event.

    • Added a scheduling fairness check to ElasticQuotaTree. When a quota with pending pods has not met its resource guarantee (min), new pods will not be scheduled for quotas that have already met their guarantees. This feature can be enabled by using the StrictFairness parameter and is enabled by default when the preemption algorithm is set to None.

    • Introduced a ScheduleAdmission capability. Pods with the alibabacloud.com/schedule-admission annotation will not be scheduled by the scheduler.

    • Supported pods with the labels alibabacloud.com/eci=true, alibabacloud.com/acs=true, and eci=true in the scheduler. For these pods, only a limited set of plugins (such as VolumeBind, VolumeRestrictions, VolumeZone, and virtual node plugins) will be evaluated. If such a pod has no PVC-based volumes, all scheduling checks will be skipped, and the pod will be passed directly to the virtual node handler.

    • Added a security check for ResourcePolicy scheduling. A Unit will be skipped if it could potentially patch pod labels in a way that conflicts with the matchLabels selector of a ReplicaSet or StatefulSet.

    v1.30.3-aliyun.6.9.4.818b6506

    2025-06-04

    • Fixed the issue that inter-pod affinity rules and the PodTopologySpread feature may fail to take effect during consecutive pod scheduling.

    • Fixed the scheduling issues that occasionally occur when ResourcePolicies are used.

    • Fixed the issue with the abnormal behavior of the ElasticQuota preemption feature.

    v1.28.12-aliyun-6.9.3.cd73f3fe

    2025-05-14

    • Optimized the behavior of the Scheduler when interacting with the elastic node pool.

    • Fixed the issue with the incorrect counting of ResourcePolicy pods in custom elastic resource priority scheduling.

    • Fixed the issue of potential disk leakage when WaitForFirstConsumer-type cloud disks are used simultaneously with serverless computing resources.

    v1.28.12-aliyun-6.8.6.5f05e0ac

    2025-04-02

    Fixed the issue that WaitForFirstConsumer-type cloud disks in ACK Serverless clusters failed to be provisioned through the CSI plugin.

    v1.28.12-aliyun-6.9.0.6a13fa65

    2025-02-28

    • Supported dynamic scheduling based on available IP addresses per node.

    • Added a plugin to perform resource validation before tasks are dequeued from Kube Queue.

    • Supported switching the preemption algorithm through component configuration.

    v1.28.12-aliyun-6.8.5.b6aef0d1

    2025-02-19

    • Fixed the duplicate disk creation issue that may occur in elastic container instances or ACS instances.

    • Fixed the max parameter issue in priority-based resource scheduling when PodLabels are declared.

    v1.28.12-aliyun-6.8.4.b27c0009

    2025-01-02

    • Custom priority-based resource scheduling

      • Supported GPU-enabled resources in ACS.

      • Fixed the leak issue of elastic container instances that may occur when using a persistent volume claim (PVC) in an ACK Serverless cluster.

    • Capacity scheduling

      • Fixed the usage errors in ElasticQuotaTree that may occur during ACS resource adjustment.

    v1.28.12-aliyun-6.8.3.70c756e1

    2024-12-16

    Cutom priority-based resource scheduling:

    • Added units that support multiple ACS types.

    v1.28.12-aliyun-6.8.2.9a307479

    2024-12-05

    Cutom priority-based resource scheduling

    • Supported configuring PodAnnotations in the units field of the configuration file.

    v1.28.12-aliyun-6.8.1.db6cdeb8

    2024-12-02

    • Improved the performance of network topology-aware scheduling.

    • Fixed the issue that Elastic Container Instance-based pods may be scheduled back to Elastic Compute Service (ECS) nodes.

    • Removed the restriction on DaemonSet pods during workload-aware scheduling.

    v1.28.12-aliyun-6.7.1.44345748

    2024-11-06

    • Custom priority-based resource scheduling

      • Supported configuring the maximum number of pods that can trigger scaling. This prevents excessive node scaling by limiting the number of Pods that initiate the scaling process.

      • Deprecated resource: elastic in Unit. Use k8s.aliyun.com/resource-policy-wait-for-ecs-scaling in PodLabels instead.

    • Topology-aware CPU scheduling

      • Fixed the following issue: Exceptions may occur when the Elastic Compute Service (ECS) instance type changes.

    v1.28.12-aliyun-6.7.0.b97fca02

    2024-11-04

    • Capacity scheduling

      • Fixed the issue that elastic quota preemption is executed without creating an ElasticQuotaTree.

    • Custom priority-based resource scheduling

      • Added support for the ACS type.

    v1.28.12-aliyun-6.6.4.e535a698

    2024-10-22

    Fixed the Invalid Score error that occasionally occurs in PodTopologySpread.

    v1.28.12-aliyun-6.6.3.188f750b

    2024-10-11

    • Optimized events in Coscheduling. The number of Coscheduling failures is added to the events.

    • Optimized notifications in virtual node-based scheduling. Warning events are no longer sent during the scheduling process.

    v1.28.12-aliyun-6.6.2.054ec1f5

    2024-09-23

    • Network topology-aware scheduling

      • Fixed the following issue: Pods cannot be scheduled after preemption in the network topology-aware scheduling.

    • NUMA topology-aware scheduling

      • Fixed the following issue: NUMA topology-aware scheduling does not take effect.

    v1.28.12-aliyun-6.6.1.348b251d

    2024-09-11

    • Supported scheduling preemptible instances in Network topology-aware scheduling.

    • SlurmOperator

      • Supported hybrid scheduling of ACK clusters with Slurm clusters.

    v1.28.12-aliyun-6.5.4.79e08301

    2024-08-20

    Fixed the following issue induced by v1.30.1-aliyun.6.5.1.5dad3be8: PodAffinity/PodAntiaffinity scheduling errors.

    v1.28.12-aliyun-6.5.3.aefde017

    2024-08-01

    • Rebased to the open source version 1.28.12.

    v1.28.3-aliyun-6.5.3.79e08301

    2024-08-01

    • CapacityScheduling

      • Fixed the following issue: The quota may be incorrectly calculated when Coscheduling and CapacityScheduling are used at the same time.

    • GPUShare

      • Fixed the following issue: Incorrect calculation of remaining resources during computing power scheduling on nodes.

    • Custom priority-based resource scheduling

      • Optimized the scale-out activity when ResourcePolicy and ClusterAutoscaler are used at the same time. Nodes are not added if all units reach their maximum number of pods.

    v1.28.3-aliyun-6.5.2.7ff57682

    2024-07-22

    • Coscheduling

      • Fixed the following issue: Incorrect quota statistics when you use elastic container instances.

    • Fixed the xxx is in cache, so can't be assumed error that occasionally occurs.

    • Fixed the following issue induced by v1.30.1-aliyun.6.5.1.5dad3be8: Pods remain pending for a long period of time.

    v1.28.3-aliyun-6.5.1.364d020b

    2024-06-27

    • Coscheduling

      • Optimized Coscheduling to accelerate the scheduling.

    • Supported pod scheduling in sequence.

    • Enhanced scheduling performance by specifying the scheduling-group.

    • Optimized scheduler plugin performance by PreEnqueue.

    v1.28.3-aliyun-6.4.7.0f47500a

    2024-05-24

    • Network topology-aware scheduling

      • Fixed occasional network topology-aware scheduling failures.

    v1.28.3-aliyun-6.4.6.f32dc398

    2024-05-16

    • GPU sharing

      • Fixed the following issue: GPU scheduling exceptions occur after the value of the ack.node.gpu.schedule label on a node in an ACK Lingjun cluster is changed from egpu to default.

    • CapacityScheduling

      • Fixed the running AddPod on PreFilter plugin error that occasionally occurs.

    • Elastic scheduling

      • Supported the wait for eci provisioning event. This event is generated when you use alibabacloud.com/burst-resource to create elastic container instances.

    v1.28.3-aliyun-6.4.5.a8b4a599

    2024-05-09

    v1.28.3-aliyun-6.4.3.f57771d7

    2024-03-18

    • GPU sharing

      • Supported submitting a ConfigMap to isolate a specific GPU.

    • Custom priority-based resource scheduling

      • Supported elastic resources.

    v1.28.3-aliyun-6.4.2.25bc61fb

    2024-03-01

    Disabled the SchedulerQueueingHints feature by default. For more information, see Automated cherry pick of #122289: fix: disable SchedulerQueueingHints feature flag by default #122291.

    v1.28.3-aliyun-6.4.1.c7db7450

    2024-02-21

    • Supported NUMA co-scheduling.

    • Custom priority-based resource scheduling

      • Supported attempt wait between units.

    • Fixed the following issue: Fewer pods can be scheduled due to an incorrect number of remaining IP addresses in IP-aware pod scheduling.

    v1.28.3-aliyun-6.3.1ab2185e

    2024-01-10

    • Custom priority-based resource scheduling

      • Fixed the following issue: When you use custom priority-based resource scheduling, the affinity and spread of Elastic Container Instance-based pods across zones do not take effect.

    • Topology-aware CPU scheduling

      • Fixed the following issue: A pod fails to start on the node as the same vCPU is assigned to the pod multiple times.

    • Elastic Container Instance-based scheduling

      • Fixed the following issue: When the alibabacloud.com/burst-resource tag is used to specify a policy, the resources are still scheduled to an elastic container instance even if the value of the tag is not eci or eci_only.

    v1.28.3-aliyun-6.2.84d57ad9

    2023-12-21

    Supported MatchLabelKeys for custom priority-based resource scheduling. This way, versions are automatically grouped when an application is released.

    v1.28.3-aliyun-6.1.ac950aa0

    2023-12-13

    • CapacityScheduling

      • Supported specifying the quota. You can use quota.scheduling.alibabacloud.com/name to specify the quota of a pod.

      • Supported queue association. This feature allows you to collect statistics on resource usage of only Kube Queue-managed pods.

      • Optimized the preemption logic. In the new version, CapacityScheduling preemption does not cause the usage of preempted pods to be lower than the minimum value or higher than the maximum value.

    • Custom elastic resource priorities

      • Updated labels for updating the unit and node of ResourcePolicy. After the unit and node are updated, the Deletion-Cost of the pod is updated as well.

      • Added IgnoreTerminatingPod. Pods that are being deleted can be ignored when the number of pods in a unit is counted.

      • Added IgnorePreviousPod. Pods whose CreationTimestamp is earlier than that of the associated ResourcePolicy can be ignored when the number of pods in a unit is counted.

      • Supported PreemptPolicy to attempt pod preemption between units.

    • GPUShare

      • Accelerated the GPU sharing-based scheduling. The 99th percentile scheduling latency of the Filter plugin is reduced from milliseconds to microseconds.

    v1.28.3-aliyun-5.8-89c55520

    2023-10-28

    Supported all features provided by earlier versions in kube-scheduler V1.28.

    Release notes for v1.26

    Version

    Release date

    Description

    v1.26.3-aliyun-6.8.7.5a563072

    2025-11-27

    Fixed an issue where NUMAAwareResource returned a score greater than 100, causing scheduling failures.

    v1.26.3-aliyun-6.8.7.fec3f2bc

    2025-05-14

    Fixed the issue of potential disk leakage when WaitForFirstConsumer-type cloud disks are used simultaneously with serverless computing resources.

    v1.26.3-aliyun-6.9.0.293e663c

    2025-02-28

    • Supported dynamic scheduling based on available IP addresses per node.

    • Added a plugin to perform resource validation before tasks are dequeued from Kube Queue.

    • Supported switching the preemption algorithm through component configuration.

    v1.26.3-aliyun-6.8.5.7838feba

    2025-02-19

    • Fixed the duplicate disk creation issue that may occur in elastic container instances or ACS instances.

    • Fixed the max parameter issue in priority-based resource scheduling when PodLabels are declared.

    v1.26.3-aliyun-6.8.4.4b180111

    2025-01-02

    • Custom priority-based resource scheduling

      • Supported GPU-enabled resources in ACS.

      • Fixed the leak issue of elastic container instances that may occur when using a persistent volume claim (PVC) in an ACK Serverless cluster.

    • Capacity scheduling

      • Fixed the usage errors in ElasticQuotaTree that may occur during ACS resource adjustment.

    v1.26.3-aliyun-6.8.3.95c73e0b

    2024-12-16

    Cutom priority-based resource scheduling:

    • Added units that support multiple ACS types.

    v1.26.3-aliyun-6.8.2.9c9fa19f

    2024-12-05

    Cutom priority-based resource scheduling

    • Supported configuring PodAnnotations in the units field of the configuration file.

    v1.26.3-aliyun-6.8.1.a12db674

    2024-12-02

    • Fixed the issue that Elastic Container Instance-based pods may be scheduled back to Elastic Compute Service (ECS) nodes.

    • Removed the restriction on DaemonSet pods during scheduling in workload-aware scheduling.

    v1.26.3-aliyun-6.7.1.d466c692

    2024-11-06

    • Custom priority-based resource scheduling

      • Supported configuring the maximum number of pods that can trigger scaling. This prevents excessive node scaling by limiting the number of Pods that initiate the scaling process.

      • Deprecated resource: elastic in Unit. Use k8s.aliyun.com/resource-policy-wait-for-ecs-scaling in PodLabels instead.

    • Topology-aware CPU scheduling

      • Fixed the following issue: Exceptions may occur when the Elastic Compute Service (ECS) instance type changes.

    v1.26.3-aliyun-6.7.0.9c293fb7

    2024-11-04

    • Capacity scheduling

      • Fixed the issue that elastic quota preemption is executed without creating an ElasticQuotaTree.

    • Custom priority-based resource scheduling

      • Added support for the ACS type.

    v1.26.3-aliyun-6.6.4.7a8f3f9d

    2024-10-22

    Optimized notifications for virtual node scheduling. Warning events are no longer sent during the virtual node scheduling process.

    v1.26.3-aliyun-6.6.3.67f250fe

    2024-09-04

    • SlurmOperator

      • Optimized the scheduling performance of the plugins.

    v1.26.3-aliyun-6.6.2.9ea0a6f5

    2024-08-30

    • InterPodAffinity

      • Fixed the following issue: Pods rescheduling is not triggered when you remove taints from newly created nodes.

    v1.26.3-aliyun-6.6.1.605b8a4f

    2024-07-31

    • SlurmOperator

      • Supported hybrid scheduling with Kubernetes clusters and Slurm clusters.

    • Custom priority-based resource scheduling

      • Optimized the feature to avoid unnecessary nodes increase when used with a node pool that has auto scaling enabled.

    v1.26.3-aliyun-6.4.7.2a77d106

    2024-06-27

    • Coscheduling

      • Optimized Coscheduling to accelerate the scheduling.

    v1.26.3-aliyun-6.4.6.78cacfb4

    2024-05-16

    • CapacityScheduling

      • Fixed the running AddPod on PreFilter plugin error that occasionally occurs.

    • Elastic scheduling

      • Supported the wait for eci provisioning event. This event is generated when you use alibabacloud.com/burst-resource to create elastic container instances.

    v1.26.3-aliyun-6.4.5.7f36e9b3

    2024-05-09

    v1.26.3-aliyun-6.4.3.e7de0a1e

    2024-03-18

    • GPU sharing

      • Supported submitting a ConfigMap to isolate a specific GPU.

    • Custom priority-based resource scheduling

      • Supported elastic resources.

    v1.26.3-aliyun-6.4.1.d24bc3c3

    2024-02-21

    • Optimized the scoring of the NodeResourceFit plugin for Virtual Node. Currently, the NodeResourceFit plugin always gives Virtual Node a score of 0. This way, Preferred NodeAffinity can preferentially schedule ECS nodes.

    • Supported NUMA co-scheduling.

    • Custom priority-based resource scheduling

      • Supported attempt wait between units.

    • Fixed the following issue: Fewer pods can be scheduled due to an incorrect number of remaining IP addresses in IP-aware pod scheduling.

    v1.26.3-aliyun-6.3.33fdc082

    2024-01-10

    • Custom elastic resource priorities

      • Fixed the following issue: When you use custom priority-based resource scheduling, the affinity and spread of Elastic Container Instance-based pods across zones do not take effect.

    • Topology-aware CPU scheduling

      • Fixed the following issue: A pod fails to start on the node as the same vCPU is assigned to the pod multiple times.

    • Elastic Container Instance-based scheduling

      • Fixed the following issue: When the alibabacloud.com/burst-resource tag is used to specify a policy, the resources are still scheduled to an elastic container instance even if the value of the tag is not eci or eci_only.

    • CapacityScheduling

      • Supported automatic task preemption in ACK Lingjun clusters.

    v1.26.3-aliyun-6.2.d9c15270

    2023-12-21

    Supported MatchLabelKeys for custom priority-based resource scheduling. This way, versions are automatically grouped when an application is released.

    v1.26.3-aliyun-6.1.a40b0eef

    2023-12-13

    • CapacityScheduling

      • Supported specifying the quota. You can use quota.scheduling.alibabacloud.com/name to specify the quota of a pod.

      • Supported queue association. This feature allows you to collect statistics on resource usage of only Kube Queue-managed pods.

      • Optimized the preemption logic. In the new version, CapacityScheduling preemption does not cause the usage of preempted pods to be lower than the minimum value or higher than the maximum value.

    • Custom elastic resource priorities

      • Updated labels for updating the unit and node of ResourcePolicy. After the unit and node are updated, the Deletion-Cost of the pod is updated as well.

      • Added IgnoreTerminatingPod. Pods that are being deleted can be ignored when the number of pods in a unit is counted.

      • Added IgnorePreviousPod. Pods whose CreationTimestamp is earlier than that of the associated ResourcePolicy can be ignored when the number of pods in a unit is counted.

      • Supported PreemptPolicy to attempt pod preemption between units.

    • GPUShare

      • Accelerated the GPU sharing-based scheduling. The 99th percentile scheduling latency of the Filter plugin is reduced from milliseconds to microseconds.

    v1.26.3-aliyun-5.9-cd4f2cc3

    2023-11-16

    • Optimized the display of the cause of the scheduling failure due to the invalid disk type.

    v1.26.3-aliyun-5.8-a1482f93

    2023-10-16

    • Supported scheduling pods to Windows nodes.

    • Optimized Coscheduling to accelerate the scheduling of multiple concurrent tasks and reduce blocked tasks.

    v1.26.3-aliyun-5.7-2f57d3ff

    2023-09-20

    • Fixed the following issue: When GPU sharing is used to schedule pods, kube-scheduler occasionally fails to admit pods.

    • Added a plugin to kube-scheduler to detect the available IP addresses on a node. If no IP addresses are available on the node, pods are no longer scheduled to the node.

    • Added a topology-aware scheduling plugin to kube-scheduler. This plugin can schedule pods to the same topological domain and automatically retries scheduling on multiple topological domains.

    • kube-scheduler updates the usage and request information about ElasticQuotaTree every one second.

    v1.26.3-aliyun-5.5-8b98a1cc

    2023-07-05

    • Fixed the following issue: Pods occasionally remain in the Pending state for a long time when Coscheduling is used.

    • Optimized user experience for using Coscheduling and elastic node pools at the same time. If some pods in a pod group cannot be scheduled or created due to incorrect node selector configurations, the other pods in the PodGroup will not trigger scale-out activities.

    v1.26.3-aliyun-5.4-21b4da4c

    2023-07-03

    • Fixed the issue the max parameter in a ResourcePolicy does not take effect.

    • Mitigated the impact of a large number of pending pods on the kube-scheduler performance. This update brings the throughput of kube-scheduler up to a level close to that when the cluster does not contain pending pods.

    v1.26.3-aliyun-5.1-58a821bf

    2023-05-26

    Supported updating fields such as min-available and Matchpolicy for PodGroups.

    v1.26.3-aliyun-5.0-7b1ccc9d

    2023-05-22

    • Supported specifying the maximum number of replicated pods in the Unit field when you configure priority-based resource scheduling.

    • Supported topology-aware GPU scheduling.

    v1.26.3-aliyun-4.1-a520c096

    2023-04-27

    Prevented cluster-autoscaler from adding nodes when the elastic quota is exhausted or when pods are insufficient for gang scheduling.

    Release notes for v1.24

    Version

    Release date

    Description

    v1.24.6-aliyun-6.4.7.e7ffcda5

    2025-05-06

    • Fixed the issue with the occasional Max count error in ResourcePolicy.

    • Fixed the issue of potential disk leakage when WaitForFirstConsumer-type cloud disks are used simultaneously with serverless computing resources.

    v1.24.6-aliyun-6.5.0.37a567db (Whitelist enabled)

    2024-11-04

    Custom priority-based resource scheduling

    • Added units that support the ACS type.

    v1.24.6-aliyun-6.4.6.c4d551a0

    2024-05-16

    • CapacityScheduling

      • Fixed the running AddPod on PreFilter plugin error that occasionally occurs.

    v1.24.6-aliyun-6.4.5.aab44b4a

    2024-05-09

    v1.24.6-aliyun-6.4.3.742bd819

    2024-03-18

    • GPU sharing

      • Supported submitting a ConfigMap to isolate a specific GPU.

    • Custom priority-based resource scheduling

      • Supported elastic resources.

    v1.24.6-aliyun-6.4.1.14ebc575

    2024-02-21

    • Optimized the scoring of the NodeResourceFit plugin for Virtual Node. Currently, the NodeResourceFit plugin always gives Virtual Node a score of 0. This way, Preferred NodeAffinity can preferentially schedule ECS nodes.

    • Supported NUMA Coscheduling.

    • Custom priority-based resource scheduling

      • Supported attempt wait between units.

    • Fixed the following issue: Fewer pods can be scheduled due to an incorrect number of remaining IP addresses in IP-aware pod scheduling.

    v1.24.6-aliyun-6.3.548a9e59

    2024-01-10

    • Custom priority-based resource scheduling

      • Fixed the following issue: When you use custom priority-based resource scheduling, the affinity and spread of Elastic Container Instance-based pods across zones do not take effect.

    • Topology-aware CPU scheduling

      • Fixed the following issue: A pod fails to start on the node as the same vCPU is assigned to the pod multiple times.

    • Elastic Container Instance-based scheduling

      • Fixed the following issue: When the alibabacloud.com/burst-resource tag is used to specify a policy, the resources are still scheduled to an elastic container instance even if the value of the tag is not eci or eci_only.

    • CapacityScheduling

      • Supported automatic task preemption in ACK Lingjun clusters.

    v1.24.6-aliyun-6.2.0196baec

    2023-12-21

    Supported MatchLabelKeys for custom priority-based resource scheduling. This way, versions are automatically grouped when an application is released.

    v1.24.6-aliyun-6.1.1900da95

    2023-12-13

    • CapacityScheduling

      • Supported specifying the quota. You can use quota.scheduling.alibabacloud.com/name to specify the quota of a pod.

      • Supported queue association. This feature allows you to collect statistics on resource usage of only Kube Queue-managed pods.

      • Optimized the preemption logic. In the new version, CapacityScheduling preemption does not cause the usage of preempted pods to be lower than the minimum value or higher than the maximum value.

    • Custom elastic resource priorities

      • Updated the labels for updating the unit and node of ResourcePolicy. After the unit and node are updated, the Deletion-Cost of the pod is updated as well.

      • Added IgnoreTerminatingPod. Pods that are being deleted can be ignored when the number of pods in a unit is counted.

      • Added IgnorePreviousPod. Pods whose CreationTimestamp is earlier than that of the associated ResourcePolicy can be ignored when the number of pods in a unit is counted.

      • Supported PreemptPolicy to attempt pod preemption between units.

    • GPUShare

      • Accelerated GPU sharing-based scheduling. The 99th percentile scheduling latency of the Filter plugin is reduced from milliseconds to microseconds.

    v1.24.6-aliyun-5.9-e777ab5b

    2023-11-16

    • Optimized the display of the cause of the scheduling failure due to the invalid disk type.

    v1.24.6-aliyun-5.8-49fd8652

    2023-10-16

    • Supported scheduling pods to Windows nodes.

    • Optimized Coscheduling to accelerate the scheduling of multiple concurrent tasks and reduce blocked tasks.

    v1.24.6-aliyun-5.7-62c7302c

    2023-09-20

    • Fixed the following issue: When GPU sharing is used to schedule pods, kube-scheduler occasionally fails to admit pods.

    v1.24.6-aliyun-5.6-2bb99440

    2023-08-31

    • Added a plugin to kube-scheduler to detect the available IP addresses on a node. If no IP addresses are available on the node, pods are no longer scheduled to the node.

    • Added a topology-aware scheduling plugin to kube-scheduler. This plugin can schedule pods to the same topological domain and automatically retries scheduling on multiple topological domains.

    • kube-scheduler updates the usage and request information about ElasticQuotaTree every one second.

    v1.24.6-aliyun-5.5-5e8aac79

    2023-07-05

    • Fixed the following issue: Pods occasionally remain in the Pending state for a long time when Coscheduling is used.

    • Optimized user experience for using Coscheduling and elastic node pools at the same time. If some pods in a pod group cannot be scheduled or created due to incorrect node selector configurations, the other pods in the PodGroup will not trigger scale-out activities.

    v1.24.6-aliyun-5.4-d81e785e

    2023-07-03

    • Fixed the issue that the max parameter in a ResourcePolicy does not take effect.

    • Mitigated the impact of a large number of pending pods on the kube-scheduler performance. This update brings the throughput of kube-scheduler up to a level close to that when the cluster does not contain pending pods.

    v1.24.6-aliyun-5.1-95d8a601

    2023-05-26

    Supported updating fields such as min-available and Matchpolicy for Coscheduling.

    v1.24.6-aliyun-5.0-66224258

    2023-05-22

    • Supported specifying the maximum number of replicated pods in the Unit field when you configure priority-based resource scheduling.

    • Supported topology-aware GPU scheduling.

    v1.24.6-aliyun-4.1-18d8d243

    2023-03-31

    Supported using elastic resources to schedule pods to ARM-based virtual nodes.

    v1.24.6-4.0-330eb8b4-aliyun

    2023-03-01

    • GPU sharing:

      • Fixed the kube-scheduler status error during GPU-accelerated node downgrades.

      • Fixed the issue that you cannot allocate all GPU memory of a GPU-accelerated node.

      • Supported preempting pods on GPU-accelerated nodes.

    • Coscheduling:

      • Supported claiming gangs by using PodGroups or calling the Koordinator API.

      • Supported controlling gang scheduling retries by claiming a matchpolicy.

      • Supported gang groups.

      • The naming of gangs must meet the rules for DNS subdomains.

    • Added custom parameters to support Loadaware-related configurations.

    v1.24.6-3.2-4f45222b-aliyun

    2023-01-13

    Fixed the issue that the GPU memory used by pods cannot be displayed as normal due to the incorrect shared GPU memory information.

    v1.24.6-ack-3.1

    2022-11-14

    • Enabled the score feature for GPU sharing by default. The score feature is disabled by default in earlier versions.

    • Supported load-aware scheduling.

    v1.24.6-ack-3.0

    2022-09-27

    Supported capacity scheduling.

    v1.24.3-ack-2.0

    2022-09-21

    • Supported GPU sharing.

    • Supported Coscheduling.

    • Supported Elastic Container Instance-based scheduling.

    • Supported intelligent CPU scheduling.

    Release notes for v1.22

    Version

    Release date

    Description

    v1.22.15-aliyun-6.4.5.e54fd757

    2025-05-06

    • Fixed the issue with the occasional Max count error in ResourcePolicy.

    • Fixed the issue of potential disk leakage when WaitForFirstConsumer-type cloud disks are used simultaneously with serverless computing resources.

    v1.22.15-aliyun-6.4.4.7fc564f8

    2024-05-16

    • CapacityScheduling

      • Fixed the running AddPod on PreFilter plugin error that occasionally occurs.

    v1.22.15-aliyun-6.4.3.e858447b

    2024-04-22

    • Custom priority-based resource scheduling

      • Fixed the following issue: When you delete a ResourcePolicy, the status is occasionally abnormal.

    v1.22.15-aliyun-6.4.2.4e00a021

    2024-03-18

    • CapacityScheduling

      • Fixed occasional preemption failures in ACK Lingjun clusters.

    • Supported adding GPU cards in a cluster to the blacklist by using ConfigMaps.

    v1.22.15-aliyun-6.4.1.1205db85

    2024-02-29

    • Custom priority-based resource scheduling

      • Fixed occasional concurrency conflicts.

    v1.22.15-aliyun-6.4.0.145bb899

    2024-02-28

    • CapacityScheduling

      • Fixed incorrect quota statistics due to specific quota features.

    v1.22.15-aliyun-6.3.a669ec6f

    2024-01-10

    • Custom priority-based resource scheduling

      • Fixed the following issue: When you use custom priority-based resource scheduling, the affinity and spread of Elastic Container Instance-based pods across zones do not take effect.

      • Supported MatchLabelKeys.

    • Topology-aware CPU scheduling

      • Fixed the following issue: The same vCPU is assigned to a pod multiple times. Therefore, the pod fails to start on the node.

    • Elastic Container Instance-based scheduling

      • Fixed the following issue: When the alibabacloud.com/burst-resource tag is used to specify a policy, the resources are still scheduled to an elastic container instance even if the value of the tag is not eci or eci_only.

    • CapacityScheduling

      • Supported automatic task preemption in ACK Lingjun clusters.

    v1.22.15-aliyun-6.1.e5bf8b06

    2023-12-13

    • CapacityScheduling

      • Supported specifying the quota. You can use quota.scheduling.alibabacloud.com/name to specify the quota of a pod.

      • Supported queue association. This feature allows you to collect statistics on only resources of Kube Queue-managed pods.

      • Optimized the preemption logic. In the new version, CapacityScheduling preemption does not cause the usage of preempted pods to be lower than the minimum value or higher than the maximum value.

    • Custom elastic resource priorities

      • Updated labels for updating the unit and node of ResourcePolicy. After the unit and node are updated, the Deletion-Cost of the pod is updated as well.

      • Added IgnoreTerminatingPod. Pods that are being deleted can be ignored when the number of pods in a unit is counted.

      • Added IgnorePreviousPod. Pods whose CreationTimestamp is earlier than that of the associated ResourcePolicy can be ignored when the number of pods in a unit is counted.

      • Supported PreemptPolicy to attempt pod preemption between units.

    • GPUShare

      • Accelerated GPU sharing-based scheduling. The 99th percentile scheduling latency of the Filter plugin is reduced from milliseconds to microseconds.

    v1.22.15-aliyun-5.9-04a5e6eb

    2023-11-16

    • Optimized the display of the cause of the scheduling failure due to the invalid disk type.

    v1.22.15-aliyun-5.8-29a640ae

    2023-10-16

    • Supported scheduling pods to Windows nodes.

    • Optimized Coscheduling to accelerate the scheduling of multiple concurrent tasks and reduce blocked tasks.

    v1.22.15-aliyun-5.7-bfcffe21

    2023-09-20

    • Fixed the following issue: When GPU sharing is used to schedule pods, kube-scheduler occasionally fails to admit pods.

    v1.22.15-aliyun-5.6-6682b487

    2023-08-14

    • Added a plugin to kube-scheduler to detect the available IP addresses on a node. If no IP addresses are available on the node, pods are no longer scheduled to the node.

    • Added a topology-aware scheduling plugin to kube-scheduler. This plugin can schedule pods to the same topological domain and automatically retries scheduling on multiple topological domains.

    • kube-scheduler updates the usage and request information about ElasticQuotaTree every one second.

    v1.22.15-aliyun-5.5-82f32f68

    2023-07-05

    • Fixed the following issue: Pods occasionally remain in the Pending state for a long time when Coscheduling is used.

    • Optimized the user experience of PodGroups in elastic node pools. If some pods in a pod group cannot be scheduled or created due to incorrect node selector configurations, the other pods in the PodGroup will not trigger scale-out activities.

    v1.22.15-aliyun-5.4-3b914a05

    2023-07-03

    • Fixed the issue the max parameter in a ResourcePolicy does not take effect.

    • Mitigated the impact of a large number of pending pods on the kube-scheduler performance. This update brings the throughput of kube-scheduler up to a level close to that when the cluster does not contain pending pods.

    v1.22.15-aliyun-5.1-8a479926

    2023-05-26

    Supported updating fields such as min-available and Matchpolicy for PodGroups.

    v1.22.15-aliyun-5.0-d1ab67d9

    2023-05-22

    • Supported specifying the maximum number of replicated pods in the Unit field when you configure priority-based resource scheduling.

    • Supported topology-aware GPU scheduling.

    v1.22.15-aliyun-4.1-aec17f35

    2023-03-31

    Supported using elastic resources to schedule pods to ARM-based virtual nodes.

    v1.22.15-aliyun-4.0-384ca5d5

    2023-03-03

    • GPU sharing:

      • Fixed the kube-scheduler status error during GPU-accelerated node downgrades.

      • Fixed the issue that you cannot allocate all GPU memory of a GPU-accelerated node.

      • Supported preempting pods on GPU-accelerated nodes.

    • Coscheduling:

      • Supported claiming gangs by using PodGroups or calling the Koordinator API.

      • Supported controlling gang scheduling retries by claiming a matchpolicy.

      • Supported gang groups.

      • The naming of gangs must meet the rules for DNS subdomains.

    • Added custom parameters to support Loadaware-related configurations.

    v1.22.15-2.1-a0512525-aliyun

    2023-01-10

    Fixed the issue that the GPU memory used by pods cannot be displayed as normal due to the incorrect shared GPU memory information.

    v1.22.15-ack-2.0

    2022-11-30

    • Supported custom parameter settings.

    • Supported load-aware scheduling.

    • Supported priority-based scheduling. You can use this feature to schedule pods to node pools based on priorities.

    • Supported sharing the computing power of GPUs.

    v1.22.3-ack-1.1

    2022-02-27

    Fixed the issue that GPU sharing and scheduling do not work when the cluster contains only one node.

    v1.22.3-ack-1.0

    2021-01-04

    • Supported intelligent CPU scheduling.

    • Supported Coscheduling.

    • Supported capacity scheduling.

    • Supported Elastic Container Instance-based scheduling.

    • Supported GPU sharing.

    Release notes for v1.20

    Version

    Release date

    Description

    v1.20.11-aliyun-10.6-f95f7336

    2023-09-22

    • Fixed occasional incorrect statistics on quota usage in ElasticQuotaTree.

    v1.20.11-aliyun-10.3-416caa03

    2023-05-26

    • Fixed the cache error that occasionally occurs during GPU sharing in earlier Kubernetes versions.

    v1.20.11-aliyun-10.2-f4a371d3

    2023-04-27

    • Supported using elastic resources to schedule pods to ARM-based virtual nodes.

    • Fixed the issue that load-aware scheduling does not work as expected when the CPU usage of pods on a node exceeds the CPU request of the pods.

    v1.20.11-aliyun-10.0-ae867721

    2023-04-03

    Added support for the Matchpolicy field in Coscheduling.

    v1.20.11-aliyun-9.2-a8f8c908

    2023-03-08

    • CapacityScheduling: Fixed the kube-scheduler status error caused by quotas with the same name.

    • Supported cloud disk scheduling.

    • GPU sharing and scheduling:

      • Fixed the kube-scheduler status error during GPU-accelerated node downgrades.

      • Fixed the occasional issue that you cannot allocate all GPU memory of a GPU-accelerated node.

      • Supported preempting pods on GPU-accelerated nodes.

    • Topology-aware CPU scheduling: Pods that have CPU scheduling enabled are not scheduled to nodes that have NUMA disabled.

    • Added custom parameters.

    v1.20.4-ack-8.0

    2022-08-29

    Fixed bugs.

    v1.20.4-ack-7.0

    2022-02-22

    Supported priority-based scheduling. You can use this feature to schedule pods to node pools based on priorities.

    v1.20.4-ack-4.0

    2021-09-02

    • Supported load-aware scheduling.

    • Supported Elastic Container Instance-based scheduling.

    v1.20.4-ack-3.0

    2021-05-26

    Supported intelligent CPU scheduling based on sockets and L3 cache (last level cache).

    v1.20.4-ack-2.0

    2021-05-14

    Supported capacity scheduling.

    v1.20.4-ack-1.0

    2021-04-07

    • Supported intelligent CPU scheduling.

    • Supported Coscheduling.

    • Supported topology-aware GPU scheduling.

    • Supported GPU sharing.

    Release notes for v1.18

    Version

    Release date

    Description

    v1.18-ack-4.0

    2021-09-02

    Supported load-aware scheduling.

    v1.18-ack-3.1

    2021-06-05

    Supported node pools for Elastic Container Instance-based scheduling.

    v1.18-ack-3.0

    2021-03-12

    Supported scheduling based on both Elastic Container Instance and ECS.

    v1.18-ack-2.0

    2020-11-30

    Supported topology-aware GPU scheduling and GPU sharing.

    v1.18-ack-1.0

    2020-09-24

    Supported intelligent CPU scheduling and Coscheduling.

    Release notes for v1.16

    Version

    Release date

    Description

    v1.16-ack-1.0

    2020-07-21

    • Supported intelligent CPU scheduling in clusters that run Kubernetes 1.16.

    • Supported Coscheduling in clusters that run Kubernetes 1.16.