Table of Contents generated with DocToc
Download v1.10.9 in the v1.10.9 release page.
karmada-webhook
: Fixed panic when validating ResourceInterpreterWebhookConfiguration with unspecified service port. (#5965, @seanlaii)karmada-controller-manager
: Fixed the issue of missing work queue metrics. (#5983, @XiShanYongYe-Chang)karmada-controller-manager
: Fixed the bug of WorkloadRebalancer doesn't get deleted after TTL. (#5993, @deefreak)
None.
Download v1.10.8 in the v1.10.8 release page.
- The feature
Failover
now has been disabled by default, which should be explicitly enabled to avoid unexpected incidents. (#5947, @XiShanYongYe-Chang)
If you are using the feature Failover
, please enable it explicitly by adding the --feature-gates=Failover=true,<other feature>
flag to the karmada-controller-manager
component. If you are not using this feature, this change will have no impact.
karmadactl
: Fixedkarmada-metrics-adapter
use the incorrect certificate issue when deployed via karmadactlinit
. (#5860, @seanlaii)karmada-controller-manager
: Fixed the corner case where the reconciliation of aggregating status might be missed in case of component restart. (#5883, @liangyuanpeng)karmada-controller-manager
: Fixed the problem of ResourceBinding remaining after the resource template is deleted in the dependencies distribution scenario. (#5952, @XiShanYongYe-Chang)karmada-scheduler
: Avoid filtering out clusters if the API enablement is incomplete during re-scheduling. (#5931, @XiShanYongYe-Chang)
- The base image
alpine
now has been promoted from3.20.3
to3.21.0
. (#5921)
Download v1.10.7 in the v1.10.7 release page.
karmada-controller-manager
: Fixed the issue that status aggregation against the resource template might be missed due to slow cache sync. (#5844, @chaosi-zju)karmadactl
: The--force
option ofunjoin
command now try to clean up resources propagated in member clusters. (#5845, @chaosi-zju)
None.
Download v1.10.6 in the v1.10.6 release page.
karmada-aggregated-apiserver
: User can append a "/" at the end when configuring the cluster's apiEndpoint. (#5556, @spiritNO1)karmada-controller-manager
: Ignored StatefulSet Dependencies with PVCs created via the VolumeClaimTemplates. (#5687, @seanlaii)karmada-scheduler
: Fixed unexpected modification of originalResourceSummary
due to lack of deep copy. (#5725, @RainbowMango)karmada-scheduler
: Fixes an issue where resource model grades were incorrectly matched based on resource requests. Now only grades that can provide sufficient resources will be selected. (#5729, @RainbowMango)karmada-search
: Modify the logic of checking whether the resource is registered when selecting the plugin. (#5736, @seanlaii)
None.
Download v1.10.5 in the v1.10.5 release page.
karmada-controller-manager
: Fixed the error of cluster status old condition update will overwrite the newest condition. (#5401, @XiShanYongYe-Chang)
- The base image
alpine
now has been promoted fromalpine:3.20.2
toalpine:3.20.3
.
Download v1.10.4 in the v1.10.4 release page.
Helm
: fix wrongClusterResourceBinding
scope inMutatingWebhookConfiguration
. (#5262, @XiShanYongYe-Chang)
- The base image
alpine
now has been promoted fromalpine:3.20.1
toalpine:3.20.2
. (#5268) - Bump golang version to
v1.21.13
(#5371, @zhzhuang-zju)
Download v1.10.3 in the v1.10.3 release page.
karmada-controller-manager
: fix the issue of residual work in the MultiClusterService feature. (#5211, @XiShanYongYe-Chang)
karmada-scheduler
: GroupClusters will sort clusters by score and availableReplica count. (#5180, @mszacillo)
Download v1.10.2 in the v1.10.2 release page.
karmada-controller-manager
: Fixed the issue that the default resource interpreter doesn't accurately interpret the numbers of replicas. (#5108, @whitewindmills)
- The base image
alpine
now has been promoted fromalpine:3.20.0
toalpine:3.20.1
. (#5093)
Download v1.10.1 in the v1.10.1 release page.
karmada-scheduler-estimator
: Fixed theUnschedulable
result returned by plugins to be treated as an exception issue. (#5027, @RainbowMango)karmada-controller-manager
: Fixed an issue that the cluster-status-controller overwrites the remedyActions field. (#5043, @XiShanYongYe-Chang)
None.
Download v1.10.0 in the v1.10.0 release page.
This release introduced a workload rebalancing capability. It can actively trigger a brand fresh rescheduling to establish an entirely new replicas distribution across clusters.
In some scenarios, the current distribution of replicas might not always be ideal, such as:
-
Replicas were migrated due to cluster failover, but the cluster has now recovered.
-
Replicas were migrated due to application-level failover, but each cluster now has sufficient resources to run the replicas.
-
For the
Aggregated
scheduling strategy, replicas were initially distributed across multiple clusters due to resource constraints, but now a single cluster is sufficient to accommodate all replicas.
With the workload rebalancing capability, users can trigger a workload rebalancing on demand if the current replicas distribution is not optimal.
For a detailed description of this feature, see the User Guide, and for a specific demonstration, see the Tutorial.
(Feature contributors: @chaosi-zju)
Due to historical design reasons, the name of the resource template will be used as the name of the label, thereby accelerating the retrieval of resources. Since Kubernetes limits the label to no more than 63 characters, this indirectly restricts the length of the resource template, seriously preventing users from migrating workload from legacy cluster to multiple clusters.
The work that got rid of this restriction started from release 1.8, we did sufficient preparatory work in both releases 1.8 and release 1.9 to ensure that users using the old version of Karmada can smoothly upgrade to the new version.
See [Umbrella] Use permanent-id to replace namespace/name labels in the resource for more details.
(Feature contributors: @liangyuanpeng, @whitewindmills, @XiShanYongYe-Chang)
- Introduced
ServiceAnnotations
to theKarmada
API to provide an extra set of annotations to annotate karmada apiserver services. (#4679, @calvin0327) - Add a short name for resourceinterpretercustomizations CRD resource. (#4872, @XiShanYongYe-Chang)
- Introduce a new API named
WorkloadRebalancer
to support rescheduling. (#4841, @chaosi-zju)
- The following labels have been deprecated from release
v1.8.0
and now have been removed:resourcebinding.karmada.io/uid
clusterresourcebinding.karmada.io/uid
work.karmada.io/uid
propagationpolicy.karmada.io/uid
clusterpropagationpolicy.karmada.io/uid
- The following labels now have been deprecated and removed:
resourcebinding.karmada.io/key
replaced byresourcebinding.karmada.io/permanent-id
clusterresourcebinding.karmada.io/key
replaced byclusterresourcebinding.karmada.io/permanent-id
work.karmada.io/namespace
replaced bywork.karmada.io/permanent-id
work.karmada.io/name
replaced bywork.karmada.io/permanent-id
resourcebinding.karmada.io/depended-id
karmadactl
: The flag--cluster-zone
, which was deprecated in release 1.7 and replaced by--cluster-zones
, now has been removed. (#4967, @RainbowMango)
karmada-operator
: Fixed thekarmada-search
can not be deleted issue due to missingapp.kubernetes.io/managed-by
label. (#4674, @laihezhao)karmada-controller-manager
: Fixed deployment replicas syncer in case deployment status changed before label added. (#4721, @chaosi-zju)karmada-controller-manager
: Fixed incorrect annotation markup when policy preemption occurs. (#4751, @XiShanYongYe-Chang)karmada-controller-manager
: Fixed the issue of EndpointSlice residual in case of the karmada-controller-manager restart. (#4737, @XiShanYongYe-Chang)karmada-controller-manager
: Fixed deployment replicas syncer in case thatstatus.replicas
haven't been collected from member cluster to template. (#4729, @chaosi-zju)karmada-controller-manager
: Fixed the problem that labels cannot be deleted via Karmada propagation. (#4784, @whitewindmills)karmada-controller-manager
: Fixed the problem that work.karmada.io/permanent-id constantly changes with every update. (#4793, @whitewindmills)resourceinterpreter
: Avoid delete the key with empty value in object (lua table). (#4656, @chaosi-zju)karmada-controller-manager
: Fixed the bug of mcs binding losing resourcebinding.karmada.io/permanent-id label. (#4818, @whitewindmills)resourceinterpreter
: Prune deployment revision annotations. (#4946, @a7i)karmada-controller-manager
: Fix depended-by label value exceed 63 characters in dependencies-distributor. (#4989, @XiShanYongYe-Chang)karmadactl
: Fix register cluster creates agent missing--cluster-namespace
parameter. (#5005, @yanfeng1992)
- Bump google.golang.org/protobuf from 1.31.0 to 1.33.0 fix CVE(CVE-2024-24786) concerns. (#4715, @liangyuanpeng)
- Upgrade rsa key size from 2048 to 3072. (#4955, @chaosi-zju)
- Replace
text/template
withhtml/template
, which adds security protection such as HTML encoding and has stronger functions. (#4957, @chaosi-zju) - Grant the correct permissions when creating a file. (#4960, @chaosi-zju)
karmada-controller-manager
: Using the natural ordering properties of red-black trees to sort the listed policies to ensure the higher priority (Cluster)PropagationPolicy being processed first to avoid possible multiple preemption. (#4555, @whitewindmills)karmada-controller-manager
: IntroduceddeploymentReplicasSyncer
controller which syncs Deployment's replicas from the member cluster to the control plane, while previoushpaReplicasSyncer
been replaced. (#4707, @chaosi-zju)karmada-metrics-adapter
: Introduced the--profiling
and--profiling-bind-address
flags to enable and control profiling. (#4786, @chaosi-zju)karmada-metrics-adapter
: UsingTransformFunc
to trim unused information to reduce memory usage. (#4796, @chaunceyjiang)karmadactl
: Introduced--image-pull-policy
flag to theinit
command, which will be used to specify the image pull policy of all components. (#4815, @XiShanYongYe-Chang)karmada-controller-manager
: PropagateSecret
of typekubernetes.io/service-account-token
. (#4766, @a7i)karmada-metrics-adapter
: Add QPS related parameters to control the request rate of metrics-adapter to member clusters. (#4809, @chaunceyjiang)thirdparty
: Showstatus.labelSelector
for CloneSet. (#4839, @veophi)karmada-controller-manager
: Add finalizer for propagation policy. (#4836, @whitewindmills)karmada-scheduler
: Introduce a mechanism to scheduler to actively trigger rescheduling. (#4848, @chaosi-zju)karmada-operator
: Allow the user to specifyimagePullPolicy
in Karmada CR when installing via karmada-operator. (#4863, @seanlaii)karmada-controller-manager
: Support update event in WorkloadRebalancer. (#4860, @chaosi-zju)karmada-controller-manager
: Remove cluster specificPersistentVolume
annotationvolume.kubernetes.io/selected-node
. (#4943, @a7i)karmada-webhook
: Add validation on policy permanent ID. (#4964, @whitewindmills)karmada-controller-manager
: Support auto delete WorkloadRebalancer when time up. (#4894, @chaosi-zju)karmadactl
: Integrated with OIDC authentication for cluster operation auditing and access control. (#4883, @guozheng-shen)
- Bump golang version to
v1.21.8
. (#4706, @Ray-D-Song) - Bump Kubernetes dependencies to
v1.29.4
. (#4884, @RainbowMango) - Karmada is now built with Go1.21.10. (#4920, @zhzhuang-zju)
- The base image
alpine
now has been promoted fromalpine:3.19.1
toalpine:3.20.0
.
Helm Chart
: Update operator crd when upgrading chart. (#4693, @calvin0327)- Upgrade
bitnami/common
dependency in karmada chart from1.x.x
to2.x.x
. (#4829, @warjiang)
Thank you to everyone who contributed to this release!
Users whose commits are in this release (alphabetically by username)
- @a7i
- @Affan-7
- @B1F030
- @calvin0327
- @chaosi-zju
- @chaunceyjiang
- @dzcvxe
- @Fish-pro
- @grosser
- @guozheng-shen
- @hulizhe
- @Jay179-sudo
- @jwcesign
- @khanhtc1202
- @laihezhao
- @liangyuanpeng
- @my-git9
- @RainbowMango
- @Ray-D-Song
- @rohit-satya
- @seanlaii
- @stulzq
- @veophi
- @wangxf1987
- @warjiang
- @whitewindmills
- @wzshiming
- @XiShanYongYe-Chang
- @yanfeng1992
- @yike21
- @yizhang-zen
- @zhzhuang-zju