Thu Apr 13 11:32:32 UTC 2023 - Priyanka Saggu <priyanka.saggu@suse.com>
- Update to version 1.24.13:
* Release commit for Kubernetes v1.24.13
* releng/go: Update images, dependencies and version to Go 1.19.8
* wait again on pending state
* cacher allow context cancellation if not ready
* Drop development dependencies from test targets
* apiserver cacher: don't accept requests if stopped
* Clear front proxy headers after authentication is complete
* Make prerelease tag optional in CI versions
* Annotate CI version regexes
* Drop unused regex grouping
* Delete unused version regex function
* kubelet: Fix fs quota monitoring on volumes
* fsquota: only generate pod uuid is nil
* Change where transformers are called.
* Route controller should update routes with NodeIP changed When a node reboots or kubelet restarts, it is possible that its IP is changed. In this case, node route should be updated with the correct IP. In this PR, it checks if the IP in an existing route is the same as the actual one. If not, it marks it as "update" so the old route will be deleted and a new one will be created. There's a new field EnableNodeAddresses, which is a feature gate for specific cloud providers to enable after they update their cloud provider code for CreateRoute().
* client-go/cache: update Replace comment to be more clear
* client-go/cache: rewrite Replace to check queue first
* releng: Update images, dependencies and version to Go 1.19.5
* Explicitly call rand.Seed() method
* Improve vendor verification works for each staging repo
* Bump Konnectivity to v0.0.35
* Add pod to dsw if termination is not completed during reconstruction #issues/113979
* integration: migrate taint tests
* integration: migrate scoring tests
* integration: migrate preemption tests
* integration: migrate plugings tests
* integration: migrate extender tests
* integration: scheduler: migrate PDB from v1beta1 to v1
* Fix issues in volumesnapshot test for ephemeral storage
* update golangci-lint for go 1.19
* golang: Update to 1.19
* Adjust for os/exec changes in 1.19
* Update golangci-lint to 1.46.2 and fix errors
* Windows Kube-Proxy implementation for internal traffic policy.
* Fix a regression that scheduler always go through all Filter plugins
* Fix SPDY proxy authentication with special chars
* Creating Ingress IP loadbalancer alone when all the endpoints are terminating. KEP1669
* change k8s.gcr.io/pause to registry.k8s.io/pause
* Update golang.org/x/net 1e63c2f
* image pull event include duration with waiting
* kubelet: make the image pull time more accurate in event
* update structured-merge-diff to 4.2.3
* regression test for exponential recursion bug on CRDs
* Fix endpoint reconciler failing to delete masterlease
* kubeadm: remove v1.25 etcd "3.5.6-0" for v1.24
* use etcd 3.5.6-0 after promotion
* changelog: CVE-2022-3294 and CVE-2022-3162 were fixed in v1.23.14
* upgrade system-validators to v1.8.0 for a bugfix of cgroupv2 io check
* Introducing LoadbalancerPortMapping flags for VipExternalIP
* egress_selector: prevent goroutines leak on connect() step.
* Merge pull request #113133 from sxllwx:automated-cherry-pick-of-#113133-upstream-release-1.25
* Fixed (CVE-2022-27664) Bump golang.org/x/net to v0.1.1-0.20221027164007-c63010009c80
* Add CVE-2022-3162 to CHANGELOG-1.24.md
* tls.Dial() validates hostname, no need to do that manually
* e2e: use custom timeouts in GetSnapshotContentFromSnapshot()
* test/e2e/storage: replace hardcoded value with custom timeout in cleanup routine
* StatefulSet: Cleanup the complex defer function updating the status
* Be sure to update the status of StatefulSet even if the new replica creation fails
* added retries to winkernel proxy rules deletion
* added backend hashing to winkernel proxier
* kubelet: fix pod log line corruption when using timestamps and long lines
* kubeadm: mutate ClusterConfiguration.imageRepository to "registry.k8s.io"
* kubeadm: use registry.k8s.io instead of k8s.gcr.io
* add GetAllocatableCPUs test in cpumanager
* fix GetAllocatableCPUs in cpumanager
* e2e: restore volume lifecycle checks for csi-hostpath driver
* kubelet: fix volume reconstruction for CSI ephemeral volumes
* NodeLifecycleController: Remove race condition
* kube-proxy wait for cluster cidr skip delete events
* kube-proxy handle node PodCIDR changs
* kube-proxy: gate topology correctly
* service update event should be triggered when appProtocol in port is changed.
* filter out terminated containers in cadvisor_stats_provider
* Fix winkernel proxier setting the wrong HNS loadbalancer ID for ingress IP
* Bump konnectivity-client to v0.0.33
* Fix list estimator for lists that are executed as gets
* kubeadm: allow RSA and ECDSA format keys in preflight check
* Limit redirect proxy handling to redirected responses
* Make sure auto-mounted subpath mount source is already mounted
* Call SetupDevice only if Volume is not globally Mounted
* Fixes kubelet log compression on Windows
* Add zone field to vsphere test cloudconfig
* Reduce default gzip compression level from 4 to 1 in apiserver
* exec auth: support TLS config caching
* Add an option for aggregator
* Update go-runner to v2.3.1-go1.18.6-bullseye.0
* Update kube-cross image to v1.24.0-go1.18.6-bullseye.0
* Fix problem in updating VolumeAttached in node status
* Call queueSet::boundNextDispatchLocked enough
* Always log APF InitialSeats and FinalSeats values
* Marshal MicroTime to json and proto at the same precision
* Windows: ensure runAsNonRoot does case-insensitive comparison on user name
* Tolerate sub-microsecond eventTime changes on update
* Improve kubectl display of invalid errors
* fix unmatch reason when updating pod status
* fix nestedPendingOperations mount and umount parallel bug
* client-go/rest: check if url is nil to prevent nil pointer dereference
* Revert "client-go: remove no longer used finalURLTemplate"
* Skip "instance not found" error for LB backend address pools
* Update cel-go to v0.10.2.
* fix a memory leak problem when calling DryRunPreemption
* Fix JobTrackingWithFinalizers when a pod succeeds after the job fails
* Use CheckAndMarkAsUncertainViaReconstruction for uncertain volumes
* Remove volume from found during reconstruction if mounted
* Add unit test for verifying if processReconstructedVolumes works as expected
* Fix code to process volumes which were skipped during reconstruction
* Keep track of each pod that uses a volume during reconstruction
* allow namespace admins to use leases to encourage migration off of configmaps
* Fix: filter out unsatisfied nodes when calling AddPod in PodTopologySpread
* Fix `kubeadm upgrade plan` issue with FQDN nodes names
* Add rate limiting when calling STS assume role API
* Fix kubelet panic when accessing metrics/resource endpoint
* Fixing issue in generatePodSandboxWindowsConfig for hostProcess containers by where pod sandbox won't have HostProcess bit set if pod does not have a security context but containers specify HostProcess.
* Add retry logic for Unix Domain sockets on Windows
* Execute the Run function of kubelet, no log output after failure
