File kubernetes1.19.changes of Package kubernetes1.19
-------------------------------------------------------------------
Wed Jan 20 10:01:24 UTC 2021 - dmueller@suse.com
- Update to version 1.19.7:
* use default unkown sock for kubeadm cmd if cri detect is not needed
* cherry-pick part of #97451: fix nodeport quota check failure during creating clusterip
* Release reserved GCE IP address after ensure completes.
* Ensure reproducible builds when build through docker
* Fix cadvisor machine metrics
* Create OWNERS for most of the API Priority and Fairness impl
* Fix bug in CPUManager with race on map acccess
* fix azure file secret not found issue
* fix: azure file latency issue for metadata-heavy workload
* Update CHANGELOG/CHANGELOG-1.19.md for v1.19.6
* Add more logging for Mount error
-------------------------------------------------------------------
Tue Jan 05 12:02:26 UTC 2021 - rbrown@suse.com
- Update to version 1.19.6:
* volumebinding: report UnschedulableAndUnresolvable status instead of an error when PVC not found
* E2E stress test suite for VolumeSnapshots
* [go1.15] Use go-runner:buster-v2.2.2 image (built on go1.15.5)
* [go1.15] staging/publishing: Set default go version to go1.15.5
* [go1.15] Update to go1.15.5
* [go1.15] hack/tools: Update to k/repo-infra@v0.1.3 (supports go1.15.5)
* [go1.15] build: Update to k/repo-infra@v0.1.3 (supports go1.15.5)
* Flush FibreChannel devices before deleting
* Fix FibreChannel volume plugin corrupting filesystem on detach
* vendor: update cAdvisor to v0.37.3
* Update CHANGELOG/CHANGELOG-1.19.md for v1.19.5
* fix migration logic
* Add AcceleratorStats to cri_stats_provider
* coredns dep.Severity is newdefault, not newDefault
* Update ingress conformance test for finalizers
* bump cadvisor to 0.37.2
* Restore beta os/arch labels on initial node registration
* fix bug: concurrent map writes error
* range_allocator: Test (lack of) double counting
* cidrset: Add test for double counting
* Fix double counting of IP addresses
* Choosing the right source VIP for local endpoints
* Bump node-problem-detector to v0.8.5
* allow configuring ReadIdelTimeout and PingTimeout via env var
* Add a unit test testing the HTTP/2 health check help the REST client detects broken TCP connections.
* configure the ReadIdleTimeout and PingTimeout of the h2 transport
* update golang.org/x/net and golang.org/x/sys, equivalent of #96549
* fix: resize Azure disk issue when it's in attached state
* Remove ready directory which created in empty volumeMounter setUp func
* Updating to cadvisor v0.37.1
* Fix cacheWatcher leak when time jump to the future and jump back
* update e2e kubectl test
* prune type in preserve-unknown-fields objects
* apiextensions: prune array type without items in published OpenAPI
* fix pull image error from multiple ACRs using azure managed identity
* fix kube-proxy cleanup
* change GetFullQualifiedPluginNameForVolume to distinuish different drivers
* Fix a bug that DefaultPreemption plugin is disabled when using scheduler policy
* Update CHANGELOG/CHANGELOG-1.19.md for v1.19.4
* use patch instead of replace to test the dry-run option
* Fix --dry-run invocation in kubectl e2e
* Clean up remaining ns flag
* Use namespace flag passed to RunKubectl* methods
* Update max data disk count with new instance types
* fix the panic when kubelet registers if a node object already exists with no Status.Capacity or Status.Allocatable
* report UnschedulableAndUnresolvable status instead of an error when PVCs can't find bound persistent volumes
* do not allow inflight watermark histograms to fall too far behind
* Fix bug in JSON path parser where an error occurs when a range is empty
* Build files
* Allow priority to be set for kubelet process on Windows
* DelegatingAuthorizationOptions: exposes and sets a default timeout for SubjectAccessReview client
* Do not update managedFields timestamp when they don't change
* Mask bearer token in logs when logLevel >= 9
-------------------------------------------------------------------
Fri Nov 13 13:03:45 UTC 2020 - rbrown@suse.com
- Update to version 1.19.4:
* Do not skip externalLB update if some nodes are not found.
* Honor disabled LocalStorageCapacityIsolation in scheduling
* Disable watchcache for events
* Fix seccomp PSP docker/default annotation handling
* Re-add the event recorder in the release test
* Don't clear the cached resourcelock when errors occurs on updates
* Add failing test showing release is not working properly
* Add lease release tests in leader election
* Generate complete leader election record to resolve leader election issues with LeaseLocks
* Fix a bug that Pods with topologySpreadConstraints get scheduled to nodes without required labels
* If we set SelectPolicy MinPolicySelect on scaleUp behavior or scaleDown behavior,Horizontal Pod Autoscaler doesn`t automatically scale the number of pods correctly
* fix: add missing patch flag for kubeadm init phase control-plane
* Update the PIP when it is not in the Succeeded provisioning state during the LB update.
* Do not assume storageclass is still in-tree after csi migration
* Fix `kubectl describe ingress` format
* Fixes high CPU usage in kubectl drain
* Ignore deleted pods.
* Automated cherry pick of #94781: Don't attempt to detach an FC device if we don't know its name
* add a local tlsTransportCache to available_controller
* don't cache transports for incomparable configs
* Fix fcpath
* Update CHANGELOG/CHANGELOG-1.19.md for v1.19.3
* fix upgrade test
* fix azure disk data loss issue on Windows
* fix azure disk attach failure for disk size bigger than 4TB
* cloud node controller: handle empty providerID from getProviderID
* Updating kube-proxy to trim space from loadBalancerSourceRanges
* Don't create a new sandbox for pod with RestartPolicyOnFailure if all containers succeeded
-------------------------------------------------------------------
Thu Oct 15 17:06:17 UTC 2020 - rbrown@suse.com
- Update to version 1.19.3:
* Azure: fix node removal race condition on VMSS deletion
* Fix reporting network_programming_latency metrics in kube-proxy
* make download-or-bust compatible with both sha512/sha1
* replace sha1 with sha512
* avoid potential secret leaking while reading .dockercfg
* Mask Ceph RBD adminSecrets in logs when logLevel >= 4
* upgrade test for BoundServiceAccountTokenVolume
* fix detach azure disk issue when vm not exist
* vsphere: improve logging message on node cache refresh event
* Fix UpdateSnapshot when Node is partially removed
* kubeadm: make the CP join handling of kubeconfig similar to "init"
* kubeadm: warn but do not error out on missing CA keys on CP join
* Return the Kubernetes version which stopped serving deprecated APIs by default
* fix: detach azure disk broken on Azure Stack
* Ensuring EndpointSlices are recreated after Service recreation
* Handle nil elements when sorting, instead of panicking
* use more granular buckets for azure api calls
* do not mutate endpoints in the apiserver
* Remove HeadlessService label in endpoints controller before comparing
* test: add unit-test for TranslateCSIPVToInTree.
* fix azure file migration panic
* kubeadm: relax the validation of kubeconfig server URLs
* portforward: Fix UDP-only ports calculation
* make kube::util::find-binary not dependent on bazel-out/ structure
* output go_binary rule directly from go_binary_conditional_pure
* hack/lib/util.sh: some bash cleanups
* bazel: Replace --features with Starlark build settings flag
* [go1.15] staging/publishing: Set default go version to go1.15.2
* [go1.15] build: Use go-runner:buster-v2.0.1 (built using go1.15.1)
* [go1.15] Update to go1.15.2
* [go1.15] hack/tools: Update to k/repo-infra@v0.1.1 (supports go1.15.2)
* [go1.15] build: Update to k/repo-infra@v0.1.1 (supports go1.15.2)
* Update CHANGELOG/CHANGELOG-1.19.md for v1.19.2
* count of etcd object should be limited to the specified resource
* Track pods with required anti-affinity
* Ensure getPrimaryInterfaceID not panic when network interfaces for Azure VMSS are null
* Fix misusage of RLock in timeCache lru.Cache.Get()
* support ipv6 in e2e policy tests
* Allow 404 error on lb deletion in azure
* fix: azure disk resize error if source does not exist
* chore: add diskclient.Update interface
* chore: add diskclient.ListByResourceGroup interface
-------------------------------------------------------------------
Wed Oct 7 19:04:00 UTC 2020 - Dirk Mueller <dmueller@suse.com>
- use obsinfo again for commit id extraction, as this matches
upstream's build commit id (just in case).
-------------------------------------------------------------------
Wed Oct 7 14:37:30 UTC 2020 - Dirk Mueller <dmueller@suse.com>
- fix git commit determination (bsc#1177289)
-------------------------------------------------------------------
Thu Sep 17 09:05:50 UTC 2020 - rbrown@suse.com
- Update to version 1.19.2:
* Fix index out of range panic for kubectl alpha debug
* kubectl describe pod: use ReportingController as an event source
* Update CNI plugins to v0.8.7
* Add roundtrip tests for metrics repo
* Cleanup custom metrics conversion functions
* Update CHANGELOG/CHANGELOG-1.19.md for v1.19.1
* Sort list of formats for --logging-format description to make it deterministic
-------------------------------------------------------------------
Thu Sep 10 09:41:24 UTC 2020 - rbrown@suse.com
- Update to version 1.19.1:
* build/lib/release: Explicitly use '--platform' in building server images
* build/common.sh: Remove extraneous reference to debian-base image
* Update default etcd server to 3.4.13
* kubeadm: remove the CoreDNS check for supported image digests
* Update snapshot controller to use k8s.gcr.io
* kubeadm: make the scheduler and KCM connect to local endpoint
* Fixed reflector not recovering from "Too large resource version" errors with API servers 1.17.0-1.18.5
* let panics propagate up when processLoop panic
* kubeadm: Fix `upgrade plan` for air-gapped setups
* Add impersonated user to system:authenticated group
* cleanup: print warning message after timeout
* Revert "cleanup: decrease log level from warn to v3"
* cleanup: decrease log level from warn to v3
* Remove duplicate nodeSelector
* test(kubelet): add a regression test to verify kubelet would not panic
* fix(kubelet): protect `containerCleanupInfos` from concurrent map writes
* fix(azure): check error returned by scaleSet.getVMSS
* Fix issue on skipTest in storage suits
* Use NLB Subnet CIDRs instead of VPC CIDRs in updateInstanceSecurityGroupsForNLB
* Add PR #89069 Action Required
* Update CHANGELOG/CHANGELOG-1.19.md for v1.19.0
-------------------------------------------------------------------
Mon Sep 7 08:36:36 UTC 2020 - Richard Brown <rbrown@suse.com>
- kubernetes%{baseversion}-kubeadm now obsoletes kubernetes%{baseversionminus1}-kubeadm
-------------------------------------------------------------------
Fri Aug 28 13:01:23 UTC 2020 - Richard Brown <rbrown@suse.com>
- Introduce kubernetes$FOO-client-common package to make -client truely parralel installable
-------------------------------------------------------------------
Thu Aug 27 11:42:58 UTC 2020 - rbrown@suse.com
- Update to version 1.19.0:
* Updating EndpointSliceMirroring controller to listen for Service changes
* test(kuberuntime): deflake TestRemoveContainer
* build: Update to go-runner:buster-v2.0.0
* build: Migrate go-runner image building to k/release
* Reserve Static IP in Ingress test
* Add rbac patch permissions for system:controller:glbc ingresses/status
* Update json-patch to v4.9.0 tagged release
* synchronize agnhost image / app version
* track agnhost version in dependencies.yaml
* Make sampleAndWaterMarkHistograms not fall very far behind
* fix e2e test
* etcd 3.4.12
* Deflake TestServiceRegistryUpdateDryRun test
* Fixing a flaky EndpointSliceMirroring integration test
* Updating EndpointSlice controller to wait for all caches to be synced
* Deflake TestSubresourcePatch
* Speed up master reconciler tests
* Remove flaky audit e2e test
* Fix building with GOFLAGS=-v
* Fix issue in evaluating symlink path for subpath
* move ixdy to emeritus
* move krzyzacy to emeritus
* Fix bug for inconsistent lists served from etcd
* publishing: Drop 'go' field from 'controller-manager' 'master' branch
* publishing: Use go1.15.0 as default version
* publishing: Use go1.13.15 for 1.18, 1.17, and 1.16 release branches
* Refactor locks on registeredNodesLocks
* fix(pluginwatcher): watch socket dir before traversing it
* test(kubectl): deflake run_kubectl_apply_tests
* upgrade rules_docker
* Fix validation of ingress rules with wildcard host
* Add missing changelog entries for v1.18.7
* Add missing changelog entries for v1.17.10
* Keep track of remaining pods when a node is deleted.
* Remove direct accesses to cache's node map
* CHANGELOG: Update directory for v1.18.8 release
* [go115] Require go1.15 in build helper scripts
* CHANGELOG: Update directory for v1.17.11 release
* Add sigs for root folders
* CHANGELOG: Update directory for v1.16.14 release
* [go1.15] Update to go1.15
* [go1.15] hack/tools: Update to k/repo-infra@v0.0.12 (supports go1.15.0)
* [go1.15] build: Update to k/repo-infra@v0.0.12 (supports go1.15.0)
* [e2e]Bind csi driver namespace to test namespace to avoid namespace collision
* Promote spiffxp to build/ approver
* Updating EndpointSlice controllers to return if error encountered
* e2e storage: avoid flaky test failure when watch dies
* Validate ingress TLS secretName in v1
* e2e services wait for endpoint and endpoint slices
* [Flaky test]kubectl log test never restart pod
* cluster/images/etcd: Build etcd@3.4.9-3 using debian-base@v2.1.3
* build: Update to debian-base@v2.1.3 and debian-iptables@v12.1.2
* Updating EndpointSlice e2e tests to accept duplicate EndpointSlices
* test(endpoints): deflake TestWatchHTTPTimeout
* Sync cache before starting scheduler test
* forget the pod when the reserve plugins fail
* Update SMD to v4.0.1 and kube-openapi
* Wait for mock CSI Driver bringup to perform e2e test
* scheduler_perf: label nodes for pod affinity
* Skip LB test per schedulable node count
* fix(azure): set default value of VmssVirtualMachinesCacheTTLInSeconds during init to prevent data race
* [go1.15] Update to go1.15rc2
* generated: hack/update-hack-tools.sh && hack/update-vendor.sh
* [go1.15] hack/tools: Update to k/repo-infra@v0.0.11 (supports go1.15rc2)
* [go1.15] build: Update to k/repo-infra@v0.0.11 (supports go1.15rc2)
* Sign up dims for additional review roles
* Update to json-patch 4.8.0
* Deflake LimitRange e2e test
* Add test cluster check when starting csi proxy
* [go1.15] Use errors.As to unwrap net errors
* Fix namespace controller cleanup orphaning
* Add LabelSelector validation in Pod Affinity/AntiAffinity Filter and Score plugins
* Add tests for out of band attachments
* Add missing termination metrics to priority-and-fairness handler
* e2e delete namespaces after finish
* Add missing ppc64le architecture in the get-kube* scripts
* Requeue service after endpoint deletion
* Do not evict pods which tolerate all NoExecute taints
* fix: deflake run_kubectl_apply_tests
* Deflake TestUpdateNodeStatusWithLease - guard cached machineInfo
* CHANGELOG: Update directory for v1.19.0-rc.4 release
* Update CHANGELOG/CHANGELOG-1.19.md for v1.19.0-rc.4
* Fix /readyz to contain informer-sync
* Move bgrant0607 to emeritus status.
* Update e2e csi images to k8s.gcr.io
* Deflake scheduler PostBind integration test
* cloud-node-lifecycle controller: add missing instancev2 calls for node exists and node shutdown
* cloud provider: remove redundant implementation of InstanceMetadataByProviderID since instanceV2 is disabled for all providers
* cloud provider: remove provider ID references and improve documentation
* test/e2e: fail test rather than flooding logs if PVC watch is closed prematurely
* fix TestVolumeBinding unit test
* Deflake TestExpectationsOnRecreate
* Add kind/regression label to the pull request template
* Update Cluster Autoscaler version to 1.19.0
* Make AttachDisk idempotent again
* integration: Wait for PV to be available
* Fixing memory leak in EndpointSliceMirroring EndpointSlice tracker
* Fix an issue when rotated logs of dead containers are not removed.
* Remove unuseful error message about updating pod conditions not owned by kubelet
* Makefile: update test-e2e-node help
-------------------------------------------------------------------
Wed Aug 26 08:18:37 UTC 2020 - Richard Brown <rbrown@suse.com>
- Initial Packaging
