kubernetes1.20-kubelet-common-1.20.2-lp153.1.21 >  A `Mڸ/=„cH sbg8.Z%OѻR416Zyi\~p3e~_,i$*qׯ5?gS+^gRh,N{8HZ鴋>aqc<]\'~B=P~TEҺK+c<~OD>U{w]v?ߨЪ2X<ɿG Ѿ01AzAv&?b2$Ijd5558437dc6358d13d59477c6827e34090972a97370914c024ef0d1274b9c091810f2111641269c41d803bd23f5ad2a2bbab2273S`Mڸ/=„dsusH yb .v{Ar)UIߗS$љn}?!˧͜OEܽi'ÿZ2d0k6S HE*D%8ۚ>\&Yh˱Щ|, YGi,w`K\;B\}i~"„aٔlW,RvRD7)\aO 0Rc^շ$`װY2jBi̿Hg~RQ>pIn?nd ' 2 L  6<Bc  0 T x   ,t(T (  (8$98:=g>g ?g@gFg$Gg8HgIgXgYg\h]h\^iybjckldkekfklkukvl4wmxmXymzmmmmnCkubernetes1.20-kubelet-common1.20.2lp153.1.21Kubernetes kubelet daemonManage a cluster of Linux containers as a single system to accelerate Dev and simplify Ops. kubelet daemon`Ms390zp29openSUSE Leap 15.3openSUSEApache-2.0https://bugs.opensuse.orgSystem/Managementhttps://kubernetes.io/linuxs390x if [ -x /usr/bin/systemctl ]; then test -n "$FIRST_ARG" || FIRST_ARG="$1" [ -d /var/lib/systemd/migrated ] || mkdir -p /var/lib/systemd/migrated || : for service in kubelet.service ; do sysv_service=${service%.*} if [ ! -e /usr/lib/systemd/system/$service ] && [ ! -e /etc/init.d/$sysv_service ]; then mkdir -p /run/systemd/rpm/needs-preset touch /run/systemd/rpm/needs-preset/$service elif [ -e /etc/init.d/$sysv_service ] && [ ! -e /var/lib/systemd/migrated/$sysv_service ]; then /usr/sbin/systemd-sysv-convert --save $sysv_service || : mkdir -p /run/systemd/rpm/needs-sysv-convert touch /run/systemd/rpm/needs-sysv-convert/$service fi done fi PNAME=kubelet SUBPNAME=-kubernetes1.20 SYSC_TEMPLATE=/usr/share/fillup-templates/sysconfig.$PNAME$SUBPNAME # If template not in new /usr/share/fillup-templates, fallback to old TEMPLATE_DIR if [ ! -f $SYSC_TEMPLATE ] ; then TEMPLATE_DIR=/var/adm/fillup-templates SYSC_TEMPLATE=$TEMPLATE_DIR/sysconfig.$PNAME$SUBPNAME fi SD_NAME="" if [ -x /bin/fillup ] ; then if [ -f $SYSC_TEMPLATE ] ; then echo "Updating /etc/sysconfig/$SD_NAME$PNAME ..." mkdir -p /etc/sysconfig/$SD_NAME touch /etc/sysconfig/$SD_NAME$PNAME /bin/fillup -q /etc/sysconfig/$SD_NAME$PNAME $SYSC_TEMPLATE fi else echo "ERROR: fillup not found. This should not happen. Please compare" echo "/etc/sysconfig/$PNAME and $TEMPLATE_DIR/sysconfig.$PNAME and" echo "update by hand." fi if [ -x /usr/bin/systemctl ]; then test -n "$FIRST_ARG" || FIRST_ARG="$1" [ -d /var/lib/systemd/migrated ] || mkdir -p /var/lib/systemd/migrated || : if [ "$YAST_IS_RUNNING" != "instsys" ]; then /usr/bin/systemctl daemon-reload || : fi for service in kubelet.service ; do sysv_service=${service%.*} if [ -e /run/systemd/rpm/needs-preset/$service ]; then /usr/bin/systemctl preset $service || : rm "/run/systemd/rpm/needs-preset/$service" || : elif [ -e /run/systemd/rpm/needs-sysv-convert/$service ]; then /usr/sbin/systemd-sysv-convert --apply $sysv_service || : rm "/run/systemd/rpm/needs-sysv-convert/$service" || : touch /var/lib/systemd/migrated/$sysv_service || : fi done fi [ -z "${TRANSACTIONAL_UPDATE}" -a -x /usr/bin/systemd-tmpfiles ] && /usr/bin/systemd-tmpfiles --create /usr/lib/tmpfiles.d/kubelet.conf || : test -n "$FIRST_ARG" || FIRST_ARG="$1" if [ "$FIRST_ARG" -eq 0 -a -x /usr/bin/systemctl ]; then # Package removal, not upgrade /usr/bin/systemctl --no-reload disable kubelet.service || : ( test "$YAST_IS_RUNNING" = instsys && exit 0 test -f /etc/sysconfig/services -a \ -z "$DISABLE_STOP_ON_REMOVAL" && . /etc/sysconfig/services test "$DISABLE_STOP_ON_REMOVAL" = yes -o \ "$DISABLE_STOP_ON_REMOVAL" = 1 && exit 0 /usr/bin/systemctl stop kubelet.service ) || : fi test -n "$FIRST_ARG" || FIRST_ARG="$1" if [ $1 -eq 0 ]; then # Package removal for service in kubelet.service ; do sysv_service="${service%.*}" rm "/var/lib/systemd/migrated/$sysv_service" || : done fi if [ -x /usr/bin/systemctl ]; then /usr/bin/systemctl daemon-reload || : fi if [ "$FIRST_ARG" -ge 1 ]; then # Package upgrade, not uninstall if [ -x /usr/bin/systemctl ]; then ( test "$YAST_IS_RUNNING" = instsys && exit 0 test -f /etc/sysconfig/services -a \ -z "$DISABLE_RESTART_ON_UPDATE" && . /etc/sysconfig/services test "$DISABLE_RESTART_ON_UPDATE" = yes -o \ "$DISABLE_RESTART_ON_UPDATE" = 1 && exit 0 /usr/bin/systemctl try-restart kubelet.service ) || : fi fi$Dop :,^-AAA큤A큤A큤A큤AA`L`L`M_к`L`L`L`L`M_E_E_E`L`M_E`L`L`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@rootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootrootkubernetes1.20-1.20.2-lp153.1.21.src.rpmkubernetes-kubelet-commonkubernetes1.20-kubelet-commonkubernetes1.20-kubelet-common(s390-64) @    /bin/sh/bin/sh/bin/sh/bin/sh/bin/shcri-runtimekubernetes-kubelet1.20rpmlib(CompressedFileNames)rpmlib(FileDigests)rpmlib(PayloadFilesHavePrefix)rpmlib(PayloadIsXz)3.0.4-14.6.0-14.0-15.2-1kubernetes-kubelet-common4.14.1`@_T__м@_м@dmueller@suse.comrbrown@suse.comRichard Brown rbrown@suse.comRichard Brown - Update to version 1.20.2: * move all variables in sampleAndWaterMarkHistograms::innerSet * 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 the deadlock in priority and fairness config controller * Cherry pick 443 and 448 from cloud provider azure * Fix bug in CPUManager with race on map acccess * clean up executing request on panic * fix azure file secret not found issue * fix: azure file latency issue for metadata-heavy workload * Update CHANGELOG/CHANGELOG-1.20.md for v1.20.1 * Add more logging for Mount error- Update to version 1.20.1: * Revert "Use host IP instead of localhost for control plane component kubeconfig files." * etcd version for 1.19 is 3.4.13 for cve fixes * Flush FibreChannel devices before deleting * Fix FibreChannel volume plugin corrupting filesystem on detach * vendor: update cAdvisor to v0.38.6 * Update CHANGELOG/CHANGELOG-1.20.md for v1.20.0 * Revert "iAdd host IP to etcd listen client URLs." * fix migration logic * Add AcceleratorStats to cri_stats_provider- Add obsoletes to -client-common to facilitate smooth upgrades from older versions- Rebase opensuse-version-checks.patch - Update to version 1.20.0: * APF e2e: disable drown-out tests temporarily * coredns dep.Severity is newdefault, not newDefault * addressing review comments and supports parallel run * Update ingress conformance test for finalizers * vendor: update cAdvisor to v0.38.5 * fix: change disk client API version for Azure Stack * service.spec.AllocateLoadBalancerNodePorts followup * [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) * Restore beta os/arch labels on initial node registration * matches specific usernames instead of "*" * range_allocator: Test (lack of) double counting * cidrset: Add test for double counting * Fix double counting of IP addresses * ap&f e2e: eliminates client-side rate-limiting * fix bug: concurrent map writes error * Bump node-problem-detector to v0.8.5 * Revert "check volume directories instead of mounts for cleanupOrphanedPodDirs" * Revert "plumb context with request deadline" * Revert "use default value when the specified timeout is 0s" * Revert "add e2e tests for request timeout" * Deflake ThrottledLogger test * fix: resize Azure disk issue when it's in attached state * kube-aggregator: fix apiservice availability gauge * Fix TestStartingResourceVersion flakiness * make sure managedFields are written * generated * unit and integration tests * apiserver dedups owner references and adds warning * Add linuxonly on one multivolume test * gce: move iptables rule to mangle * APF e2e tests: add request drown-out fairness test * add e2e tests for request timeout * use default value when the specified timeout is 0s * CHANGELOG: Update error link in 1.20 * CHANGELOG: Update directory for v1.20.0-beta.2 release * APF e2e tests: rename request drown-out priority client names * Mark some storage tests as LinuxOnly * Add GC unit tests * Log cluster-scoped owners referencing namespaced owners, avoid retrying lookups forever * Queue non-matching children for deletion when a virtual node is marked as observed * Handle virtual delete events when children don't agree on owner coordinates * Make node removal conditional in processGraphChanges * Enqueue dependents for deletion when their ownerReference does not match observed parent coordinates * Short-circuit attemptToDelete loop for virtual nodes that are removed or observed * Replace virtual node with observed node if identity differs * Refactor identityFromEvent * Avoid marking virtual nodes as observed when they haven't been * Switch GC absentOwnerCache to full reference * Add GC integration race test * Plumb event recorder to garbage collector controller * e2e SCTP test not depend on kubenet * e2e/node: increase timeouts seconds to 5 for liveness probe restart test * correct e2e test predicates conflict hostport * Add a deprecation note to k/k/cluster/log-dump directory * Update topology tests for windows * kubelet: dockershim should return grpc status with DeadlineExceeded code * e2e dualstack test fixes * convert the runtimeclass API tests to conformance * APF e2e tests: use snake_case label * APF e2e tests: move common code into helper functions * Reduce volume name length for vsphere * CHANGELOG: add a hyperlink to issue 86282 * 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. * plumb context with request deadline * Fixes fake client test generation * Integrate defaults marker and remove ContainerPort defaulter * Remove StripDefaults from BuildSwagger * Add default for protocol and test that it works * Strip defaults in new places * Re-generate with defaults * Update kube-openapi * flowcontrol bootstrap: give catch-all PL more concurrency share * APF matching: fallback to catch-all if nothing matches * flowcontrol bootstrap: make exempt PL last * APF: graduate API and types to beta * Mixed protocol support for Services with type=LoadBalancer (#94028) * Fix test name in e2e resource metrics api test * Test CRUD operations on RuntimeClasses API * vendor: cadvisor v0.38.4 * Relax matching on pod_memory_working_set_bytes metrics * configure the ReadIdleTimeout and PingTimeout of the h2 transport * Match pod resource metrics to pod not container * add more e2e sctp tests * remove wrong test for SCTP connectivity * Update the route table tag in the route reconcile loop * Fixes sigfault in case of empty TopologyInfo * Update generated files * Add service.spec.AllocateLoadBalancerNodePorts * update golang.org/x/net and golang.org/x/sys * Update 1.18 changelog with changes from v1.18.11 * fixup! unblock resources that the storage version manager depends on * fixup! add storage version garbage collector * generated * GC integration test * unblock resources that the storage version manager depends on * make storage version manager wait for lease creation * add storage version garbage collector * Fix cacheWatcher leak when time jump to the future and jump back * Correct rebase issues * Implement shutdown manager in kubelet * Add systemd package to interface with dbus * Corrected CSIDriver validation rebase issues * Remove duplicate CSIDriver name validation * Adjust CSIDriver validation to check objectmeta * Enable logging and drop permissive targets for CSI mock driver * Update the mock driver to use 4.0.2 * Included e2e test for CSIDriver FSGroupPolicy * Move CSIVolumeFSGroupPolicy to beta * Relax validation for CSIVolumeFSGroupPolicy * Fixed quantization and made monotic time reversal not panic * Log defaulted kube-scheduler component config at startup * Updates related to PR feedback * core/v1: document that topologyKeys requires the ServiceTopology feature gate * Use K8s in the README * stop serving deleted APIs * Increase watch timeout when scaling Deployment Replicas * CHANGELOG: Update directory for v1.18.12 release * scheduler: Implement resource metrics at /metrics/resources * quantity: Allow quantity to be converted to float64 * plumb service account token down to csi driver * remove generator from service in kubectl * add myself to sig-network-api-reviewers * Add datapolicy tags to staging/src/k8s.io/client-go/ * cleanup: fix log capitalization in scheduler * Add --experimental-logging-sanitization flag to Kubelet * Run ./update-all.sh * Implement e2e tests for pod scope alignment * Add tests for getPodDeviceRequest() for devicemanager * Add tests for GetPodTopologyHints() for devicemanager * Update topology hints tests to use pod object for devicemanager * Add tests for GetPodTopologyHints() for cpumanager * Refactor topology hints tests for cpumanager * Move scope specific tests from topologymanager under particular scopes * Move common tests from topologymanager under scope * Update topologymanager tests after adding scopes * Implement devicemanager.GetPodLevelTopologyHints() function * Implement the cpumanager.GetPodTopologyHints() function * Update logging to use a format util * Implement topology manager scopes * Add GetPodTopologyHints() interface to Topology/CPU/Device Manager * Add flag value validation of TopologyManagerPolicy * Add flag value validation of TopologyManagerScope * Add kubelet configuration flag 'topology-manager-scope' * Support custom tags for cloud provider managed resources * use uncommon ports for e2e network test * fix pull image error from multiple ACRs using azure managed identity * add e2e test for dual-stack secondary service IPs * Verify iptable rules are applied for tcp, udp and icmp * Choosing the right source VIP for local endpoints * Configure StackdriverLogging Windows service to restart on failure. * Adding sample files to demonstrate how cloud provider leverage CCM. * Updating EndpointSlice strategy to cover alpha NodeName field * Support high availability ports * Allow debugging kubelet image pull times * Separate in-tree gcepd driver for windows * Cleaning up EndpointSlice update validation tests * Removing "IP" from supported EndpointSlice address types in kube-proxy * Updating ControlPlane to support NodeName field * Updating EndpointSlice controllers to support NodeName field * Adding EndpointSliceNodeName feature gate * Adding NodeName to EndpointSlice API, deprecation updates * update bazel * Promote Deployment lifecycle e2e test to Conformance * CHANGELOG: Update directory for v1.19.4 release * handle the case for slow cronjob lister, add unit tests * convert to stardard lister, use []*batchv1.Job instead of []batchv1.Job * actually retry if we failed to reconcile some objects * Update storage test suits for Windows * Promote Pod/PodStatus lifecycle e2e test to Conformance * Fix failures in TestBindPlugin and TestPreemptWithPermitPlugin * Set 0 sync period in scheduler integration test * CHANGELOG: Update directory for v1.17.14 release * GA of RuntimeClass feature gate and API * put a message, not a stack, in the log on a timeout * Update snapshot CRDs * Set priority of Event v1 higher than v1beta1 * Implement TopologyInfo and cpu_ids in podresources * Convert podDevices to struct * Generate podresources API for TopologyInfo and cpu_ids * Add TopologyInfo and cpu_ids into podresources * Change GetDevices interface * Revert "Merge pull request #92312 from Sh4d1/kep_1860" * fix all Sting method not check nil in "k8s.io/apimachinery" * fix the validation logic for Job/CronJob RestartPolicy field * fix note delete to update according to code * Fix a bug that DefaultPreemption plugin is disabled when using scheduler policy * fix(test::npd): provide NPD with proper kubeconfig * Bump kas to v0.0.14 * Capture defaulted plugin configs from framework * move lease controller to k8s.io/component-helpers/apimachinery * Bug Fix for process_start_time_metric initialization * add V(4) log when apiserver lease was deleted before this controller reacts * generated * integration test * add apiserver lease garbage collector * update violation_exceptions.list and make generated * add cronjob_controllerv2.go * Add CRI v1 proto * dualstack: Use Agnhost in place of BusyBox * Downward API hugepages * added new runtimeclass test and changed Disruptive to Serial * kubeadm: mark the "master" label/taint as deprecated * dualstack: use correct IPFamily list for conntrack checks in e2e * move service controller config to k8s.io/cloud-provider/controllers/service/config * Use FilteredDialContext with quobyte API * Update quobyte client API to v0.1.8 * import restrictions: allow k8s.io/kubelet to import credentialprovider apis * hack/.golint: ignore golint for new kubelet and credentialprovider APIs package * pkg/credentialprovider: export URL parsing and matching helper functions * pkg/credentialprovider: add initial exec-based credential provider plugin * kubelet: add initial credentialprovider v1alpha1 APIs * feature gates: add KubeletCredentialProviders feature gate * kubelet: support alpha credential provider exec plugins * kubelet: update pkg/kubelet/apis/config/OWNERS to include api approvers and reviewers * kubelet: add CredentialProviderConfig API * Change at which level klog.Fatal is invoked * Forbid creating clusters with more than 100 nodes without vpc-native * Change the logic of pod volumes existence check during kubelet cleanupOrphanedPodDirs, cleanupOrphanedPodCgroups and PodResourcesAreReclaimed * remove label dependency on k8s api in Azure * Remove duplicate import * Add --experimental-logging-sanitization flag to control plane components * FsgroupChange policy test suite * Improve observability of node authorizer: * Adding some metrics to the graph * Adding log message when node authorizer has synced * Remove ready directory which created in empty volumeMounter setUp func * cloud-provider: update docs and guidance for InstanceV2 and Zones * fixing issue where SMB share paths cannot resolve with CRI-containerD on Windows * Ignore specific Pod update events in scheduler * Support customize load balancer health probe protocol * Move fsGroupChangePolicy feature to beta * Add WindowsContainerResources to UpdateContainerResourcesRequest * Change snapshot test to use v1 apis * Modify storage snapshottable and disruptive test for Windows * fix: change storage account client API version for Azure Stack * Fixed failure: ProvisioningFailed: Failed to provision volume with StorageClass "standard": invalid AccessModes [ReadWriteOnce ReadOnlyMany ReadWriteMany]: only AccessModes [ReadWriteOnce ReadOnlyMany] are supported * Set enable konnectivity service to true by default * fixup! apiserver correctly validates encoding/decodable versions * HTTP Prove: Removes Accept-Encoding header from http probe * Update snapshot CRDs to v1 * Enable ConfigurableFSGroupPolicy feature gate * Add dangling volume check for vsphere * update e2e kubectl test * prune type in preserve-unknown-fields objects * Add AddedAffinity to the NodeAffinityArgs * Remove --redirect-container-streaming functionality (#95935) * Fix go lint on folder apimachinery/pkg/runtime/serializer/protobuf * Graduating AppProtocol to GA * Fix command and arg in NPD e2e * Call MountDevice only once * Mark MountDevice as uncertain after failed resize * Restore staging path creation * Add unit test for staging path creation * kubelet: move pkg/kubelet/cri/.import-restrictions to pkg/kubelet/cri/streaming * kubelet: add feature gate check for exec probe timeouts * features: add ExecProbeTimeout feature gate * kubelet: allow dockershim exec timeouts to be longer than 10s * kubelet: add e2e test for exec readiness probe timeout * kubelet: stop skipping docker exec livenessprobe timeout test * kubelet: respect probe exec timeout from CRI by returning utilexec.CodeExitError * kubelet: respect dockershim exec timeout * Implement log sanitization * Update bazel * apiextensions: adapt error tests to fixed validation messages * apiextensions: switch validation to kube-openapi * bump(k8s.io/kube-openapi) * stops puting a stacktrace of aborted requests in the logs * require APIServerIdentity to be enabled to run StorageVersionAPI * apiserver correctly validates encoding/decodable versions * updater correctly updates storageversion status * return a Status formatted JSON response * generated * Add an integration test. * make some rbac and scheduling post start hooks tolerate the apiserver bootstrap delay caused by installing storage versions. * Add a generic filter that blocks certain write requests before StorageVersions are updated during apiserver bootstrap. * Collect storage versions as ResourceInfo when installing API endpoints. * Add a feature gate * add andyzhangx as reviewer * Add nodeSelector for konnectivity daemonSet * fix ingress comparaison * fix defaulting * add nil case in proxy * fix build * fix nit in validation * fix tests * Update generated * fix reviews * Update generated * fix review * fix typo * add owner for feature gate * fix rebase * fix review * Add tests * Add route type field to loadbalancer status ingress * Update docs and fix redundant logic of scheduler perf * remove kube-proxy/config/v1alpha1 from .golint_failures * Fix go lint on folder apimachinery/pkg/runtime/serializer/json * Change the features removal note * Promote TokenRequest e2e test to Conformance * cleanup dated wording "NominatedNodeName annotation" * fix formatting * generated * add kube-apiserver-lease-controller poststart hook * Add constant PodReadyTimeout to e2e test * Adjusted timings and management of pods in e2e test * Create Pod+PodStatus resource lifecycle test * Update UTs * Fixing expected pod subdomain to match framework.TestContext.ClusterDNSDomain * e2e: add APF flowcontrol request drown-out test * Less restrictions for AWS NLB health check config * scheduler: remove FrameworkFactory. * PV e2e: fix race in NFS recycling test * Add support to size memory backed volumes * update max azure data disk count map * kubeadm: fix the lint failure where return value is not checked * Generated changes * Move the remaining kubectl bits to k8s.io/kubectl * Introduce kubectl-convert plugin * Adding config extension to CCM. * remove annotation cache sync because of code removed * Increase the timeout to allow summary test pod to start * removes filter metrics test retries * Use topology labels instead of old beta names (#96033) * CRs: Default non-nullable nulls * Move kubectl get-context validate logic to Validate function * endpointslice API: rename 'accepting' condition to 'serving' condition * endpointslice controller: add test cases to TestSyncServiceFull for terminating endpoints * endpointslice controller: refactor TestSyncServiceFull to use test tables * endpointslice API strategy: drop disabled fields 'accepting' and 'terminating' * endpointslice controller: set new conditions 'accepting' and 'terminating' * feature gate: add gate EndpointSliceTerminatingCondition * endpointslice API: add accepting and terminating conditions * kubelet: Use CRI SecurityProfile for Seccomp * Switch GCP list calls to paginated calls. * dualstack: cleanup IsIPv4 duplicates in favor of utils * dualstack: incorporate IsIPv4 updates from utils repo * abort if namespace doesn't exist or terminating * proxy: validate each CIDR config seperately and check for errors * minor changes in adding tests for checking metrics labels * make flags of TokenRequest required * default `service-account-extend-token-expiration` to true * E2E stress test suite for VolumeSnapshots * Update Microsoft/go-winio to released version * APF metrics: set StabilityLevel to ALPHA * kubectl debug: allow set-image-only invocation * Introduce a simple datapolicy library * Remove the dependency between create priorityclass command and generators * Deflake existing configmap count * update features to indicate beta in comment * cleanup: use i18n.T on all command descriptions * Revert "add e2e test for Service ExternalIPs" * CHANGELOG:Fix kubelet flag enable-cadvisor-json-endpoints * client-go/rest: fix finalURLTemplate for url base == "/" * Correctly fix clearing conntrack entry on endpoint changes (nodeport) * Remove Const IPVSProxyMode * update changelog-1.20 * test images: Authenticate in order to push images * Support multiple standard load balancers in one cluster * Use user facing field names in validation message * report UnschedulableAndUnresolvable status instead of an error when PVCs can't find bound persistent volumes * separate RootCAConfigMap from BoundServiceAccountTokenVolume * Add mrunalp as node approver * test images: Removes -p yes flag from qemu-user-static script call * optimise defaultpreemption: enumerate fewer candidates * apiserver/filters test: fix data race and do not leak goroutines * APF: use snake_case in metric labels * Volume snapshot e2e test to validate VolumeSnapshotContent and PVC finalizer * minor changes to tests for checking metrics labels based on review comments * check if kubectl version required values are empty * Exclude KEP-1933 from verify-all.sh until after alpha status. * add GVK to fake dynamic client to match actual behavior * demonstrate existing generated client and fake client behavior is consistent * remove dead apiserver field * Adding an owner for addon/dns folder * fix: pass bearer token to curl using -H instead of --oauth2-bearer * Remove alpha from kubectl debug * do not allow inflight watermark histograms to fall too far behind * test images: set DOCKER_CLI_EXPERIMENTAL=enabled * Remove maximum volume limit comment which is easily outdated. * DelegatingAuthenticationOptions: allows for setting a timeout for the TokenReview client that is used by for the webhook authenticator * Update staging/src/k8s.io/kubectl/pkg/cmd/create/create_rolebinding_test.go * pause image: Disable DiagTrack service on Windows image * replace string casting with fmt.sprintf in test * Fix paging issues when Azure API returns empty values with non-empty nextLink * change plugin name in fsgroupapplymetrics of csi and flexvolume to distinguish different driver * test images: sets HOME=/root in cloudbuild.yaml * Update CHANGELOG link of older releases * Remove TaintBasedEvictions Feature Gate * Replace calls to cs.CoreV1().Nodes().Create() with createNode() * Add a function to ensure created nodes are present in scheduler cache * local-up-cluster.sh: Use config file instead of flags for kubelet * Add a verbosity option to exec * KEP-1933: add static analysis target to hack/ * kubectl debug: Allow mutating image names * Increasing withTimeout for ReplicationController Lifecycle test steps * Propose seccomp/apparmor protobuf type definitions for CRI graduation * Update Klog dependency * Remove service load balancer feature gate * resource-metrics: add pod metrics e2e test * reenable e2e_node services & debugging improvements * resource-metrics: add pod/sandbox metrics to endpoint * Removing Alpha annotation as the feature is graduating to Beta in v1.20 * Add runtime representation of []v1.PreferredSchedulingTerm * Move pkg/kubectl/cmd/auth under staging/src/k8s.io/kubectl/pkg/cmd/auth * use patch instead of replace to test the dry-run option * CHANGELOG: Update directory for v1.20.0-beta.1 release * fix CPU time of pod stats with cs.CPU.Time * Skip the sig-storage e2e test as early as possible * Remove useless variable and if * kubectl debug: add tests for Complete,Validate * Move Snapshot to GA * register controllermanager.config.k8s.io group. * fix(test::npd): fix node problem detector test * fix n to nodeInfo easy to understand * Update cobra dependency to v1.1.1 * Adjusted e2e test watch timeouts, logging and pod management * Update usage information for --http-override * Bump agnhost version * Add override option to netexec * Add a redirect handler to netexec * Add optional code to netexec echo endpoint * Honor disabled LocalStorageCapacityIsolation in scheduling * Build files * [kubelet] Allow priority to be set for kubelet process on Windows * add e2e test for Service ExternalIPs * e2e cases for apf * Move helpers from pkg/registry/rbac/reconciliation and pkg/registry/rbac/validation under k8s.io/component-helpers * Add runtime representation of v1.NodeSelector * Remove variadic argument from storage interface * Revert "Updated golang/x/net. Also updated golang.org/x/sys" * Add multi request test * Fix issue in missing metrics of terminated requests * Fix bug in JSON path parser where an error occurs when a range is empty * set webhook retry backoff parameters for kubelet * thorw error if webhook retry backoof is not specified * make backoff parameters configurable for webhook * fix vendor/k8s.io/apimachinery/pkg/api/meta staticcheck * Updated golang/x/net. Also updated golang.org/x/sys, as required by hack/lint-dependencies.sh. * Disable watchcache for events * local-up-cluster.sh: Remove ineffective parameters * Pass all packages to conversion-gen * refactor: migrate health checks of control-plane off insecure port in tests * Implementing ExternalTrafficPolicy: local in winkernel kube-proxy via DSR * add a jitter to bound token renewal * Add datapolicy tags to cmd/kubeadm directory * devicemanager: fix race in stub * Add datapolicy tags to pkg/volume/ * change bucket from 0.0s to 0.0001s * refector service some e2e cases to make it runing in multi providers * Added tests to check metrics labels * Move CCM to staging k8s.io/cloud-provider * mv TokenRequest and TokenRequestProjection to GA * Populate ClusterIPs on read * test: remove flacky ut * Add json-response flag to porter * local-up-cluster.sh: Pass CLUSTER_CIDR to kube-proxy * HTTP Probe: Add 'Accept' header by default * Add keep_time_key setting to fluentd configuration * Enable filter latency tracking for request filters * Measure how much time a request spends in server filter(s): * hack/local-up-cluster.sh: fix API_PORT * exec credential provider: exec -> client.authentication.k8s.io/exec * kubectl flush profiling when get a sigterm * Cleanup non-namespaced objects in e2e test during interrupts * remove dead negotiation methods * fake dynamic client: support *List kinds * Move informer_factory to staging * Remove behaviors * add tests that update services while gate is off * Add datapolicy tags to staging/src/k8s.io/kubectl * Add datapolicy tags to staging/src/k8s.io/legacy-cloud-providers * Mark SecondaryRangeName as deprecated. * Move MatchNodeSelectorTerms to k8s.io/component-helpers * exec credential provider: k8s.io/client-go/tools/auth/exec helper * exec credential provider: ProvideClusterInfo and kubeconfig shadow * Remove FieldMatchingFlags * Adjust conversion generator to new converter changes * Add datapolicy tags to pkg/scheduler/ * Add datapolicy tags to staging/src/k8s.io/kube-aggregator/ * Add datapolicy tags to test/e2e/framework * Add datapolicy tags to pkg/apis * kubeconfig: add explicit path, if specified to loading precedence * Introduce api-extensions category in k8s apiserver * refactor(apiserver): ignore the insecure flags * Allow for configuring etcd progress notify interval on GCE * Fix seccomp PSP docker/default annotation handling * fixes max-min fairness * local-up-cluster.sh: Pass SERVICE_CLUSTER_IP_RANGE to controller manager * fix kubectl debug link error * Added config parameter for CPU threads * Updating EndpointSliceMirroring e2e test to accept multiple slices * pin to latest k8s version * more versions * rev versions * use current method signatures * Update csi-proxy version * add audit-log-compress to apiserver * Update PriorityClass conformance test to cover DeleteCollection * SetCondtion updates generation * Update testing-manifests/storage-csi owners file * CHANGELOG: Update directory for v1.20.0-beta.0 release * 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 * Wipe some fields on service "type" updates * Make some methods into non-methods * If image has stackdriver agent installed, use it. * remove the deprecated client that we stopped generating * generated * Update function setting master node size for GCE * dualstack endpoints integration tests * fix case when HC timeout is 0 * pv controller test: more test cases * pv controller test: use sub tests * pv controller test: enable klog output * cleanup: fix some error log capitalization * fix unbound variable on upgrade * add godoc for events to events.k8s.io * Event: Document TTL and best-effort-ness * Add the pod_resources_endpoint_requests_total metric * Update podresources api e2e_node tests * tombstone-ing IPFamily field(15) * fake dynamic client: document that List does not preserve TypeMeta in UnstructuredList * Change function signature for MatchNodeSelectorTerms * Use host IP instead of 127.0.0.1 for kube-apiserver healthcheck. * Webhook: handle error when calling wait.ExponentialBackoff * Fix cacher test after bumping fakeBudget timeout to 2 seconds * Move pkg/apis/core/v1.IsScalarResourceName under pkg/scheduler/util * fixed addons fluentd-elasticsearch statefulset format error * PV controller: don't delete PVs when PVC is not known yet * follow up for #94109 * Enable Volume Expansion tests for Windows * scheduler: make Profile an interface. * Promote verify PriorityClass endpoints e2e test to Conformance * add e2e test for services with hostNetwork endpoints * e2e use functional options to configure NetworkingTest * fix e2e service test container listening port * dual stack services (#91824) * Kubelet now implements the V1 podresources API * Update generated files * Add podresources v1 API * Fix cacher test flakiness * Make versionconverter functions private * Reuse SSA type converter for resources in the same API Group * Change pvc describe test * add an APIServerIdentity feature gate * DelegatingAuthorizationOptions: exposes and sets a default timeout for SubjectAccessReview client * Fix the kube-proxy comment so that the document can be generated correctly * Fix a lot of typos in Azure codes * pause image: Stricter registry prefix regex * apiserver: use canonical egress selection names in EgressSelectorConfiguration API docs * apiserver: update TestReadEgressSelectorConfiguration to use 'controlplane' egress selector name * apiserver: add validation for EgressSelection names in EgressSelectorConfiguration API * apiserver: support egress selection name 'controlplane' and deprecate 'master' * Update test/e2e/storage owners file * Report a metric for time taken to perform recursive permission change * Log PodExec stdout + stderr * Add e2e test for the newPV Controller metric * removed whitspace * added xpack gem to enable ilm support in fluentd-es-image * Add yaml util to unmarshal numbers into int/float * Update the frontend IP config when the service's `pipName` annotation is changed * fix typo in e2e test * Add MrHohn as an owner of ip-masq-agent addon * Bump ip-masq-agent version to pick up CVE fixes * ipvs: check for existence of scheduler module and fail if not found * fix: do not hardcode nginx image URL in rc e2e test case * Grant group KUBE_POD_LOG_READERS_GROUP access to read pod logs on gke control-plane. * Add configuration options to specify --detect-local-mode on kube-proxy. * Fix static checks for pkg/controller/podautoscaler * proxy: label kube_proxy test with more unique label * Optimize NormalizeScore for PodTopologySpread * Add 5k nodes benchmark for pod topology spreading * PV Controller: PV plugin and mode metrics * style: update comments in topology manager * Optimize string building for NamespacedName * Clean code: optimize some logs of mount * extend request interval to make session affinity cases stable * Add jingxu97 to volume/util owners * Delete framework/v1alpha1 folder and change remaining import paths * Fix a bug that Pods with topologySpreadConstraints get scheduled to nodes without required labels * Fix staticcheck failures on apiserver/plugin/pkg/{authenticator, authorizer} * Alter wording to describe pods using a pvc * Deflake PostFilter integration test * allow component-helpers to import util and klog * generated * generalize lease controller * move node lease controller to component-helpers * Add integration test for Default PodTopologySpread * bump qemu version * Rename flags * Replacing factory tests with scheduler.New and options * Add SETUP_KONNECTIVITY_SERVICE flag * Separate network proxy flag for apiserver egress and starting pods * kubeadm: validate node-cidr-mask are correct * kubeadm validate maximum service subnet size * write checkpoint only when allocated devices updated. * kubeadm: validate podSubnet against node-cidr-mask * Properly quote flags passed to Cluster Autoscaler * Update nodelocaldns yaml to use 1.15.16 image * remove unused const failedExpiration * If we set SelectPolicy MinPolicySelect on scaleUp behavior or scaleDown behavior,Horizontal Pod Autoscaler doesn`t automatically scale the number of pods correctly * format incorrectAddresses in klog * fix: add missing patch flag for kubeadm init phase control-plane * fix loop bug for verify attached volume * test: Add service cluster IP range unit test * prepare node performance tests to move to separate job * ipset: Address a TODO, add test for TestEntry() with IPv6 address * Fix labels for spreading benchmark * Added functionality and API for pod autoscaling based on container resources * Add more Pods and relax skew in E2E spread test * Adjust the acceptable value of UsageNanoCores * Return non-zero code on logexporter failures * Fix default values for logrotate in /var/log/ * Add integration test for kubectl debug * Fix golint failures in client-go/transport * Update the PIP when it is not in the Succeeded provisioning state during the LB update. * Support cross compilation only on amd64. * Always set relevant variables for cross compiling * Use host IP instead of localhost for control plane component kubeconfig files. This is a part of work to allow control plane components to be moved off hostNetwork. * Allow deletion of unhealthy pods if enough healthy * iAdd host IP to etcd listen client URLs. Allow kube-apiserver to use host IP to connect to etcd. Update etcd/migrate to allow additional client listening URLs. * Add error detail in log * CHANGELOG: Update directory for v1.20.0-alpha.3 release * fix the fake cloud provider * Warn that hyper-v featuregate will be deprecated in 1.20 and will be removed in 1.21 * Do not skip externalLB update if some nodes are not found. * Promote Replication Controller lifecycle e2e test to Conformance * Promote Replication Controller Lifecycle * Move client_builder to k8s.io/controller-manager * Graduate DefaultPodTopologySpread to beta * e2e: topomgr: fix ginkgo log * netpol: Add CRUD tests for NetworkPolicy API * test images: Adds Windows support (part 2) * refactor: migrate node e2e tests off insecure port * Generated bazel * Generated conversions * exec credential provider: wire in cluster info * Fix dependency linting * Kube-proxy: Perf-fix: Shrink INPUT chain * Fix flaky unit test Test_Run_Positive_VolumeMountControllerAttachEnabledRace data race * conntrack log delete operations * Bump github.com/Azure/go-autorest/autorest/adal to 0.9.5 * kube-proxy: log stale services operations * staging/publishing: cleanup import restrictions for csi-translation-lib * tests: Refactor agnhost image pod usage - common (part 2) * test, e2e: Remove duplication when using PodClient.CreateSync * fix test flakes * kubectl: allow users to use args with KUBECTL_EXTERNAL_DIFF * e2e don't use hardcoded name for containers name * avoid apiserver/cloud-provider/controller-manager deps from k8s.io/csi-translation-lib * Add client listening URLs flag to etcd migrate. * test: Fix using deprecated default cluster IPs * Disable test for large clusters * The function shouldRecordEvent will panic when the value of input object is nil * Fix catch all regex and missing DryRun Options * fix unbound variable issue in verify.sh * support multiple bind records (fie nodelocaldns test regression), by first replacing PILLAR_ and then replacing other vars. * Add back openapi gen for generic types and clean up doc.go * fix a nil pointer ref in slice allocation * warn user about resource being deleted * Use gsutil to download kube binaries and release * staging/publishing: remove release-1.16 rules * test images: Switches to buildx * Add new way to generate leader election lock * fix staticcheck vendor/k8s.io/metrics/pkg/client/custom_metrics/multi_client.go:49:4: ineffective break statement. Did you mean to break out of the outer loop? (SA4011) vendor/k8s.io/metrics/pkg/client/custom_metrics/versioned_client.go:38:2: var codecs is unused (U1000) * kube-proxy ensure KUBE-MARK-DROP exist but not modify their rules * Remove redundant variable * Switch events conformance metadata from v1.19 to v1.20 * Fix windows node startup failures. * CHANGELOG: Update directory for v1.18.10 release * CHANGELOG: Update directory for v1.17.13 release * kubelet container status calculation doesn't handle suddenly missing data properly * enable benchmark by matching test config * remove debug print from create quota * Move SCTP to GA * remove secondary client retries in e2e tests * Add support for create ingress in kubectl * Fix metrics reporting in kube-apiserver * Fix --dry-run invocation in kubectl e2e * fix: smb valid path error * Fix staticchecks in vendor/k8s.io/client-go * Replace non-ascii string under test/ * add methods to deallocate and starts vms in a scale set * Removing flaky tag from test * Use fluentbit as logging agent. Replace stackdriver * e2e: node: topomgr: avoid plugin leak on test fail * add a local tlsTransportCache to available_controller * Move dirExists() to kubelet_test * Map SelectorSpreadPriority to PodTopologySpread plugin * CHANGELOG: Update directory for v1.19.3 release * Fix fcpath * Ignore deleted pods. * Make copied `localtime` file write-able * Reorder default Filter plugins to have UnschedulableAndUnresolvable first * Add unit tests for dockershim/exec.go * test: Fix deprecated --dry-run parameter * Enable XFS tests for recent GKE COS versions * CHANGELOG: Update directory for v1.20.0-alpha.2 release * PriorityClass lifecycle tests * Fix log dumping for GKE * Add `create ingress` command to `cmd/kubectl` * Support the node label `node.kubernetes.io/exclude-from-external-load-balancers` * Remove dependency between create rolebinding * kube-eviction: Fix SI of process quantity * add systemd mount options interface to support the no-systemd mount * Move scheduler interface to pkg/scheduler/framework * Fix vsphere disk detach failing * - fix golint errors by adding defition comments - remove /pkg/apiserverinternal from .golint_failures * [auth]Check verbs for nonresourceurls and resources * add a OWNERS file under storageversion * generated * add the storageversion.Manager interface * kubeadm: add some output to the generate-csr command * Fix the busybox image tag in kubemark hollow nodes * fix azure disk attach failure for disk size bigger than 4TB * Add PULL_REQUEST_TEMPLATE.md and state the repo does not accept any direct changes * Run hack/update-vendor.sh * Add rule for component-helpers into staging/publishing/rules.yaml * Removing GetPodPriority from pkg/api and importing PodPriority from k8s.io/component-helpers * Duplicate a single helper from pkg/apis * simpler addition of nodeport basic validation * Replace Branch Manager/Patch Release Team with Release Manager * OWNERS_ALIASES: Add xmudrii to release-engineering-reviewers * make new command functions to keep private when they are not required to be public * wrap errors from NodeLabel, NodePorts, NodePreferAvoidPods and NodeResourcesBalancedAllocation plugins * fix azure disk data loss issue on Windows * fix upgrade test * Add defaultingType to PodTopologySpreadArgs * Don't depend on DNS in NetworkPolicy tests * kubelet: fix iptables setup under dual-stack * Improve logging of iptables canary test * Update addons volumesnapshots owners file * Clarify that we don't audit events due to performance impact * don't cache transports for incomparable configs * Update write-pki-data to give read permissions to KUBE_PKI_READERS_GROUP, for components running as non-root to be able to read the credentials. * Update to Calico v3.16.2 * Update Windows Pause version to 1.4.0 * Move cmd/controller-manager to k8s.io/controller-manager and cloud specific configs to k8s.io/cloud-provider. * API server: fix default_konnectivity_socket_path typo. * cleanup: fix golint errors in /pkg/kubelet/stats * cloud node controller: handle empty providerID from getProviderID * always collect containerd logs * Upgrade snapshot controller to 3.0.0 * log-dump.sh: Fix shellcheck issues * test-integration: Fix using deprecated default cluster IPs * Initialize k8s.io/kubernetes/staging/src/k8s.io/component-helpers * Update the example for kubectl port-forward * Upgraded aws-sdk-go to v1.35.5 to include more regions, i.e. Millan * kubelet: Set dual-stack hostNetwork pod IPs on dual-stack nodes * utilnode: fix incorrect documentation about node name vs hostname * kubelet: allow specifying dual-stack node IPs on bare metal * kubelet: Remove unnecessary sorting in NodeAddress tests * allocate service-account flowschema to global-default * Fix reporting network_programming_latency metrics in kube-proxy * Allow configuration of etcd healthcheck timeout * add ipv6 support to the image issue-74839 * set lastterminationstate for container status even when CRI fails to return termination (or any) data * Remove the dependency between create quota command and generators * make download-or-bust compatible with both sha512/sha1 * Fix documentation on EndpointSliceCache map * Adding marosset to test/e2e/windows/OWNERS approver * Add SELinux labels for kubelet on Fedora CoreOS * add CAP_NET_RAW capability to CRI stats summery test * e2e hostexec commands does not need sudo * Update after code review * Update cluster/gce/util.sh * Fix some shellcheck warnings/errors in cluster/gce/util.sh * Mask bearer token in logs when logLevel >= 9 * Update test/e2e/framework/network/utils.go add logging about min/max interval Co-authored-by: Antonio Ojea * scheduler framework: document how to make plugins configurable * Send node startup scripts to console and journal * Disable the Accelerator metrics by default * Adding Bazel dependencies * Adding failure scenario for long FQDN and setHostnameAsFQDN feature * add unit tests for network plugin manager metrics * clarify help for --api-version: note it's API group/version * include APIVersion in output of 'kubectl api-resources' * Move cloud related feature gates to controller-manager * Replace AreLabelsInWhiteList with isSubset * mark node-problem-detector as serial * Cleaned up and fixed: the potentially misleading comments around Event struct * Update DialFromNode to return values as is done w/ other tests. Update comments to clarify missing probability check function (since they need to be updated anyways b/c of the return value introduced) * Azure: fix node removal race condition on VMSS deletion * removed links to soon deprecated helm stable repo & added artifact hub link * get kibana up to speed with es * rev version of stateful set * v1.18 support changes * kubectl: Update triage/support label references to kind/support * Fix golint failures in pkg/registry/core/replicationcontroller * kubelet: do dual-stack iptables rules * use new fluentd image in daemonset * updated fluentd and all plugins * fix staticcheck for kubectl pkg files * Add gate to install CSI proxy * SetHostnameAsFQDN will be beta in v1.20, enable feature gate by default. * test: e2e: fix race in pods test * Fixes high CPU usage in kubectl drain * proxy: Add tests for kube-proxy config defaulting * test/e2e/framework/:use the term 'Control Plane' in comment * Do not update managedFields timestamp when they don't change * reconcile NLB attributes on service creation * fix detach azure disk issue when vm not exist * Mask Ceph RBD adminSecrets in logs when logLevel >= 4 * Update triage/support label references to kind/support * Update addon-manager makefile to use the new staging repo - for real * Remove unused FailfWithOffset() * Verify pod termination with E2E PreStop hook * Move predicates and priorities configuration creation to Policy mapping * Make the creation of namespace using POST and PATCH consistent * vsphere: improve logging message on node cache refresh event * Tag LabelSelector with +structType=atomic * Bump to latest kube-openapi and SMD to pick up structType=atomic support * just log essential pod info * Add error text to kube-system wait error * Change kubectl clusterinfo to non offensive words * e2e can't use both pod.Name and pod.GenerateName * Fix test "[sig-windows] DNS should support configurable pod DNS servers" * Fix lint errors in pkg/contoller/endpoint * Update Makefile to point to the new addon manager gcr repo * Fixed percentage behavior in instr * remove dns_common.go delete configmap stuff * kubelet: remove some redundant iptables option checking * Remove federation tests entirely because they break clusters and aren't relevant anymore * Don't add empty AZ labels to OpenStack volumes * resolve PR comment ( add nil check ) * fixes test/integration/ttlcontroller staticcheck * Add CSI proxy log to fluentd * Update Windows image patch version * Updated symbol 'framework.GetAllMasterAddresses' to not use word 'master' * Updated symbol 'framework.GetMasterHost' to not use word 'master' * fix: use ">" instead of ">=" in resource allocation * add configuration for controller migration. * Watch bookmarks may contain version of objects of other types * Support sharing one IP address for multiple services * Add unit tests for feature "Support sharing one IP address for multiple services" * Graduate SelectorIndex to GA * correct 'information' spelling errors * correct 'admission' spelling errors * add note for kubeadm cert renew * Fix UpdateSnapshot when Node is partially removed * Add LookForStringInPodExecToContainer that takes container name parameter, modify LookForStringInPodExec to call the new function. * Fix test name TestDump * kubeadm: deprecate self-hosting support * make update * fix integration tests * fix error message * fix goroutine that lives too long * wrap errors in selectorspread and podtoplogyspread plugin * add unit tests for getStorageAccounts in azure_storageaccount.go * Lint ttl_controller * Fix staticcheck failure for vendor/k8s.io/client-go/discovery/cached/memory * kubectl: deprecate --delete-local-data * wrap errors in service affinity plugin * [pkg/watch/json]: remove dead code * wrap errors in taint-toleration plugin * pause image: moves wincat binary location * Adds support for building Windows pause image * tests: Refactors agnhost image pod usage - network * tests: Refactor agnhost image pod usage - common (part 1) * gce: redirect handshake server requests to metadata-concealment too * Disable one subpath test for windows * Add csiproxy log * Collect debugs before framework.Failf * Enable per-zone logexporter pods creation * e2e/storage: disable caching when writing/reading to block PVCs * e2e: add option to create pods with different image in pod.Config * Clean up nits in delete cascade * wrap errors from DefaultPreemption, ImageLocality and NodeAffinity plugins * Fix LookForStringInPodExec(...) to take container name as argument, and refactor function usage Ref: PR 92127 review discussion https://github.com/kubernetes/kubernetes/pull/92127/files#r447853904 * switching to stable version * Fix kube-addon-manager overwriting resources with EnsureExists * apiserver self request metric * address instance_test nits * It's an 'Instance' of apiserver * upgrade test for BoundServiceAccountTokenVolume * 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 * using structured logging in scheduler framework runtime * agnhost image: use actual DNS domain instead of hardcoded cluster.local * add aojea to test reviewers * Use the exported struct member directly * fix: correct glbc ClusterRole * Updated vendor by `hack/update-vendor.sh` * Update bazel * Update dependencies * Remove kubeconform * Remove dependency on behaviors for conformance generation * Fix typo & documentation on kustomize.md * Remove unmaintained stackdriver logging e2e test * replace sha1 with sha512 * Use pager's context instead of TODO * kubelet: add unit tests for imagePullSecrets keyring * Keep the imagePullPolicy of kubectl alpha debug consistent with the default * Remove ext2 + ext3 tests * Add failure logic to tests that rely on side-effect-free dial functions. * Adding back in the breadth-first-polling logic. "Revert "Merge pull request #93837 from jayunit100/DialFromContainerB"" * Update the use message for kubectl debug * Graduate kubeadm alpha certs command * Fail test when Cinder volume deletion fails * Add extra log when Cinder volume deletion fails * Delete pre-provisioned Cinder volume by ID * deprecate scheduler metrics BindingLatency and SchedulingAlgorithmPreemptionEvaluationDuration * modify static check * Replace e2essh on e2e service tests * using kubectl apply to create metric adapter * Allow the lifecycle of kube-proxy to be managed independently of the startup scripts for GCE * Fix `kubectl describe ingress` format * check readyz before adding endpoint for kube-apiserver * Bump network proxy images to v0.0.12 * Bump konnectivity-client to v0.0.12 * tests: Refactors agnhost image pod usage in tests * Revert "Merge pull request #93837 from jayunit100/DialFromContainerB" * CHANGELOG: Update directory for v1.20.0-alpha.1 release * Move Kubelet Summary API to staging repo * removing deprecated scheduler metrics * refactor(kubeadm): make `alpha kubeconfig user` command accpet --config * wait until the iptables rules are programmed * Add test for listing Leases from all namespace * use GetInstanceProviderID to get instance provider ID * Remove mattjmcnaughton as a sig-node reviewer * Enhance the prompt information of verifyRunAsNonRoot, add pod, container information * Update nodelocaldns yaml to use 1.15.14 image * handle longer vendor paths without go files * Show error in status if preserve unknown fields is true for nonstructural schemas * `find . -type f \( -name "*.go" -or -name "*.md" \) -print0 | xargs -0 gsed -i 's/the the /the /g'` * endpoinslices must mirror services labels * kubectl: add a space between effect and operator when printing tolerations * Return the Kubernetes version which stopped serving deprecated APIs by default * log-dump.sh: Do not modify logexporter-daemonset.yaml in-place. * Take into account latency incurred in server filters * test images: uses nanoserver * Remove kubeadm audit package * Do not assume storageclass is still in-tree after csi migration * fix kube-proxy cleanup * scheduler_perf: use time.Ticker in throughput measurement * Remove testcluster check for csi proxy * Add cheftako to KCM owners. * Add an unit test for requests including value after token * Limit the max number of splitting * fix: detach azure disk broken on Azure Stack * [pkg/api/podsecuritypolicy]: fixup typo * Add a check for crio service before starting node e2e tests * apiextensions: prune array type without items in published OpenAPI * test: add unit-test for TranslateCSIPVToInTree. * apiserver: fix healthz vs. livez vs. readyz log output * test(kubelet): deflake TestRotateShutsDownConnections * kubeadm: update vendor to exclude kustomize import * kubeadm: remove the --experimental-kustomize feature * update-bazel * use more granular buckets for azure api calls * fix duplicate testcase names * Adds filtering of hosts to DialContexts. * update storageos vendor for FilteredDial change * kubeadm: remove the --kubelet-version flag for "upgrade node" * Adding cheftako to CCM owners. * scheduler_perf: refactor to allow arbitrary workloads * Tests for empty constraints array when DefaultPodTopologySpread is enabled * cluster/images/etcd: Build etcd:3.4.13-1 image * build: Update to debian-base:buster-v1.2.0 * build: Update to debian-iptables:buster-v1.3.0 * test(apply): deflake run_kubectl_apply_tests * fix azure file migration panic * Add support for s390x * test images: Fixes echoserver s390x image * Change code to use staging/k8s.io/mount-utils * enable gce InstancesV2 * implement gce InstancesV2 interface * Update csi proxy to v0.2.1 * cloud node controller: implement with workqueues and node lister * CHANGELOG: Update directory for v1.19.2 release * CHANGELOG: Update directory for v1.18.9 release * CHANGELOG: Update directory for v1.17.12 release * test/e2e: Busybox image is not being templatized * move dashpole to emeritus in kubelet * LockToDefault the ExternalPolicyForExternalIP feature gate * Fix resource location for ipv6 pods * Refactor kubectl without stdin test * e2e sctp support for ipv6 * Fix e2e autoscaling namespace error * reduce cloud api calls in cloud-node-controller by passing instanceMetadata to updateNodeAddress * Ensuring EndpointSlices are recreated after Service recreation * Increasing acceptable timeout for EndpointSlice garbage collection * Delete namespace parameter in create adapter * verify-generated-swagger-docs: remove unnecessary build * kubeadm: relax the validation of kubeconfig server URLs * Graduate WinOverlay to Beta * count of etcd object should be limited to the specified resource * Run gofmt * run hack/update-vendor.sh * Move podresources api to k8s.io/kubelet/pkg/apis * optimize the use of informer for scheduler * Revert "conntrack e2e test debug connections" * ingore apparmor on non Linux operating systems. * test: ensure WaitForCacheSync is called after starting sharedInformerFacotry * Move the RuntimeClass tests out of node-kubelet-orphans * Test watchcache being updated in multietcd setup * Enable progress notify events in watchcache * Allow tracking resource version for reflector store * Implement etcd3 progress-notify feature in etcd3 layer * Pipe newFunc to etcd3 storage layer * Update kubeconfig command-line help message for kube-proxy * Clean up remaining ns flag * Revert "Revert "Switch cos version to M85"" * ubernetes_lite.go: remove image argument from SpreadServiceOrFail * Don't attempt to detach an FC device if we don't know its name * follow up cleanup after SupportPodPidLimits GA * Mount kubelet and container runtime rootdir on LSSD * avoid potential secret leaking while reading .dockercfg * clean up podpreset deprecated client * generated * remove generated podpreset * remove pod presets * test/{images,utils/image}: Add justaugustus as reviewer * cluster/images: Add justaugustus as reviewer * hack: Add justaugustus as reviewer * .github: Add justaugustus as reviewer * kubeadm init phase upload-certs: add flag --kubeconfig * portforward: Fix UDP-only ports calculation * chore: add network rule support in Azure account creation * Fix misusage of RLock in timeCache lru.Cache.Get() * e2e test support microk8s * test(iptables): deflake TestRestoreAllWaitOldIptablesRestore * feat(iptables): be able to override iptables-1.4-compatible lock path * test flake: fix data race in csi_test.go * Move podPassesBasicChecks() to VolumeBinding plugin * Remove pvcLister from genericScheduler * add retry for creating metrics grabber to fix test flakes * Remove tests related to HTTPS support for ingress-gce * 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) * Graduate ServiceAccountIssuerDiscovery to beta * avoid logging token in RunDeleteTokens * gpu device plugin uses EnsureExists addon mode * Non-zero cfs quota period duration requires feature flag * Use namespace flag passed to RunKubectl* methods * Update etcd to dd1b699fc489 * Fix some shellcheck warnings/errors in cluster/gce/util.sh * kubeadm: remove stray "alpha phase" command * kubeadm: remove the "alpha kubelet config enable-dynamic" command * Fix index out of range panic for kubectl alpha debug * rename some files * Fix minor comment in a script * Wrap errors on VolumeBinding plugin * Wrap errors on DefaultBinder plugin * Wrap errors when running Bind plugins * Wrap errors when running PreBind plugins * Remove field disablePreemption from internal scheduler codebase * Run unit tests 2 instead of 3 times via bazel * Add the storageversion.Manager interface * pillar_dns_server remove change so as not to break tooling in the near term in the kubernetes/dns repo * Fix staging/src/k8s.io/apimachinery/pkg/labels golint findings * Use EphemeralContainers for storage validation * Hold error in framework's Status * Increase time it takes for second pod to startup for offline resizing * Read PV object from apiserver to prevent flake * Migrate scheduler, controller-manager and cloud-controller-manager to lease lock. * fix the bug that kubeadm tries to call 'docker info' even if the CRI socket was for another CR * Update csi-proxy to use beta version * remove feature gate SupportIPVSProxyMode. * Revert "Switch cos version to M85" * bump the version of k8s.io/system-validators to 1.2 * kubeadm: Update versions for 1.20 * Handle nil elements when sorting, instead of panicking * kubeadm: print warnings on invalid cert period instead of erroring out * CHANGELOG: Update directory for v1.19.1 release * Prevent deletion of namespace again * Deprecate Dockershim * Tolerate NotFound errors when deleting snapshots * storage E2E: explicitly wait for PV deletion after ephemeral test * Add vCenter info metric * Avoid unnecessary calls if other error occurred * test(portforward): deflake TestGetListener * Add roundtrip tests for metrics repo * generated * Let kube-apiserver host the storage version API * fixup: add podLister as a member field of DefaultPreemption * Preemption plugin to fetch pod from informer cache * Switch cos version to M85 * generated * fix import paths * add internal.apiserver.k8s.io/v1alpha1 to known gvs * update doc.go and register.go * move apiserverinternal types to kube-apiserver * Initialize scheduler's podInformer in sharedInformerFactory * Sort list of formats for --logging-format description to make it deterministic * skip TestClientReceivedGOAWAY * Count storage requests as whole byte values in quota * Add Stephen Augustus as Reviewer for staging/publishing * e2e: fix deployment non-unique env vars to avoid SSA error * Remove pkg/api/endpoints * Stop container before remove for Docker * fix golint failures for staging/src/k8s.io/client-go/examples/workqueue * Revert "fix cluster/log-dump/log-dump.sh shellcheck failures" * deferredResponseWriter returns after calling the Close() method * test(tools::events): deflake TestEventSeriesf * Fix typo in comment of hack/verify-shellcheck.sh * Fix staticchecks ST1005,S1002,S1008,S1039 in pkg/kubelet * Remove unnecessary double-pointer * Add metrics for azure service operations (route and loadbalancer). * Add tests for daemonset view history * Add commend for printHistory function * test(watch-tool): deflake TestRetryWatcherToFinishWithUnreadEvents * dedup the printHistory logic in DaemonSetHistoryViewer,StatefulSetHistoryViewer * Describe sts on rollout history if the revision * kubectl: Use Fprintf * fix golint for pkg/volume/azure_dd * test(workqueue): deflake TestMetrics * build/lib/release: Explicitly use '--platform' in building server images * build/common.sh: Remove extraneous reference to debian-base image * test(kuberuntime): deflake TestRecordOperation * prefer NoError/Error over Nil/NotNil * Deflake serving options, avoid hard-coding ports * Deflake port forward tests to avoid hard-coded local ports * Deflake cpumanager checkpoint unit tests * Use unique socket name per cm test * fix(azure::cache): TimedCache.Getter should be called once on the same key * test(azure::cache): deflake TestCacheNoConcurrentGet * Run slow kubeadm upgrade tests in parallel * Ensure kubeadm tests have unique names * Avoid mutating global variables in kubeadm certs phases * test(iptables): deflake TestRestoreAllGrabOldLock * Skip TestGOAWAYConcurrency * test(apiserver): deflake TestClientReceivedGOAWAY * Deflake TestSetup * Deflake TestHTTPProbeProxy * kubelet: remove alpha warnings for CNI flags * Add more tests for LRU cache lookup * Move ResourceQuota admission to k8s.io/apiserver * do not mutate endpoints in the apiserver * remove DefaultIngressClass feature gate for 1.20 * Cleanup custom metrics conversion functions * feat(kubeadm): specify resource requests in etcd pod spec * Stop ignoring unit test flakes, require multiple successful passes * kubeadm: remove the CoreDNS check for supported image digests * Use ExpectNoError(err) * allow to map the same container port to different host ports * Updating winkernel kube-proxy OWNERS file * consistently use double quotes in proto files * Deflake TestServiceRegistryExternalTrafficHealthCheckNodePortUserAllocation * add more testcases for GOAWAY filter * prefer double quotes and consistent usage in .proto files. * Improve ability to debug Events e2e failure * Added support for register-with-taints setting #93608 * AWS NLB/ELB health check config based on service annotations * kubeadm: adjust the logic around etcd data directory creation * fix vendor/k8s.io/apimachinery/pkg/runtime/serializer/versioning staticcheck * test(csi): deflake TestAttacherWithCSIDriver * Remove HeadlessService label in endpoints controller before comparing * Remove default conversions * Update security contacts for sig-cli * Allow to disable logrotation of kubernetes and pod logs * Fix some shellcheck warnings/errors in cluster/gce/util.sh * Fix some shellcheck warnings/errors in cluster/gce/util.sh * Fix some shellcheck warnings/errors in cluster/gce/util.sh * Enhancement on the testcase to cover more possibilities * fix kubeadm update coredns with skip pending pod * address review comments * Update kubectl to use new vendored github.com/fvbommel/sortorder dependency * Fix broken dependency vbom.ml/util * Add exp backoff for connection refused errors * proxy/userspace: clean up and consolidate testcase setup * proxy/userspace: use waitgroups instead of sketchy atomic ops in testcases * Start CSI proxy as service * remove --verbose_failures from .bazelrc * fix a few missed references * CHANGELOG: Update directory for v1.16.15 release * update scripts * fix doc file * rename * partial * move port definitions to a common location * remove iperf e2e test ip family tags * modify DNS-1123 error messages to indicate that RFC 1123 is not followed exactly * kubeadm: Fix `upgrade plan` for air-gapped setups * Avoid the API call to update pod if nothing is changed * moving files from k8s.io/util/mount into staging/src/k8s.io/mount-utils * fix: return error with fewer mount options on Windows * fix: smb remount issue * Remove driver letter assignment during volume format * Revert "Merge pull request #166 from jingxu97/May/drivename" * chore: add more logging for mklink on Windows * fix: remove unnecessary readlink check in IsLikelyNotMountPoint on Windows * Remove driver letter assignment during volume format * Switch to klog v2 * ignore golint for some stutter that we want to keep as-is * fix bad spelling * Revert xfs_repair fix * Fix windows MountSensitive error * Fix subPath mountpint check * Fix mount_windows build error * Introduce paramater for sensitive mount options. * FormatAndMount unit test only checks for MountErrorValue now and closed gaps for some error values * Add more detailed error output when disk formatting fails * Split MajorMinor into two fields * update after review * use xfs_repair to check and repair xfs filesystem * Return typed error when Mount Fails * Validate the existence of filesystem before attempting to mount it (linux) * feature: implement Output method for FakeCmd * Fix golint errors * Update doc.go to show k8s.io/utils * Extract pkg/util/mount and drop BUILD * bind metadata proxy to 0.0.0.0 * Add e2e multiple endpoint services test * spdy: add optional periodic Pings on the connection * fix vendor/k8s.io/apimachinery/pkg/conversion staticcheck * Handle eviction of pods in deleted namespace * drain: eviction creates in a deleting namespace will throw a forbidden error * Fixed reflector not recovering from "Too large resource version" errors with API servers 1.17.0-1.18.5 * Stop setting SelfLink in kube-apiserver. * Fix FakeClock::Reset to always succeed * add myself to sig-node-reviewers * kubeadm: make the scheduler and KCM connect to local endpoint * kubeadm: Ensure etcd manifests are regenerated upon upgrade * Add WatchListPageSize to cache.Config * Allow 404 error on lb deletion in azure * add lock for csi node update * fix typos in cmd/kubeadm * clean up testutil/metrics content * Update comments in pkg/scheduler/framework/v1alpha1/interface.go * Update `kubectl diff --server-side` test. * Skip TestLoopbackHostPortIPv6 run on non-IPv6 env * relax matcher for ResourceMetricsAPI test * fix staticcheck errors in resourcequota * update test to match validation filter of pods * Omit optional field when empty * Update CNI plugins to v0.8.7 * fix golint failures * Mark SCTPConnectivity tests as disruptive. * Add tests for SCTP network policies. * Ensure getPrimaryInterfaceID not panic when network interfaces for Azure VMSS are null * Improve the prompt message when kubeadm init is successful * fixed golint error in pkg/apis/admissionregistration * Add 1.19.0 API testdata * Remove 1.17.0 API testdata * chore: 📦 update generated files * Update max data disk count with new instance types * Getting rid of the Salt DNS replacements, addded / back. * tests: Refactor agnhost image pod usage * Improve docs for client-go warning handlers * Override hostname with instance name * Fix testpattern check * chore(openapi-spec): 📦 regenerate * Moving e2e tests to e2e_node directory * Combine switch case into one case * Ran hack/update-bazel.sh but discarded changes to root build file * Move common portion of dynamic snapshot out of switch statement * Remove custom hash function, generate a random UUID for snapshot and snapcontent name instead * Stack snapshot tests into one to reuse snapshot resource and reduce time taken on prow. * Quick commit * Remove snapshot class from prepv test because it is not needed. * Combine two tests with overlap to reduce overall snapshot test suite time * Add pre provisioned snapshot tests * start kube-apiserver and webhook server only once to shorten the webhook audit test time * Minor cleanup in snapshot test suite. * Add a link to liggitt's deflake docs in Flake template (#94243) * Deprioritize extensions/v1beta1 in discovery * Update cri-tools to v1.19.0 * Update default etcd server to 3.4.13 * fix(azure): check error returned by scaleSet.getVMSS * fix cluster/log-dump/log-dump.sh shellcheck failures * Remove azure-load-balancer-disable-tcp-reset annotation * Remove duplicate nodeSelector * fix linter issues for pkg/kubelet/cri/remote and staging/src/k8s.io/cri-api/pkg/apis/testing * Fail early on stress test and improve logging * kubectl wait add ignore not found flag * Support kubectl delete foreground * remove deprecated kubelet endpoint /metrics/resource/v1alpha1 * Update options.go * Initialize candidate directly instead of iterating the array of candidates * Move brendandburns to to emeritus status. * fix vendor/k8s.io/apimachinery/pkg/apis/meta/v1/unstructured/unstructuredscheme staticcheck * only update Azure data disks when attach/detach * Enable IP Tags on Azure LoadBalancer public ips * Add PR #89069 Action Required * Update etcd makefile to build 3.4.13 image * Update etcd client to v3.4.13 * update drain message to include namespace for pod in kubectl get compatible format * CHANGELOG: Update directory for v1.19.0 release * Fix issue on skipTest in storage suits * Remove file_mode args for Windows test * Update Calico to v3.15.2 * Set snapshotType for tests with NTFS. * Update CHANGELOG-1.18.md * Add logging if container or pod fails to be killed * Adding label NodeAlphaFeature to include tests in Node Testgrid * Add impersonated user to system:authenticated group * Verify running mirror pod has running containers * Only process all nodes when incoming pod has no preferred affinity * test(endpointslice): deflake TestSyncEndpoints * Remove unnecessary conversion * Detect change of volume attachability in the middle of attaching * Promote PidLimits to GA * Ensure backoff step is set to 1 for Azure armclient * [pkg/registry/admissionregistration]: fixup wrong copy&paste godoc * Remove StartupProbe featuregate and related logic * Make ProjectedVolumeSource optional * Track pods with required anti-affinity * Updating kube-proxy to trim space from loadBalancerSourceRanges * let panics propagate up when processLoop panic * Allow to configure clustername in creation of kubeconfig using kubeadm command * Make similar buckets for api and etcd request duration histogram * .github: remove stackoverflow from support issue template * Refactored pkg/controllers/resourcequota * fix: use sensitiveOptions on Windows mount * Cleanup kubelet TODOs that are no longer pertinent. * kubelet, cgroupv2: do not create /sys/fs/cgroup/sys with cgroupfs * Update wait function w/ descriptions of global node states, TODO for kubelet API. * ARM client: survive empty response and error * delete tests which use container manager stub * tests: Use Windows absolute path on Windows pods * Only lock statuses map when status isn't successful * Do not allow manual changes to manageFields via subresources * Change nodeInfolist building logic in scheduler * do not create StatefulSet pods when PVC is being deleted * fix misleading comment in expand_volume.go * Simplify and de-lint GCE Windows kube-up docs. * Add context to x509 verify failures * Read ssh username from env variable * Fix the logging message * Use /usr/bin/env bash in Makefiles and scripts * Scheduler auth stop using legacy scheme * allow adding annotations to pod when using kubectl run and add tests * kubeadm: Separate argument key/value in log msg * update ResourceMetricsAPI node-e2e test * docs(api): 📝 Fix HPA docs with addition of "External" type * fix vendor/k8s.io/apimachinery/pkg/util/framer staticcheck * Removed broken link to Analytics * Make CSI mock driver log parse more flexible * Verify that an ingress with empty TLS is valid * Update client-go out of cluster example * Update yaml files to use seccomp GA syntax * go fmt fix * remove some notes about scheduler/algorithm * get: -o yaml, json set ServerPrint false * Add test for `pkg/kubelet/util/util_windows_test.go#GetAddressAndDialer` * back out conversion of DialFromNode->DialFromTestContainer * kubelet: assume that swap is disabled when /proc/swaps does not exist * e2e storage: skip multi-volume generic ephemeral volume test * Reduce test volume sizes * support ipv6 in e2e policy tests * modify the warning log format from %d to %v * Update snapshot controller to use k8s.gcr.io * Seperate registry and feature gates * names unnamed testcases, adds name to t.Errorf msg * Improve running time of TestSchedulerWithVolumeBinding * Add maxTries logging statement, otherwise the numbers are really hard to interpret * implement breadth first try * fix test/e2e_node staticcheck * Fix error messages * Fix Poll variable name * Fix waiting for PVCs to get Bound * Number of failed jobs should exceed the backoff limit and not big equal. Remove patch in e2e test of backoff limit due to usage of NumRequeues * fix vendor/k8s.io/apimachinery/pkg/apis/meta/v1/unstructured staticcheck * Fix some typo * golint * fix vendor/k8s.io/apimachinery/pkg/labels staticcheck * Enable completion for `kubectl config delete-context` * vsphere: remove inactive members from OWNERS * sig-network: remove inactive members from OWNERS_ALIASES * cluster/addons/dashboard: remove inactive members from OWNERS * apiserver: remove inactive members from OWNERS * Fast return when no any matched anti-affinity terms * Update seccomp e2e test for GA * Fix an error in NodeUnschedulable plugin comment * Format Codes * minor logic fix * e2e test intra-pod breadth first logging and summary * fix incorrect comment in runtime ObjectKind interface * Fix job's backoff limit for restart policy Never, rely on number of failures instead of number of NumRequeues * test(kubelet): add a regression test to verify kubelet would not panic * fix(kubelet): protect `containerCleanupInfos` from concurrent map writes * refactor(*): update pvc quota name typo in comment * fix: azure disk resize error if source does not exist * chore: add diskclient.Update interface * chore: add diskclient.ListByResourceGroup interface * Remove false positive warning in kubeadm cmd * cloud provider: add zone/region to InstanceMetadata * Sort kubectl top output when --sort-by and --containers are used together * add testcases for kubelet getters * fix a typo in the comment * Use v1helper.GetPersistentVolumeClass for compatibility * Add tests for managed fields tracking. * leaderelection: Remove doubled space in "attempting to acquire leader lease" log * benchmark-dockerized.sh: set KUBE_ROOT * kubeadm: reset don't unmount /var/lib/kubelet if it is mounted * PodReasonUnschedulable is not a pod condition type * Fix kubelet log message when starting a container * kubectl describe pod: use ReportingController as an event source * Fix TestScorePlugin: numScore should be accessed with atomic.LoadInt32 * expose Run[Pre]ScorePlugins functions in PluginRunner interface * Cleanup wait forever loops in pv_controller_test.go * Strip unnecessary security contexts on Windows * Updating the Reviewers / Approvers for WinKernel Proxier * fix the remote endpoint cleanup logic * Add chendave to sig-scheduling REVIEWERS * Adjust default replica count to 2 replicas * Run make verify with python3 to fix publishing bot issue * Document blocking behaviour of RunOrDie and Run * removed k8s-master label * Add example of using resource builder to load a manifest file * tests: Fixes variable expansion false positive test * revert auto-completed package names * remove legacy leftovers of portmapping functionality that was moved to CNI * mark this test serial due to race conditions * Clean up daemonset test names * Remove phantoms from dump_requests output * deps: opencontainers/go-digest -> v1.0.0 * Update k8s.io/utils dependency and use ebtables from there * kubectl: Allow --patch-file to be passed to `kubectl patch` * reduce type conversion and correct variable name * Fix verbs reporting in kube-apiserver metrics * kubectl: improve the warning message when doing kubectl apply to the resource without expected annotation * Fix label reference in deployment checks * Use NLB Subnet CIDRs instead of VPC CIDRs in updateInstanceSecurityGroupsForNLB * Update readiness checks * Update ready condition * chore: upgrade Azure/go-autorest to v14.2.0 use autorest v0.11.1 * extend ShouldCallHook benchmark to verify performance imporvement * [kubectl] Fail when local source file doesn't exist * nodelocaldns.yaml: remove force_tcp option which used for external dns query * Create Deployment resource lifecycle test * apf integration test: ensure no rejection * skip mismatched webhookAccessor and object * return err directly when nodename is not consistent in cert * add bootstrap policy for monitoring roles * unit test * move to cadvisor.MachineInfo * Delete ILB FR in case of changes to port/proto. * fix windows container root validate * hack: ensure go version is not specified for master in publishing rules * publishing: remove go 1.13.9 for master branch for controller-manager * add test case for kubeadm memory check * add more testing cases * Shorten watch restart test, run in parallel * ipvs: log error if scheduler does not exist and fallback to rr * fix golint failures in pkg/kubeapiserver/options, fix some incorrect replace of receiver name * fix golint failures in pkg/kubeapiserver/options, rename receiver name of BuiltInAuthorizationOptions to o * fix golint failures in pkg/kubeapiserver/options, use API Server in commemts instead of APIServer * e2e node: fix kubelet service restart failure * Mention background deletion in kubectl delete cascade help * fix bazel build file * correct the sandboxId attribute in unit tests * address review comments * add sandbox deletor to delete sandboxes on pod delete event * Check for sandboxes before deleting the pod from apiserver * fix(kubectl): warn users that flags cannot be placed before plugin * Avoid using socket for hints * regen * update build * remove serializers from codec factory * update * clean up unused var containerCache * cleanup: print warning message after timeout * Revert "cleanup: decrease log level from warn to v3" * cleanup: decrease log level from warn to v3 * AWS: increase io1 volume IOPs limit * Add a preflight check that the control-plane node has at least 2GB RAM * Handle int -> float conversion in FromUnstructured * Fixed potential nil dereference in hostpath unit test * Fixed error string should not be capitalized * add GetAddressAndDialer unit test * [k8s.io/apiserver/pkg/server/filters]: fixup go vet warning * [kubelet/dockershim] : Use local copy for range iterator * fix(staticcheck): fix static error of package `/pkg/volume/gcepd` * Filters on kubemark nodegroups when selecting node template to enable hollow node clusters that contain real nodes. * Move MasterUpgrade() out from e2e framework * Use temporary directory when creating temporary file in tests * generated * add vendorless to client-gen * use canonical import path in client-gen so that package comments are * ipvs: ensure selected scheduler kernel modules are loaded * Fix NetworkPolicy describe for egress-all policies * Export WaitForCSIDriverRegistrationOnAllNodes to be used by external csi driver repos * [e2e/storage] fix range issue in getCSINodeLimits * kubelet: add operations count and error count metrics to network plugin manager * fix HandleCrash() order * fix golint failures in pkg/kubeapiserver/options * remove duplicate path import * AWS cloudprovider allow nlb-ip and external type lbs to be managed externally * Move WatchUntilWithoutRetry() from e2e framework * Add metav1.SetMetaDataLabel func * Remove resize map code * lock-free broadcaster, use chan to ensure thread safety * [test/kubelet]: Fix deadlock in watch manager test * kubectl: remove the dependency between autoscale command and generators * cluster/addons/metadata-proxy: cleanup inactive members from OWNERS * test/images/pets: remove OWNERS to cleanup inactive members * apimachinery: remove inactive members from OWNERS * [pkg/controller/certificates]: remove staled func comments * Remove DeprecatedMightBeMasterNode() * TestCreateInvokeAdmissionControl: remove unnecessary goroutine in sequential processing * Add unittest coverage for boottime_util_linux.go * dual-stack: make nodeipam compatible with existing single-stack clusters when dual-stack feature gate become enabled by default * Adds additional documentation for job status * Set the oom_score_adj of guaranteed pod to -997 * Resolves staticcheck failures for component-base/metrics * Change the node name from "machine" to "node" * Fix a typo in PreBindPlugin comment * Replaced repair with returning error, in delta_fifo.go * Fix description of conversion generator * Don't create a new sandbox for pod with RestartPolicyOnFailure if all containers succeeded * add defer * SIG node owner files clean up * Refactoring: Reduce unnecessary lines * Fix a typo in comment * code clean for podgc * fixed e2e vsphere statefulsets test * Repair instead of panic when data corruption detected in DeltaFIFO * conntrack e2e test debug connections * disable goffuzz in binaries * Add get-users and delete-user to kubectl config * fix typo in runtime/interfaces.go * Return nil as error when instance is not found so that node_controller could delete the node * Cherry pick the fix https://github.com/kubernetes/kubernetes/pull/70291 * Find what fails pull-kubernetes-e2e-gce-ubuntu-containerd * Review update * Sync with master * Remove previously added '' no longer needed * No quotes needed/wanted for CURL_RETRY_CONNREFUSED * Add fix for run-kube-controller-manager-as-non-root * Updates after review * Update cluster/gce/gci/configure-helper.sh * Update cluster/gce/gci/configure-helper.sh * Update cluster/gce/gci/configure-helper.sh * Update cluster/gce/gci/configure-helper.sh * Mitigate newly added shellcheck issues * Updates after code review * Update after code review * Code review update * Fix shellcheck w/e in cluster/gce/gci/configure-helper.sh * changes in imports was unintentional * fix S1000 simplify ch switch cases * fix S1000 simplify ch switch cases * fix S1008 simplify bool returns * fix S1002 omit comp to bool * Warn when creating roles with custom verbs kubectl * Fix static check failures for staging/src/k8s.io/apiserver/pkg/server/healthz/healthz_test.go * cleanup tempfiles in unit test * Enhance apiextensions-apiserver in standalone mode * fix CHANGELOG typo * Mirror pod without OwnerReference should not be created * Add unit test to kubectl/cmd/run * remove stale TODO after this PR: #92204 * add line break when describe hpa * basic regression test of runDockershim * fix func name NewCreateCreateDeploymentOptions * Fix error check logic * Fix broken format in CHANGELOG-1.18 * Remove contemplation of invariant violations from delta_fifo.go * fix expanding rbd volumes without ceph.conf * Adding tests for Kubelet pod update functions * Update Calico cpva to v0.8.3 * cleanup: remove useless methods * Modify the kubelet document * fix golint errors for pkg/kubelet/dockershim * cleanup: no need nil check before range * Add lease release tests in leader election * Generate complete leader election record to resolve leader election issues with LeaseLocks * fix: comments(only create parent dir) * make kubectl/kubeadm completion script support busybox * Add node-local-dns headless service * Made containerd-namespace flag can be used when kubelet config start cmd * fix comments * fix: ignore dir check in csi node stage/publish * Link to krew's new project site * Remove --include-uninitialized flag documentation * Use canonical URL for the krew project * return success if cordon node by replace * kubectl/drain add support for custom pod filters * do not wait for node ready when starting kube-proxy in 'nokubelet' mode * validate KubeletCgroups and KubeReservedCgroup- Initial Package/bin/sh/bin/sh/bin/sh/bin/shs390zp29 1621380355 1.20.21.20.2-lp153.1.211.20.2-lp153.1.21 kubernetes1.20manifestskubernetes1.20kubeletkubelet.servicekubelet.service.dkubelet.confrckubeletkubernetes1.20-kubelet-commonCHANGELOG-1.20.mdCONTRIBUTING.mdREADME.mdsysconfig.kubelet-kubernetes1.20kubernetes1.20-kubelet-commonLICENSEkubelet.1.gzkubeletvolume-plugin/etc//etc/kubernetes1.20//run//usr/bin//usr/lib/systemd/system//usr/lib/tmpfiles.d//usr/sbin//usr/share/doc/packages//usr/share/doc/packages/kubernetes1.20-kubelet-common//usr/share/fillup-templates//usr/share/licenses//usr/share/licenses/kubernetes1.20-kubelet-common//usr/share/man/man1//var/lib//var/lib/kubelet/-fmessage-length=0 -grecord-gcc-switches -O2 -Wall -D_FORTIFY_SOURCE=2 -fstack-protector-strong -funwind-tables -fasynchronous-unwind-tables -fstack-clash-protection -gobs://build.opensuse.org/openSUSE:Leap:15.3/standard/3bef44ebbdd15b709d57f3ef066aeb75-kubernetes1.20cpioxz5s390x-suse-linuxdirectoryPOSIX shell script, ASCII text executableASCII textexported SGML document, UTF-8 Unicode text, with very long linestroff or preprocessor input, ASCII text (gzip compressed data, max compression, from Unix)R;yܕ;3>{Iutf-8ad39dd3528734d7484138f63e2e760de48658eefd2b79e13b119d3d619f58796?p7zXZ !t/] crt:bLL ŔS@CRct1r8ΫJzP2] ϮHM#bh8v-B7Fã/\xht߄e@$JnJyzno0Ր9svu+:د7!1‚UӐTv}Pj_iB6k `SKpt_8x ERE6;8#f+,ֈTW*w͆W/PԽc_JMj%uްd1=<:%x3|%n }@#u]ޯ 7+{.Ymh_<*¾ ?\lf.%D6XXTG0ׯ?QrإY1GڕV²zW t#jq+JFOE<1+Gf2~~" >LTggOw%HbU|k8jhgTO7L栂U3Dm[{Zg "LI}Gʓh!Ai8&_P[찕j4u3z<AMѷqY~K+:8hg,hu YY> go6)gu9jڼ%Jell#1 zp`ESW(Rhػ':ijd<:>>:/d 32gp'PȒC%HU݌$@) ugd~x|bܕ{%QchY>IvYE.νk~0M6,p>'8| ؊( vU Ս٪yjtz ÐA1C2<'?ND/^_3DwtwxW2O(xR+.$}r -~K9w={U012B%O̝"x3t)6. )BE`H]'t? %y2U۷&ȭJjAɆս]WG1 .ѱ'?Pnэq-Г,=?m]Y*xwYmG&W`hR?[\|ΘiOT)'JW@Ȋ0 j-!/WA(w S VNQ@>xmScssTeGGɸzuQEԁ%<۝MޑZĀMߋ^% Vz9]H>41jHXh{[mQ쌹3aJd\l\~7?ov%;ftd rJ(&@HcZn>W-c A(E]W(㺸duBM5E5sJ߻PnZД-~:o@^`7;?zCpHK*dt:C-O ȝ1 jt zI֕S{xq IOyd:<4CjTTh0\ NDb̷nKIBm|82&u%q ٵ)SLJ_VP|P99P%ylwr>\S97⦭6(vt(C3u{&5=f©4۳lg8ETee!bH #r9 jn.c]F7L&bS^[_o#`v/D6 ᅡ"OfRQsbi3ʅ+IR.aOԼ(3=JP=6# T1+w"bg~;: 2\fSa\}KLܐ>X~N3܉]s Ӵ{{=H?ZgAΝpҏʅ2,JjI&i0E(R"xwmhx ,RGJlWP/C>D|cOב(De0C %O3p7i5gMC`+b]<1[YR(FZ330VLNwXGf1]AOa/[e6 *. CTw5CfprO#T6ej0D?&vb™=i -*cA<[cf^@6P<\̠~w; cs~&+-b&JzȎDfEhļ,+MX-# _#Q wkeKr:cIV;ඡt8_x#yt}#B'Iu R;;~#Q4yTMHib~H¥1}!0B>vi|Q?7g5JHԗLK)ĹI|śX;]kƪh(mrGSF9;a:&BLXkdIF5B9@R<+ŤC}h%;ӃPUH"oH*u-53CqB 3ka`{'Q-fWY&"˕`6NcI*uڄ0dz`#Scwik XI|'3b Q]_Mܮ]VD`K%[6pz@ZӺqfIu&ZD#\LBdh2 L>[xXJvtL`rN%yZTokR7jA@6l;kY*?i[*o]WY$ x%OMϰ^C=Us m޾z!oAZEFHL+]CvWWI0b EDJ7KTILrLY s ^%)cJFuA鍗DtiЫN}-q_ls #?W) a 8n+{}#&(Y%#~APP 6) ^@Veg:lɍ7+界63RD8 [ 7+OEeD{ {k!J!SV,i@RǙһ"JcU ^_FXFe$_^foX5j2r3(nqR;m aQ\ ,Fb꼦#q}Kr lLFfq0,Vߟ]Ɍ. {p` +v¢ SEjI5p~˙a0h?l5 df(K"OH"GCAKOgϜPlbF6G-$IG9=X`״K_-!gvV6^~G^h·alpY3bu4GS.2Ł<b&2ڤ!vM~Mp^ز_]1uh_ aN1Q uVt5l gE i.)M9IZUT֭*;XeyML{28t h;~鼣E6{pLKD &WQ&f rKRc"ٷ]oK s#TpU[IasOB,ͻRg=dp s7J.fϚX{\ήv"KgΩRh\LHz"OV,,θwp #ub!S޸?:#EI=gCl7j:pjM 9I%k<73 =U<&O(!-UC|Z4캴!ϡ D8y6 M4osL"OL)K9y`maI&Dl1-UfJZ& Q~bme$BXTF!MlM%6hؑ٨u@MT)nx)>}?I"oe,K&ɲ.nߠZUdؤ.x*ʫ<zxAxL?arEEoҪa}(;ύ6xc®[*K{@j[@KiGOgD9Lv"x MU(Hp].#$5vʉݝڋ0 "EY|#?ŴK&}MƇ^l{"AU(є %LA;TBE{heҨKa+(֘bUMxp( czyP82͖"q#~38 t(3 AwzOgH-~8Epȏ3X+ʣ|)e&1h=U{(ֈ{&=BǕ{xgfZ~ϳQ¬8wyIt4)=aApHuw:>>{T,=5yCЖ2ֽpy(: ldZ˃,߸*2n-bs@x r'%! wFH4WM_FYԅhmVfׇ<9F.ڴB@I6M<' T)Q& i;prl.虖[@dr! ~tKsCݱ=C"m'(R.1iă@H?X.[y DM>,:dSB$KrKvZ"dɆ«3HPsCNAGcF.Vfv7C%xgE92ѣb^ C}-9'ԬR/F5ʥf_r,Ǖo ('U헬44`Kjď̇nDXy8ޅ:J\΃ihX4#g;+e jCK6Xz-/s7"+D|0ZٶQxDQ)lŹlK*Fn6L4׀]=B9z 0JY%tq rȦ,ɣĠFBr?DxNqOR{zN$TKx.Q G`bK~ATv4tIh 0%6HEy|iUQZ 2cY?5HZi6ę}b5i~ g30> @oEhg$ʹpcdfУZiZ)M[)@@rC{XLj>M۷eok'MUeNǶR,m66}nf1,&!ݝD_&gPuktHOW.gmMP*zu,_Ü:t iBWbk;&q8>c a 7mk2ǚJzIiGB1.t̐[`G"TQߐK{ۋ3l 8 >zxއ?ӻfW 킰]8}DYV W3yv_(,!)+atak[;Cy_ m 9Xft]cdRpVORH+Pr%[:;ݥ*OR⒲%̼c$k,5.]4"Su=A#%6OkKXAZVgS8*@3HV&}B"}F得p!R=i]OQ(=^p''< 4\dί+D;3x7,O7"˃:n(.f9mwT _C5ߘ5p0GlXԲjai.x ]\$qΤy2W\O$ڄVk1F5WVk9W GCLÿٱӝ'>}jxTUYTu͋=CQDB6,$ 4ꔺ`b2^MY;,r%=|(++[h?HUnyЎN  m>?U];Sr'y3Ho^6c0l|g/~zMꘪ.R_@ьfS$I LAB *675P*=tΰ ;%a?3`ˋ}iyW9rXz~M#6{*wys%Ը?٩OXdwUf8$2f:YR]\O~ e-\%R7- X(o:Mٌnqم|Qq/MoB>q1kk3HĔrʑ28a]'VE+J:?cbHlFG ?IjB@7LP|LlNi<"Mo5\&}YNNNI#HXp-Vm]n?kř[HO+_ ȨdQ ~Uߟb-#SC晏RÀ*])@U(Slk%6M[Y||2'z"]Ge$L>չ4n [$<96q8\LJ7R%t=7 >ϝz$^ tsdmGoˬ@b;ѱٹb,r9k: ^5y;R*:1Aձ#`^c*=%]IcRҼԾj"?۫LUm36]J'\hW,iɗpOH~':gy@rcaKlÔu7I޺ˢ[;P~E䟠DEɞ:}(g}uk f¡>&i\Z.gP2}e-JkW,^ޱ*:N=ӫ$N=NN _Jǭ!!(+H05̬0*f:W0ϺdǗF 3-G9y7neY.MGn3P)׮晖Wj`0G#-37~\XC  ʝ Aҍiڬ3Vx4MH0kd7R=~E8,8_M871 Mu>ЯR-xbMmк0=ȺsB nw_CjSɂ8~]Ll}gR:-?7&Ђ蜼Mx7g:!#SqǓiYx=o wU:Λ@]2g; g}3Ց^52[!b7EX2 A0dGBc?҆R(hk06zMj! {Zw1PIQ*}s5y ^~+s&B]eqŸ2+>2à>NT{5]z,QGB e)]Y@T!s]uc ~=ewY}ـ 4MP-LkJ}%>V wƊ2HLdcԿA{m1عվ_Rtj]ȅ ONϭ"skXLqNK7d6fwbq JNqf֔AxMs  :7T`fQڞ|9jE =P2;T,6SDlCov)$:\{H|n􈙍k4zW/q7Y@!v_??pwk^jб"oJ}E} &V Sx"&A[gmN2Ѓ <=2\CbU8$~KpvzϵuT?|(}O;Q.uOGit$K>ukkN*0c,\z'PY5{b4r'p=SsN#yZrF A;;6K`%>Y'u+kkI  ql (rBnT!e)Lx6^v4i@F!0dGw8OzBr@ZsA:8Viwp,?$*@qhq10U?hDB>"z}a[9L*C0f>݄jrCM~jXֺD>_LTaWlO(%*CavAy%X ZCuVXP.LU%PUD=U&Ts޵p`ݚg_1luæU^h)[8&- \F@ɧB*њgz2WÆI6,<'ȫ ZET wvi_bȰtﱺߣ *]gPO $hGM)(kA:FEhe > )>}~ Tsx-0{Iv/6r ӽ|FmZXvQJ 6+C=5# MlO~Qek j5?c6~Qiq ?f8(4;2İ~4W1a! y&0sjRt@ T?BGep>]sԪγ34=Z$iI @Ol5ccHo0*Ϗ<)k>DG$Kz6Um }4K~xVYSgn :zGpcnIŏ V 拓Pjm2&jwX/bN\6KuW"YI'?ֆAt  J$ɚ.ކ$d1{V 7HLxt٪0BM,=l?toYPÆf\O;xv-5o[v pȭK;Tf0.Y6߾UfXQ(iIe+ t=W{YEVcTW!{h bE ,@JmJK ENf^O}b#e#t4"D^3C:2-1($/Q_n^ٞPrXʓW"mzqߧg#ބ7Z5"Y7g'ȺEg&lxTĬ5{$rY9o5{oXB}''hH&KhXXאkt;F=d"f00 М: LG.ɒY/eMQ~9oX3[ӍC[8cTup ᭠jA{Izh p^EE[e<iJu[@aD (<8q8T&+C;Qhp85;¤ =;VOg[;g`F~'J X& J'*'zWG:ܕZC^TI/<1; Cf}p^C(Nm7ToѲXgvz-b eZ@3I} eP18\6$g8-E-ss?A_~f\X5ctRC>&C.Z:9 -[VSƘŎ;HڦWJWjB1`";I gNsA竧AjN= G/?@+5UYS Wi`,.~l  vQѰ"E .˥2}~ #Z ]D g@-qt]yBSF'gXzmx$0r,Ô7+AZrJTZ]J3ԇ;qmPdLR |1>2`}Tu"WT-nc+B| HAdWCޔϛU[9[;(GRCeQH{|a{d &N]=(`QG ݧӏU8,DexfӦ9d*_h)%ɗ˙,_jB1͂%*g_'g+6E='m\(fI A4*9Ү=#j}wLÑmdac"!]M N:ksU2 ‹_#/3#3uXg$0^.Al"d5m>P~u#.TOȜSG9B tڱNXM~yk+9+8MpW"ǀVR@GF籜hu0Ckq3)Fm|ZϚgE?Ɇ-kxɣ~SH@%1I^% \f- Bt78kk1$u v;#l9G$^+܋i#RiG_O%Zcd/j<=!Pޒ_0Y@(o\ׯI,<`H <_yxJ8ˉN|>OeG\p/n)>p;1ù*o'}_Z7h" Ȓ=Eџ(׿6 ֮GalR3Nq^K$beepU4;aI xi) FR@ f=BsHw:|4ʁ)QM XFJH7e!d3_NE q1%9q^c/fD I|$bpsBgҽpTv0xs1d\)o=>b/׷y@gEHq(t"ڠ%y0q.V 8=5-hi ߞѼf%K2-D<$~ vɉc9{  ]!ߴg7dyv~s9~ F:j#R$Q7ȋJ͵}iT"""hhbycWʷ-M֡G9oh Zx rNYR L&XȧT1x;'Ö)z#\KˣY/D- H*o݇S4xےt ֟ꛀz= bp1jZ`Y6ˆ = W;Т^ j~ho_8Gګypͭ&<23+iTh=*PNN:~}9R t9|u ~Β)2d М^oz5%Cc f}l8ۗQ!I#"%wwh&][4n|m\n$4{ѣGPeVoJkq~ڮ?ɃrggzxӸf[H+əx=7ȭf:Y&l)}% ɦC:^'e#0YKC^{8?v<Hy wn=ܗGv@ll8:_'s}AIQwhțg#,pw@0BDڼfg쭺351sN'M܆/ l:+¢Tۖ%`пsBE9Ԏoi:W I.6@8\6d6 tx^T sP?b_mAxf2gl?I$GdGS{Xޘ5L%9fբ-OemW#a\?*.VYtYՒD\[}:)C3hLvk'B'v|xA:r+ҕLܫIw[q3W\a{TI6%- 쎛`XHʓ޶`VX˹[ݯSϞ RĄ%^tK772Jo6ʱ /ﯾͦ]!P|F!O (-+Ӎ*@b7NBEӐ<@ڝ隝Jv`ji [1kc"_:x)Ֆn]mfi]P5e_onO|;zY4i5qc'wWx .N \x$b?1KEi#&$ЈLr؉&(Rr 3Hd$ׂ@0 u  K HƐjA///)cJyX6G̫vٗM̠ikSwiˆ\d"]3.c$)f*z{л΅cueGBSU#sC)P$ e²01I6̽)/.[ G' M2i'R -w PjoMLZ1Eyj']~4@NEɌeGqNtsl/a̝w3لZ8SwO1ʺ`8yƱPJ:bMdgT@#[ 0l8[Јjt]Pdq?AYlE4]uou仍84/yJ2lvEĶCe706KŒ@z+~}{PsפUQ7`EFT@ "_2fK`IH3Jy%'txw( Dm 9qeLJAŨ]ٴ4O{y跹,D/;6{LI_yV,'`u!$FVB9 rE]Ja9;.jPK5ٮn_¹ Q,-Mk)t厴{km&{"j;Zc6#xר/IK=[.^4Q _E?g?UKLd韅[ J\R>E^P@QEsnxA7s]:yjJ-ykç-M3-z(ؤkc?wzTkQo8iۼg}J.y:~5nj^Wwz:u㝍i> 6XS$JO@$_r=YO4?cX%-E\^5Kͮ ԠȒda31%r-iB32KZuLU`PF'Tۘ͋GhX%Cbj˴ ,vPW.gVVSØ+^$<4opo GZi?fhɉ0/E>a1xHRKE8fD MTB84fYH;vb̸q BZN|S<0 I:]h=}(0?moia\ohH1naA8`g3f=Nh-4t}ݾSQ½(E {5e*U]o5CW -4xgDfp 3Lm"CM/&a7>C,FOkqF/ =x5p ߥֈ-yFxB͂&4Ky uxO8ľqPP cλ-U3y0l>9 g Qo0qU3US OZ}fY  w/P(DȆZ4Oxø6. 50/`*Z.V1A.Z{geO%Wҽ¡r2TĹI&`Ñ!dWvpρR=~ff$DNϦ|% ^σǠ {qB~s"/9`K:EGAw iM(s"Ļ邈}utV䡯Gt}U1{׭x<\lxh*vBh7r,$QcI8K144eP*<|Lۿ&ؠ\&$ 65qc2\wfsΘ7`t> k;.Nysdc Q5@I9]n(EѨj^#;QNodX9. F088לōscq_9v)+]@-MTX(Ւͼ2<\"Af_3N?i,f݈GZN*1+UOF:M2EM,,Gvr<}04@L4s|5J~:dU1Sʟ"y \uDJ/ѹ 8m=̛Zflߔg29"?1:?Z{.?/[\eWz< &qFC$6xf;fZo.AUђ!-~}ZK ڙpv Գ^dhw_Rښ_e'D?+`9m/MTD E/ > 6Hˉ˵w im7WB~E.\ᒱٌODoUDrIilGT.@ ۠dD9@=f${EQǚ۽j{q%-|EH6@шfˁNZmHO"z`֨XH:¾⥺+&B\%sen~@a" R`JǓz]eG4̨:}uX$%hk>U e﹍ /!2  YR؆z)eFfuO`˶%h2׍9d~-%+d9]a4Jw(6Nӭz{F#Me(fmhͮǪZƼRAOWWM}|="06d9>'g1 `;3y 6h׳J36gK\o77R8)ig6vcu+zdA  yLMǍ|}k|l迗i4xL#BG@ŕfkNy2:4z:Ut:0of-u>뿠nl(-QemfZ0Y\z 2[lو *QY/ҁ&1/,ߛyx꼞C hhmt2/j%QK=]L/ŘEnᶵèni_(w4`rO)}=:Vjӊ$iF91RYa@{Zizqb(*2w|H`qfFp|~KL`Ŀp&pZ iW[9cлrG T?I.$ wڻ XC١+z*u-a!0iF;8,(H<[έaǼC(a`!x B|au*ٌtaGg6S}T|nbOgm(1vpnarbaHY Hf$!3ӡ|%kd_`{Ti#D)t71ѭ^3x2H=Zt[ZJgN?joҰehEeV ,?EZi!iY 0*_O聞M D B rq Gx4χIoQr@?o]>Ru#^ Uƪh#ζI2/of oOuK]lq.6U$!i.c!Ĝ:eY(v[ɧ3'ɽM=Haj3_ܕU }]PWt6j=87!Ro)n³'Exc-jhpc EtimUR Up6qjR?."DeI0uWceT3$ ?B'`ji厭7iCsȸEPm]&O~ &ڮjO^lMzLR c V %%)]AOuԭbU:L$MV6 7HyvI_iڞa@K:u%ǽ 6DdbIpǜEP͠XW&!V^A r>/hHlTA4ivxh#,.֧K3Xand[*>7Eql$gRGCp$8`w/@^vHR8?"]-~=G]awtz@R >U)t[uMd\;Tvv,Jᷱ}npYk->Ol&;k lMFKv$(5]o{ZW )MDCԭ~Bx\kRuǥ;=яt>'2TT(Ko7ϊ*&\)Ux\'h\["d>Z{r/#'%}ϡrZ;jkk2̐ mǩ \zٴFڥ`i%*iX0 <"oRzbrP;+.3X܋sK\Ş)]|" GB!C,,rؒdbǂ@l1Oxq;tшXB> <͐Z'ąԽ_JM?(gŧ?C Ž?shv=L^" Ao!E3 E۲D;ǃg 3^9qܷ8 Ʈ{t&zT+ԕw? N>?aHp.ctaMwQ3iuQ@V (5&Eq<X~|_ \g3t:EW}G[S Dy`#ϔqEMrr=Nՙc gmCR%_̕Y. y0ߌ+r')+,r#9EZ'JEv0i&00$}= ѺDB\u!,ˇ(]^B0"n]LQvAŊeV[Ig: FHB Z}㐼{"Wk,1rw'P !eHtM DVv5.ͫ:K.Q [:7P-vH: 0Zڰ 4ER}>/ 3Soe^2Р@` ix _P`-dv $7?Ce}!/`r8Dg1G3hk*lQSY$ib.Ƅ~25m%Fo֮Z:n4:/M-:#Gr$%fHaM:/DmN{\C8 ^axgXAC18| }Նw9R oX)*n8M`J x?Ӱô_q%=-PI[l˥.bf_6U ]чe9hG(r0>AOlK|=Z7 UT `@R.iHrMJai]{dyTEx*9WGNaW k0tH֦[m; N3ҳvW⺦܀C&tcDg1[VbP5f? 걢tyywd?ԃ$#};[HŽgy3mM:QTFnkpɕ^i2B'1y%2)bz1@悇gz p=ZQi_$}NJ`9)1mV6\qp2^g&wwvX2,AP"~ϗtHS±LLrTaY4Ę^<-0qq#2FFe"nM~zU 01 ӧnmv 8*Yp' ߘ ݾm$ ̖uCYy_aC->!iBۙuY5-N:4aΎeժ:D֠dYXLD2Ƹܻo"?^ߗ pRڻUsa Sk:?** 8A$!C Fݦws/qh A_b!y.i[P/0r[Vnb%32YK&d4/Ra!,?KjxΝbkej[vDa;9Iwk%)\ Ev`C?$փ V1^w=#xq`g@LϠ3Q+ xGNn+<;,rT{CӝRIv*p(dC}A k*R4. H!NeP u[b/=olݼ|jvԦr03g˕= + =8ިN1].bb6-,D5Ѧؠ,pl?lf5v,Z"m^6˖v9"';ßտg*,vS8SO ʟ!C.Dj;;@O)^9 `x$wE/iL 6,8EoGVb/2NMwg,S9I&= A卣:wCNAu$Fc؁KwxXIJ_3I ^Zc;Wu& I }Һ1 (4"$ /!nmd٤2Iټ}ˬ(?Zo.~ K:2 y*q};@-cɽ$:r-]Q/_W9=3QJ̮cIfj④ u9;5txljA)z%FYM}r0۱c420:ycdZq?@ k,Cߩܺ.vtX"gh[|םvC,T+uR;x-] C"* MŠx͠u ;Z~R[c޴=ٺ%pʜSԲ=;׺yFGg}R?A?3GH1)f07L0gᜓ_xOJɰĶe+޽s dDE e܍ׯ/87uĪ9Tca*Z?Q8.-?S>"=\_Ym ֬sn_/h J]$wN ge䁈L>:֟vͮ /Z\Pm hVa,}kR-Jg02]5Vɩ"q0QDN w5R]3\|@Nm魶5?4_ta4 LٌNp#IK%K6<mazAW\b3Kk[&Ínܿx׀e%oX:#xRzIʄJXRb%G ] C, HAgI=BX阝N0n:`ϕƮ B8oqsaN[*ҢMTf_~@LR\bopH3\-EUwAٯWUh@o: U:@٥a #% 0T^dYG81Z>mz8މ[[NF˪cn x8}.bgk %9S+1P$v? %H_|25Q]8ajHc|BL)~{PREdQhG{=Ǽ.!#K9 %@dH䵁.; 5ezm{-܃WJ%}w6a`"NQ U|O[ń>ej;7k$׾p:iZW64gf'#}7:a$Vas,7~{)=3.$*/9E-䧽MSΫ洬p nb7)gQmƒƯFmhؾ0N9 A~ୀcU߅AT%ݨ]l&B>OEn );q(oFP $zML:a4l"X,?y Z|:ުk3@Dj%A4ͦ$,/ NT{>slL!ȎZvWK>"%cZ[I FD;9.լxes=Ť4,+08!;[yG*?.=&%YAJ]~*cF݅$nHAMQjQ ~+DRl> / C $\$$XܴPO g/Ţ`G^ Un0_GR/D~ $fiLcbJF!cbsCv9 4t{ZY9}O/MDgb(B\jW>,kƣcn=3PK}'I>zCȚ*[5)7ڳ; z DA4c 1{iSoZ8:ej:.t',WVy98+:}35VS{rY(86Y-Wx ҦZ\"'Q)(+;H hwP1"? Hq\W6笷h囝b50a؃ᦂڅX";N;EN^f@EoF%JuNbA mQ GҀSLe>)GtmY"/X:g&R96avN>VBHβbzP=SY0Na`PĜ-^IJ}#q9NLMǶ1cJ>U0&B1 {NRFC61z$ sikpYȋsFxpxlaoP<+sk$hcEaW,N6D<7=>#CRp2x)*"+IƁ@D$<>iN bi: VYi7_ H+ M߲Tvw `=2H$K7jOm LqwBt>5t}Pi:XW[J"F*rQxP;-S^\yL+~\mw /(tVpoA B{Wâ&A7!!_kY! }׮xG1jZ~21Se\x$&UE+Ŀ1;G?; 6xzVF fhJ=C9}NƂ9ND0jgG+"s.D2r,\?zs]V6Z/^ޅ4lX%Bs9bƓѤc?dNg;ş@b2Հ!?h*߷ $/ի8dLdj(s4thlIq K6zdR ^# .o8o.QY(/?kqxoz9=ҌT ag{i#)7Bns_l[^ viay"E]%!E&3˪Kգ+Xp} K8*{k~SŇU:[shKR.1dh,Z:HforYYU6}yVX߀.Pn`Z'rωnzg>dȘn⦭qJ+Gr|E*NyٚJC[U`^nKT̊kӒiRns8NI%dCo8R-Tɪk!y9Φx}s?/BxMS.*5Pƀo&W"VM@ e3tcjƏ ls7Fu ש6= aGZ Bc[+1Zb.=9=ߛC[@Wvu^HG. $vsD|;0q gL ֬ 6X芞C\)f7rF|bFgSS7&fpnkO͑1EEy2ph. {QNv4 I T:ih%_Ӏ%,pd>p~̆=/Y [PgQ΁_GS\m\v&vӞ ZmJ7)h?\e),J,67^/OՏ#ֻeoKYirA#pZc'V 9t97i(Cw d tDϴT*0 曇f24Ȳt)n|d-'c?UZlQ73)z9'ĆfR)qvcvܢ 4pP9v#Tl=^"<ΦSAM+a";Ʌ5Ǵȴj: DχDT:/$wcs538ih =Pʟ濐Oo]aSQ I~Nව¦&uѹ}TyW NJb݃Ր d^t"]utڙM2uZel4Z[ Tz7pbn Ԟ]HUiǽVb4LQϖr~-o|wD6ychM̜n\ͥ~#vT_/*,]7k0@Z$;j= ބVs)#1uW'5b9amׯUny~ ,aZUSšEQV秾[86 ӅĺtV9kW| 3/_Hֹ VSW*03 QRQJ``ia{Nۦ 6ӦXyP OJF}D|`@]|c mjL{ά gbdd ~D]Mu0dn]O%}1K,4\tdsu '':UUrh)^BPm -W{48}5E\ ':9VN,Q}}Ȕ)=eFD !0y#\IY^K-0&Zc%IJG`*\m8{eXgՐP Bޤ'I|N+a7#RG_N"ep f}Rg5-s'#j[?JX<`"B*Au@=RڱُŽ=,%#zz{84 2CΟّ/ɈqxHg% {} C?Bg_ LF4Bg]D$)Koҥkm$Qֱ?44)%!@F2DΓ)%)gUoGbь l1LJU'p{eXmHGc3A3[MpB}*X.* jDwL+$dWCBƪEC?HڴsvrOkkC$Q:M(RheM=4St,2Qo4Yf|go{bO\ 2lC‡Y z{3R:Cۈ0}xl\ݏ(?\uky$FÂ㽖dKXnHuV5j3nzݮ6CJRw74y$Wu5ׁ7w'̗V+pЪ3ꝶp˪7_<uif;_?xnvNODo. Jl]0iȅ6 3$a3-ߌfp--T܌̆ &Vًˉ$ڕb8S-|IX*vlٔH5E[xCmYA^UM}RJh 3T'7֨"aTh NYJzdL_B]&'ڶq޳cgCO`|_A9ܞ D}5gLAAQX>GF|_zEӕ=$"&9"{Lu0 fI̾Cc(e$/.iiۇD|[ԫ w.P*^%m}eUΏ/e:\ZۜJ7"NtN)‡xWY )&ae\wKbۂm#9 ;G1`/?Ce;lSMCE,rŤ)9T]?1` Cpv yU |qbk21؁s /qzӗzšI'kqY 1GOY"p,i{ R|~iQgY6R^tax)?bK BVj%K SПӒQٝ= tb<{JnT8h)e ʸmψQܺQxiwKcpU]\:[̛ St-;ÔS0jyxuN䜀~zS+7v GJnKRB^~ӑdep{7pE=2ApߞXcsG z8,d?/ON5w‹v%_c VAKS鷥&EW$1{7w?4d+^x1à,J+)xR^ [Uօpm+.=FL"`Op-.R 62rƤ08+6ɿt0Ak8H7:8yghP'sJүa>NV݆Ӊ' -Vv J;IHXCm00 cWawIő]f7e*rc⤎*ȢvO[f A Z<צ5pdc(x߁Ax%RKIb[Y':6%Ce#?Ģld;aCFcA'53OmHx\a6*@9E}@' qXeSN4CAdZ|-JHe^8oCn3giOq.\S2BŞW*9\4nJ5yvI_x,f^3M6/BCJ]+/}Ps;  {h 3x/*I([/FP[z * :YZ<ZHͻ0H)ΗO-51{)s8DAK1H gk]'%҃km;/4eq7iGլS}*EQpst1&]Άoo'mz0>'բ-D{IU΢GC ~7-|MY'CaeQqqW4B5aD /w29r\'B0/-e$E`sQˠ;4 < v.4O}"l%d4|a`t I&"ţ(T[ !:v @-9_?K /Z^ƫ=Pb`5\6As{6m*y"`M,Q:׊! ݊e9vg*cQJnDZ #˳DBcq JP[X|݈RJ`L5:mq.1(*[b4;B؀i!Ϲ+ Ũ^u|t쫏#l?;sZTqD~3ky=9""T;4rf3t0$HaЭIY#Ncef g>52_._sdFqgRʼ~5׾2P}Ŕ/wQ"['2z5.oW:Mxi*Ak֝4&!&z4$|gl7FXG3>YxV"V;F8ft 'z6Izj[6ߕ{.}$rpm9q4}ǐܹMK$SFW`7DPv?ooJuh  >STQb>m70$FP{4OWspuE¯~C Xn18ЅO_i)ǁxgx#?rԲ?VKg.>&8<{bJ{ѝ:h aFBe^AN!,/(΋kꉒ>eP͉V5r;V)ˍO!sWX iV$8\Qd;bMmZWB?J;w{C`ET; ]ԀѢsHpHשnJ rڼ|҉qz{r 4,n[ܹpe!}y~ T8uu&VRM.ߵJY$8Z_!Uuzٞܫ4Em:;Շ4~S=S(_Qi<^qfNTnx/cӬђܭLc?4h (:6̰c&k)(1dZ& {[t5;/Drm68@@ȜRD+{kv΀j }EAнn* s1X|gq(P\&jpdasm D`~ #/:"w]9"5>L豻Y) kr´I’e,ULσЄ+yKE٪0VL9D42+Ay `\4`^ f)!!Y]ߢ.uް[o S dr2 ̢!mh"זUJ͟"kqUHf^Ig'l 8ъ]+ma$T*q@[vIn%d1 sw+%B*U;UcE{)XvM"iE!hyId>H$GruhAx?7TuOmzkH/7GSn:o: >Ʊ.XC+aߕUJX#t.`*A%UY$Ep2]nϭ,09K$Q";x¼niqqtFa $.| ZP]`򼷿K5tw2. NkKJ%[8XmA/)b,}Y82]CЉ9!]^MPOJgu<쁸û߅W%uU DkFGr0QDƏED.zxMe"jwR?h/ (:BHA8)jPcևswض&ڔשw<%| ұWÞטsθU7yo nf3\3xKd6d1aݜhx{G,3$2dc?ӿ2;o~o2q+UWc GBa?2}Se$:E=`BB:6 ΂#.]޷~|yygXB*xPGMp'[ 21ѻGP~34AX. hZR:soiNTՃ~_mXck|>E=̦Q[6Q e(CW0'\0LMT#~ "AojfP݂A.s[Xe {}}52[lBzlnZ$g?ƨde#L!KbB,I4j$&&}cTG[ aQ*A B ˚H!92ϕh2=\ e +Xvd1kѕ`m.-G;iOw.B눘!an0<|E=/AIhT*P"r9YE1;܆bca~0j[ ]8 BؚaY?K- 3^SFۍR& P& 9}xF6}Vy3CCFw=s|ڕ"+{o7;+iQL[gHfa)t?*AW/m"9"J)'WVpGo(8 oMr;A(w77qZ%Q'Q 1"7blDdc$t6GS( [S%C(t;xy8G#bf|, -:A:^eź*W{V+YЕ'q~w&ZP-}پEE?X9oB_//GN>(~r㫵EMD-5Ő7"kB[ӆQ]SB"L"nw *qlzT3,r-8g3_UnB7񸞆UkrX$6D,¡H;G[ ma=b+c?U7v foOZyP6<]q$UhJG9q(6 Z#'R.$w+&?<>O9N{%[:>tW)4bmׇ|!jW_`nR&oxHeԀˎlT+tnwM؉Lpe>tY'ksV4E,b2\!CiBJƋWU}&G9$Y }3OP$\tX]` )p,IQ~Cx MwwXG_xg3` 9G8QɿȪl辠ZQԴL%st\hnn վ,'^R#*%y ʔ Vg2; >n s^, T_}%1mv-x׵Vc<4Cyy,5<"}ƚ?D ʷX߃Cc)s貥Mzy~{H0FE%=x!ocLS FCoep&c|bQ&!6:72zO>Ԧ??H3~ xy FJɆ0IР4 n6:T=B` o k Pu6T}#l_(Uv!{,&Vy_*&_ [ O:A77TFrcJAj<>%t ]A-0؉~Z)zwHMSsDT̖lCfuRP_ٕO #Iq ϕB72X}g L `32'O# kw꾒&RC>YywKcհEn3J93]s.֦"7pLu?:KEёP5wlJ{rOGո*ְz$UN7+[~֭NI S;|̧ьfv*Fe u/ /xO2NAoӦ.HH`Pǩ)ĒyᏢ-?WS(gVYUPj=)*K!g. @#hU`3zWfVGY T>aЇU<G!lnx^C=miɚ6mv vn1 >g0|$/7N%X(lZj? 20A]#;@((x"\#1? Sɬ`r\*bZPw@yɜL}/C˽%r<0<NTL 鯝W#дzvy!w <8duU,Vp "R7R!]Hӓ19rYJO 1֣՛6#{CF)_J"Knb='j# yq?I?oq>f%@#I!oE%v;M*b!Qm2o/z6V1-C^PR%*0nja1 |ku<6r[HMk5i v'TSǑ+ h2'$=p\DDO$Ij60[EJj, [B zeU 18xk.С#8>諏q !2# Xu (Q+L܁ t#GGC鑶y 0 ɿ~ SX^~|t>UB-;@{8cx9|2&;'u>_pe|H_ s# t# ^~Mnnb*wm:{nyI!|Oz#z!AnRfi 1"9)āXyc܈p%Bqᰎfg?e0iXӰ :6(H!>WKyH0g݋:G=fǯ)iT'6 4SpQ6aDq!Cd _+yҕUINn7FJf:rȊ24:e,H~,L3(K8a.}C YFnsrzЎa\zķg qI`#qlq~A{?@]Ih@ 'r.>oHViAkmߣM6pG^)שjS|:DY =岘;`{L㦗 :̟vr̰S $5m CϲF6޺te\dW}wsMC`=ɜW!&oJ?$$ f̝1?oOO"(5LtcABN{ yu<1y A88 QJ~̓!.c~Bި"Yox\,Uhu>F,Ꭷ.:zqX #=Bu{O6|Bz.4hϫd-b2onm}#Qq :fc S/뻺ƃ:<Pʓy1dd'Jar4*)h9|4WJw%\v 4kyDwl_G7ya Xg}6ҏ&w!aJ<9j W {n $PAOf3 V?N.-??\9~RѪ%_퍄4pʵ*FK|b^ijgm 4,La ]Z*8ʬD!S.C`0p)YVqbF "+3{I?uPw`t\1K~ӁO;-:m&[ t/}<$migXd` ~?8 lrS2zJ6~drdڽKqXqmYL YrąQ@ DL݉4qxlcwYV>!NZ^ H~YDԅUP1\H)G$ "LQ'ENHn(jLI28z,s-N) )^K)t("y$sF d2Ϟ|*o/䌂{Vg89ѳ 3t#^ ,oGgHT&*W>?5 IF-KV"'W!}L X]֊n O.z7afy^JgZ>UGpInne|EnZ*6v@`qk@72}u[%-Gzհpѩi'%*b={O7tզz|Ո ha z{[駽DvJ{\*\sRɲiZVe^݂,. OrL< btY*|$Ҭ;^jjb*cT%0Vw>5Mc~r> r#!Tȡ& (2ge"0(V5qj}SNCގ[<.>3;qyϷxg|%+k,|Vj"2"YndkW\?OtV(j0]N5||@sS}>a#:NP˩27Ff&Or;”ʉ~}9; 1W )z$n;K*jUN]޿',\ɩ(1. SԦ B&7,#،bQ [spdyjɥp:Mt|Z;?25-`q7 nv ]j{vra?F{$4ؗdP`zgS3Jre4z儽 wj=t _rI8mY.MnXDv'v j ܮJ6qEřaޒ<012&fCiHFkEukIMWcGnKTm ;jPyh_P[1Ʊ#On9~ =KVL\Fy0}+umTƬO@&,9S7sۧ&0i1^"VcjLTX^ 1CSVsc\$4_UkCϳEO-8lolӦdbeOP:r~'PCGFNfA䶲WgAR9"kf~6͑%i2j5!лDN{p]-~̊"yس3ѣ#yo@0v11\eL>TEEBWrmNd> K릠&+ _U䞠Qgٮ3=f'9'v3[#7]0"TșueٹoD2j]<\WQ8ַ5 V pr\Z`Py#'%>rzB|R@|xXP]d#"n}Ke Rca;yк,a>Ňv>q@\o/A wnj.Jl$¹,PWh*1l3Zu!COsB@h9~zn;U9~ڦF X ~ۡoCjbDϻUi~"n-@O'2VВhPOۊvUD FC\*{D4?e^|t߱ z gxI!F=_ce˜j@u" b>Wa w|%}>mz|I|vK@udaToA꽓.re[*rKfE~לh݊_'e'Xl⻶1r+ 5͓9z-36QS^ H ,5djhU¬@ޟ#7ZᣘNym$孮s< A`n9j07`[̴$ lWFDAx7k#(OVTq*]: yB_RkkXon/UmX@53d7 |: >q0]t{zϩ!{N %Yzme @ $Y+5 ,L[?~r758*5Hc@G\v*4SKgS-h^"Rd9bTyҭ"bWr:.h7$A E=|cHPF|/ppia>M|˨m &xTuPucD֫?tRGdQŹZ{0/NT3%1 #rtQ!2|FKzB9ZSިZrC8*|\#}J%\.n<]}o'@ Q6`k{pC:%w2(3|O .^2K=.t;rO&i_ʚ3GVt }l xM<4ц {NQ„ᥒhf]\l^9|C7$ iD|*(nE3T„H*u%fi 0*#'De[T=`q_%) W2I`.EW_zo G#{g ̷˚1ZdLAಘ}&TdD-ΠA(]"&oٷ$=,i~\KJx~/ 1펬gJbR{L64}뗺?X`>M_tSޕ甝C Q'a;t:16b 77n8x L &$1`iL[]yr] X ֑X }DݽEEPD7F87&Pǝ'u,h?6s`e9'H?P[,3;0{-{cp-jT%+ Il2sI y[gfw_|#{Jj -Dt LazY)dT<{983nligS= P}m*Xfn\FB5죮㒘<^#qA^'q3YH׫K i)U¶ X31z{؀k_>]TEclwVnG ?3uH³SJ&fe d r Yjʆ^Aj|Ct!Fh4n:e*jI!"fL %-\|'2LSġꀅOzʟ/o&H(I^BK:E*4˰IaOD9c$$_nل ă2*> -#U kn\ByP+ܐZT@. eE͟= Emg``aYجC_)e5ltk ;IEST5:v+OAR_΋9cʥ$3xCj|H50G nZÞ`7E]Isz(*LIE^QD[;R#w/8ѕc2}9C ir3e} r2U9p\ >59:$0Uj+R9[sOPeV#8npoЏħlO&Ȫ6p]0 7o Ўډ}m4% 0Q$ {B%ypN#Wx>ʟS1O7l-Å'("ltM Q@ s;m)`ϵら H+GĂKd8[;"PN`yL*m;}!;qS d=&w6M{-aꐞߚϨ̔N KlhiLN꽋B5q^o]ѴA!&5}&J'vjb, 'ysaz qc%M{*,lVqF%^,H8|[ܽk/Б9OZ>%KX.#3_b& ($`U4L7#>HnZQ/aA,(oܜ6Ю5eڮr*\j{{N.B ]dʧZR͜a ۴$_ET @d?ZE>l$ ^hOtM` <4?՞(Z7 8Tg D +M$,M9(^u.>HCY[,`hBnn컘0`}NyJ-A@oO`OR(b1lh^hdFbGT󱌼 \Bl{5 /wmMw ۩ߘo ;ShaCc%;+F Ϗ6b,\<;OAԾ@:]6/0#h>Ӈ,OÒk0ẁ?#9g$$r*q/yxw%@I(fE_&LY^ G}phQđ0Ns(mD̺pΣE:'?_^:*з@S+U˨r{(oz"0ޝ ^oVvkG }%NuJ]Đ /5Ԍmљ CY ӺǶ7* n3dEjZfLXf;u aJu ՖuzO@+#\+ѭRͪ&X7$ qBDS1\+1Opp=.̹\RϠ#&͝,e;%QmR G]A@dj&d[xaF60-Gq5h) 9VѱajT=E Zޛ 򣕳+xKU,)o)ƘTӱH K 8j0 FTA-Yjn$e@|% H^sV7),:v$N{bBh!m!v:BjY i\wns|jʢ.%Pb)sVtGjJ*hPilPfwP}w_yԗA.*SEQ )=LGa]S/A>"]kg]r)s)K Ɏ+~pVVkeR&Yq0X k5%]LpϘlAVdP6%mg qB %yfvAkQ9%o"K_pBM^۵G/ʐpEQRN?jn(32]vlGiAЍ3^, -xA)@-<%sq>Ft-z$c! MG8HHK!Wr\-A;'{m~/JO8iȨ#/C\ 3[ ߯їXܚm'=M,.{w2 eis=p[c1 lIc"KXdcB&X(/&XpӠG>}$ծ`,ɛ|vC6=Έ b`(  M%;c-ւs>#V