Skip to content

Latest commit

 

History

History
1560 lines (1338 loc) · 198 KB

CHANGELOG-1.14.md

File metadata and controls

1560 lines (1338 loc) · 198 KB

v1.14.1

Documentation

Downloads for v1.14.1

filename sha512 hash
kubernetes.tar.gz 87f4cb0c6c137cbd07cc0e0b7049722ed2a3f21866ac02aecf1a0e03d350d1e90d7907487bac5ef224da75d05056abfa39f5b1b3741987dde1c5165f379cb253
kubernetes-src.tar.gz ef1b4ed78ed92124fbec2b2bf54ba3b293038f9c8f1df3883071ae9430450cab9c02c6111cf171ad8d61a0aef6d236fbb9f0f1526e6c00f0452323e8c7c16305

Client Binaries

filename sha512 hash
kubernetes-client-darwin-386.tar.gz f9f14293ab8c6e6d49c5be16c2bcfad640a258d3e1ce600d6b327a4ba84c369f679b8ed65f7f96c23b7277c6cbf4fa54cc75dd8d75e4c8a3b756dc02f7e99071
kubernetes-client-darwin-amd64.tar.gz 761c6334ff33e0487feb15f9c335e0e44f36fbb1b5b95ddb8aad0383a085ce5c31f762d60b2fc4365052221b7594b5e9c046c25c9806ca93e7af9183e4474cb2
kubernetes-client-linux-386.tar.gz 58c4db0219debd85ded6dd0eac2ceac3432772805826b362d915571aec0b3f93e60eaee7181bbf28bf7fb7d93011b9849fa486f7a05b53f4ac922845f2a5deeb
kubernetes-client-linux-amd64.tar.gz 66c3a470caecfb35ce9a995a6298e689aed5fabefbdb8aca5086adff572266ae47b997eea03ff3ce0272fdb5be8e22aced3e3ae35906b5ac90cf928d7c0c974f
kubernetes-client-linux-arm.tar.gz 50f76e9cca5e056d9dabe7f27de7db72539cb33c3e24bb541e35cf32293b7614d4a22447ec6d9e6a604bfe97825f023e72934993bf144c7763f76896d57595f6
kubernetes-client-linux-arm64.tar.gz 088621d5cbb8587896b38e6e1faa84471490a1bd2371c082143aeebc0bac6498422c9175014cba22e5190dd761d4154bec91b1d1b93a09d1fae771d3bebf2227
kubernetes-client-linux-ppc64le.tar.gz d743819920dd3ac906a855af2c1a1327f991e4c295357c610b1fad5d5cd8abf5ac1296e3bf9a46fa3f8877a152e3f8fba3a5d27e51289926f7519215769c24c6
kubernetes-client-linux-s390x.tar.gz 71cdd44a0d5418500407e9eea6f7118b7384b8c9a4bafaefb78c107b23e0503393b5a831bbe8eaaab6a37b4b23be3e7c5f700b991bbb4e656a72c46198e40e35
kubernetes-client-windows-386.tar.gz ffeacdc7658da5564811047e66e2b8e370d037b8d983a2af0ceb9cf762f6214658f633fe626d6e97810f520c664c0ab8d346a8e2ce6be330787c099693d76c83
kubernetes-client-windows-amd64.tar.gz f49b8acef5c31b59dfff0d63b4e175f54f605dd5783bdd57e18cdea766719860490281d2cdf0a3ea1f77d2c3753b4ec735af06ccda7f5ca4fcab14cd26411ef2

Server Binaries

filename sha512 hash
kubernetes-server-linux-amd64.tar.gz 4775257f919bf013a92d6e3b66099d85e09be32b462189ede82874ea860ccacc3f42ff2d70e9e55b9f3b7886172bf85b26a87bc51e9d42435bfd608308b84ec6
kubernetes-server-linux-arm.tar.gz 2806099d6bdd32940a3221a076fff51eb9c79729e058a6b5ef283cfbbf71695d35b0d121643282a089d1ce3ca8599e3a6d095ad90be931bd893ac6ddae562664
kubernetes-server-linux-arm64.tar.gz 1aa3c44189f9be25181b69e6ef336955f92ceb0a35c07579e386e4c0584d4bbb5b6f7cb044ccb7002ea111f8286439f02679391f66799c78e8b9d6e100bee5e5
kubernetes-server-linux-ppc64le.tar.gz 6e91be7bf7b3cb1dc20a395cbf32c70ad867f1300d09800bb2c6854c93ff8d9cf8c180b703f3172564f0b10336ce04740015f67738fa5401992ad6e3505b1b69
kubernetes-server-linux-s390x.tar.gz ee915182eda63d67292464ed3a3e029c1749d016cd730e15dc8fd19fdcc8ee5ae7bc926681935b6e6837d555e2165c231d3554e9428ac09b051b31f5d22b07e1

Node Binaries

filename sha512 hash
kubernetes-node-linux-amd64.tar.gz df2b5010eea175fd45868c908cc769cc8fefb367899af448ef9c64db5e4a7b50db9bdba77b81943446d0abeb2d9d36d72a22a8d72042f88eecb9123c9b77c0b5
kubernetes-node-linux-arm.tar.gz 338ca18540c087c81b07bd341f390b78e446deb270d7e320ef108f9f293518c26580c17968c1a87fe7af2546ff56a9392009a354202dea1d2083b79652250da3
kubernetes-node-linux-arm64.tar.gz dd2544dd9543cb9a556def0116fdccb8b14c0e7ae07debbf10f3b4ac0669a1f38be28327114781157cc9ae06e96140b1a0650eeb707bd883ae3509e0ee873da7
kubernetes-node-linux-ppc64le.tar.gz 866fd6680f0d35c747c54369e6363b0241a233a505110a899236e1390ec7365b9ae7df4ddf7087514dc5102ce757a46b4fb218db0c081bb15c200ed526209a83
kubernetes-node-linux-s390x.tar.gz 87e7b803f1ae05bf4652fd0b2f636ce61bd1100e40ce7c5c2530407346260435a8f649a41bfbfa5cacb7a810d007ac19323056ef175f67aee469528b0a7d7e30
kubernetes-node-windows-amd64.tar.gz 0e46f70d7f1ec542b33119a9773a98dfb46d80f96a9f31f22ccc33c902e4bb102e2d2453a0fcebcfe319b331d1a78606269816f0f239b68902f7059240ca790e

Changelog since v1.14.0

Other notable changes

  • GCE/Windows: disable stackdriver logging agent to prevent node startup failures (#76099, @yujuhong)
  • Support vSphere SAML token auth when using Zones (#75515, @dougm)
  • Fix empty array expansion error in cluster/gce/util.sh (#76111, @kewu1992)
  • [stackdriver addon] Bump prometheus-to-sd to v0.5.0 to pick up security fixes. (#75362, @serathius)
    • [fluentd-gcp addon] Bump fluentd-gcp-scaler to v0.5.1 to pick up security fixes.
    • [fluentd-gcp addon] Bump event-exporter to v0.2.4 to pick up security fixes.
    • [fluentd-gcp addon] Bump prometheus-to-sd to v0.5.0 to pick up security fixes.
    • [metatada-proxy addon] Bump prometheus-to-sd v0.5.0 to pick up security fixes.
  • kube-proxy no longer automatically cleans up network rules created by running kube-proxy in other modes. If you are switching the mode that kube-proxy is in running in (EG: iptables to IPVS), you will need to run kube-proxy --cleanup, or restart the worker node (recommended) before restarting kube-proxy. (#76109, @vllry)
    • If you are not switching kube-proxy between different modes, this change should not require any action.
    • This fixes a bug where restarting the iptables proxier can cause connections to fail (kubernetes#75360).
  • kubeadm: fixes error when upgrading from v1.13 to v1.14 clusters created with kubeadm v1.12. Please note that it is required to upgrade etcd during the final v1.13 to v1.14 upgrade. (#75956, @fabriziopandini)
  • Fixes a regression proxying responses from aggregated API servers which could cause watch requests to hang until the first event was received (#75887, @liggitt)
  • Increased verbose level for local openapi aggregation logs to avoid flooding the log during normal operation (#75781, @roycaihw)
  • Update Cluster Autoscaler to 1.14.0; changelog: https://github.com/kubernetes/autoscaler/releases/tag/cluster-autoscaler-1.14.0 (#75480, @losipiuk)
  • Ensures the conformance test image saves results before exiting when ginkgo returns non-zero value. (#76039, @johnSchnake)
  • GCE Windows nodes will rely solely on kubernetes and kube-proxy (and not the GCE agent) for network address management. (#75855, @pjh)
  • kubeadm: fix "upgrade plan" not defaulting to a "stable" version if no version argument is passed (#75900, @neolit123)
  • kubeadm: preflight checks on external etcd certificates are now skipped when joining a control-plane node with automatic copy of cluster certificates (--certificate-key) (#75847, @fabriziopandini)
  • [IPVS] Allow for transparent kube-proxy restarts (#75283, @lbernail)

v1.14.0

Documentation

Downloads for v1.14.0

filename sha512 hash
kubernetes.tar.gz 0ad264a46f185a9ff4db0393508a9598dab146f438b2cfdc7527592eb422870b8f26ade7ed089359c06741d998fcd730f897eae261f922c1a26d9fdc034d270d
kubernetes-src.tar.gz c5175439decc1c5f54254572bfec3c9f61f39d6bd1cbc28d1f771f8f931b98f0c305f1871618ce7e9de9cf3bf8227e19dcf985a7e017c74d0d7ab4005b3dbd59

Client Binaries

filename sha512 hash
kubernetes-client-darwin-386.tar.gz 68bdba50a2b0be755e73e34ffc758fd419940adace096b1ddebd44a0eae2c7cdaed984965ea8f2145c1cab0be47bd6c72c2aeb73e51d449bfeb9ce1854b6c562
kubernetes-client-darwin-amd64.tar.gz 255bd93082b3ac5d69bd4e45c75c9f19efee50ad6add50837ff2987ce16cbcc485fad334c980b17f69e5a344ee50548e206f747441ad4a045aa65746c79d10ca
kubernetes-client-linux-386.tar.gz 2bd115ad2503fdfe5482e4592fcc0c8a2aee36be5205220a13c8050cd1e55dd3c08377425dbe5a03e4ffd21cf603c739ec4eaf3e5b2514a725d095df46f25d98
kubernetes-client-linux-amd64.tar.gz a551adf8019b17fce5aff2b379fab3627588978a2d628b64ba1af6f3be1b435322368b00dd04fa739d01c341420016b93239cc0d4601cee86706d81d78cb4d7f
kubernetes-client-linux-arm.tar.gz 24e771cd4074786330e07f5537259a28d0932102639326230d9161f12a8dc545638a55bc252771eb4e21e95e2c7f0918dc1238ac4dc70d3b8b33f093da7123ab
kubernetes-client-linux-arm64.tar.gz 16204f2345ab3523bbe3c868f04806a97c111d940b2594aaff67cf73b4259040c7770d5b0e7bdb7ffd7389f87e5f090ae875bd0f192b07582f59a01a1df32f5b
kubernetes-client-linux-ppc64le.tar.gz cd9ce829d585dd3331c53d35015d4017026d5efd24b9bc2f342995245628598c98bd8b1f1d706b196a7b3046a44049d4aba6efb4b1000722bfd055bd8a662f1f
kubernetes-client-linux-s390x.tar.gz 482c0a8e53b27f8922f58d89fb81842ddd9c3ffd120e635838992dc97d535e46b42e7d8c439cb739b7c1d63c8eed27d7e3bcac7126a6a96e56cc13d52f396328
kubernetes-client-windows-386.tar.gz 4446d666f999e979a7245e1b7ebf4817f7bd23aa247a38853a63b9cda473c7d4c2d376a2fd0df13ba15b740bf6b458cac14bd03dbf5a8151fc230e40c08294cf
kubernetes-client-windows-amd64.tar.gz 97f4789f21d10fd3df446e55bc489472dcd534c623bb40dc3cb20fe1edd74c1a89a50ce7caa4e5e0536f3b22d8698060bfe8c46f4adbd0e507349412e52664e8

Server Binaries

filename sha512 hash
kubernetes-server-linux-amd64.tar.gz 25739802a641517a8bbb933b69000a943e8dd38e616b8778149dd0138737abacf377683da2ff35fdd0bbb305b88bc8fc711df20a2585720a43bb674ef36b034f
kubernetes-server-linux-arm.tar.gz c1dbba77a4ff5661eb36c55182a753b88ccc9b89ca31e162b06672126743cfea115b2f8ea8658b12344c36df17958e310c1b8efbdd7800f44f013e1e6f10477d
kubernetes-server-linux-arm64.tar.gz ad346bbe2a053c1106b51e5125698737dc7b76fa3bf439e14d4b4ba1c262678fede9c507c1098aac6e14d2c742c526c8d257fefa95dd3bbb1dff959e1dc7b9aa
kubernetes-server-linux-ppc64le.tar.gz 49f9bd1c751620ecf4b5c152f287d72b36abca21fd1dfe99443d984473c6efa051a910de585c42f5447ef7c18d7dbd905a66c4f09ca6025f45e63f5e96e3ca2f
kubernetes-server-linux-s390x.tar.gz d6be847f2a0358755a69dea26181e5fc1a80ac4939b8b04a3875e1f6693553cad562452bfad21b2e380ddda1839ab846122bc3339d8bec0971f218f6e8f6dce9

Node Binaries

filename sha512 hash
kubernetes-node-linux-amd64.tar.gz 75dc99919d1084d7d471a53ab60c743dc399145c99e83f37c6ba3c241b2c0b2ecc2c0d1b94690ff912e2a15b7c5595aa1d2d24c2fc439e06d85ff0246fb43b89
kubernetes-node-linux-arm.tar.gz 49013a4f01be8086fff332099d94903082688b9b295d2f34468462656da4709360025e9d84b069410c608977ef803079af09af1f1e2678af7cb64e0fc02e9c9d
kubernetes-node-linux-arm64.tar.gz f8c0cb0c089cd1d1977c049002620b8cf748d193c1b76dd1d3aac01ff9273549c06a1e3dfe983dc40a95ee8b0719908e0cdf86ce17359b5f1b2426f2c55799a1
kubernetes-node-linux-ppc64le.tar.gz 48fc02c856a192388877189a43eb1cda531e548bb035f9dfe6a1e3c8d3bcbd0f8e14f29382da45702cb28a91126d13ede42bd6e9159e12ecbd387ca9a58f9a92
kubernetes-node-linux-s390x.tar.gz d7c5f52cf602fd0c0d0f72d4cfe1ceaa4bad70a42f37f21c103f17c3448ceb2396c1bfa521eeeb9eef5f3173d84e4268704a247edd826d765f65e9a29a4f7f72
kubernetes-node-windows-amd64.tar.gz 120afdebe844b06a7437bb9788c3e7ea4fc6352aa18cc6a00e70f44f54664f844429f138870bc15862579da632632dff2e7323be7f627d9c33585a11ad2bed6b

Kubernetes v1.14 Release Notes

1.14 What’s New

Support for Windows Nodes is Graduating to Stable (#116 )

  • Support for Windows Server 2019 for worker nodes and containers
  • Support for out of tree networking with Azure-CNI, OVN-Kubernetes and Flannel
  • Improved support for pods, service types, workload controllers and metrics/quotas to closely match the capabilities offered for Linux containers kubernetes/enhancements: #116 [kep]

Updated Plugin Mechanism for kubectl is Graduating to Stable (#579)

  • Extends functionality to kubectl to support extensions adding new commands as well as overriding specific subcommands (at any depth).
  • Documentation fixes kubernetes/enhancements: #579 [kep]

Durable Local Storage Management is Now GA (#121)

  • Makes locally attached (non-network attached) storage available as a persistent volume source.
  • Allows users to take advantage of the typically cheaper and improved performance of persistent local storage kubernetes/kubernetes: #73525, #74391, #74769 kubernetes/enhancements: #121 [kep]

Pid Limiting is Graduating to Beta (#757)

  • Prevents a pod from starving pid resource
  • Ability to isolate pid resources pod-to-pod and node-to-pod kubernetes/kubernetes: #73651 kubernetes/enhancements: #757 [kep]

Pod Priority and Preemption in Kubernetes (#564)

  • Pod priority and preemption enables Kubernetes scheduler to schedule more important Pods first and when cluster is out of resources, it removes less important pods to create room for more important ones. The importance is specified by priority. kubernetes/kubernetes: #73498, #73555, #74465 kubernetes/enhancements: #564 [kep]

Pod Ready++ (#580)

  • Introduces extension point for external feedback on pod readiness. kubernetes/kubernetes: #74434, kubernetes/enhancements: #580 [kep]

Kubeadm: Automate certificate copy between control planes in HA setups

  • Joining control plane nodes to a HA cluster can now be simplified by enabling the optional automatic copy of certificates from an existing control plane node.
  • You can now use kubeadm init --experimental-upload-certs and kubeadm join --experimental-control-plane --certificate-key. kubernetes/kubeadm: #1373 kubernetes/enhancements: #357 [kep]

Kubeadm: Expose the kubeadm join workflow as phases

  • The kubeadm join command can now be used in phases. Similar to the work that was done for kubeadm init in 1.13, in 1.14 the join phases can be now executed step-by-step/selectively using the kubeadm join phase sub-command. This makes it possible to further customize the workflow of joining nodes to the cluster. kubernetes/kubeadm: #1204 kubernetes/enhancements: kep

Known Issues

  • There is a known issue coredns/coredns#2629 in CoreDNS 1.3.1, wherein if the Kubernetes API shuts down while CoreDNS is connected, CoreDNS will crash. The issue is fixed in CoreDNS 1.4.0 in coredns/coredns#2529.
  • Kubelet might fail to restart if an existing flexvolume mounted pvc contains a large number of directories, or is full. #75019

Urgent Upgrade Notes

(No, really, you MUST read this before you upgrade)

  • kube-apiserver:

    • Default RBAC policy no longer grants access to discovery and permission-checking APIs (used by kubectl auth can-i) to unauthenticated users. Upgraded clusters preserve prior behavior, but cluster administrators wishing to grant unauthenticated users access in new clusters will need to explicitly opt-in to expose the discovery and/or permission-checking APIs:
      • kubectl create clusterrolebinding anonymous-discovery --clusterrole=system:discovery --group=system:unauthenticated
      • kubectl create clusterrolebinding anonymous-access-review --clusterrole=system:basic-user --group=system:unauthenticated
    • The deprecated --storage-versions flag has been removed. The storage versions will always be the default value built-in the kube-apiserver binary. (#67678, @caesarxuchao)
    • The deprecated --repair-malformed-updates flag has been removed (#73663, @danielqsj)
    • The /swaggerapi/* schema docs, deprecated since 1.7, have been removed in favor of the /openapi/v2 schema docs. (#72924, @liggitt)
    • The /swagger.json and /swagger-2.0.0.pb-v1 schema documents, deprecated since v1.10, have been removed in favor of /openapi/v2 (#73148, @liggitt)
    • kube-apiserver now only aggregates openapi schemas from /openapi/v2 endpoints of aggregated API servers. The fallback to aggregate from /swagger.json has been removed. Ensure aggregated API servers provide schema information via /openapi/v2 (available since v1.10). (#73441, @roycaihw)
    • The OpenAPI definitions with the prefix "io.k8s.kubernetes.pkg" (deprecated since 1.9) have been removed. (#74596, @sttts)
    • The ValidateProxyRedirects feature was promoted to Beta and enabled by default. This feature restricts redirect-following from the apiserver to same-host redirects. If nodes are configured to respond to CRI streaming requests on a different host interface than what the apiserver makes requests on (only the case if not using the built-in dockershim & setting the kubelet flag --redirect-container-streaming=true), then these requests will be broken. In that case, the feature can be temporarily disabled until the node configuration is corrected. We suggest setting --redirect-container-streaming=false on the kubelet to avoid issues.(#72552, @tallclair)
  • kubectl

    • The deprecated --show-all flag to kubectl get has been removed (#69255, @Pingan2017)
  • kubelet

    • The deprecated --experimental-fail-swap-on flag has been removed (#69552, @Pingan2017)
    • Health check (liveness & readiness) probes using an HTTPGetAction will no longer follow redirects to different hostnames from the original probe request. Instead, these non-local redirects will be treated as a Success (the documented behavior). In this case an event with reason "ProbeWarning" will be generated, indicating that the redirect was ignored. If you were previously relying on the redirect to run health checks against different endpoints, you will need to perform the healthcheck logic outside the Kubelet, for instance by proxying the external endpoint rather than redirecting to it. (#75416, @tallclair)
  • client-go

    • The deprecated versionless API group accessors (like clientset.Apps()) have been removed. Use an explicit version instead (like clientset.AppsV1()) (#74422, @liggitt)
    • The disk-cached discovery client is moved from k8s.io/client-go/discovery to k8s.io/client-go/discovery/cached/disk. The memory-cached discovery client is moved from k8s.io/client-go/discovery/cached to k8s.io/client-go/discovery/cached/memory. (#72214, @caesarxuchao)
  • kubeadm

    • kubeadm alpha preflight and kubeadm alpha preflight node are removed; you can now use kubeadm join phase preflight (#73718, @fabriziopandini)
  • The deprecated taints node.alpha.kubernetes.io/notReady and node.alpha.kubernetes.io/unreachable are no longer supported or adjusted. These uses should be replaced with node.kubernetes.io/not-ready and node.kubernetes.io/unreachable (#73001, @shivnagarajan)

  • Any Prometheus queries that match pod_name and container_name labels (e.g. cadvisor or kubelet probe metrics) should be updated to use pod and container instead. pod_name and container_name labels will be present alongside pod and container labels for one transitional release and removed in the future. (#69099, @ehashman)

Deprecations

  • kubectl

    • kubectl convert is deprecated and will be removed in v1.17.
    • The --export flag for the kubectl get command is deprecated and will be removed in v1.18. (#73787, @soltysh)
  • kubelet

    • OS and Arch information is now recorded in kubernetes.io/os and kubernetes.io/arch labels on Node objects. The previous labels (beta.kubernetes.io/os and beta.kubernetes.io/arch) are still recorded, but are deprecated and targeted for removal in v1.18. (#73333, @yujuhong)
    • The --containerized flag is deprecated and will be removed in a future release (#74267, @dims)
  • hyperkube

    • The --make-symlinks flag is deprecated and will be removed in a future release. (#74975, @dims)
  • API

    • Ingress resources are now available via networking.k8s.io/v1beta1. Ingress resources in extensions/v1beta1 are deprecated and will no longer be served in v1.18. Existing persisted data is available via the new API group/version (#74057, @liggitt)
    • NetworkPolicy resources will no longer be served from extensions/v1beta1 in v1.16. Migrate use to the networking.k8s.io/v1 API, available since v1.8. Existing persisted data can be retrieved via the networking.k8s.io/v1 API.
    • PodSecurityPolicy resources will no longer be served from extensions/v1beta1 in v1.16. Migrate to the policy/v1beta1 API, available since v1.10. Existing persisted data can be retrieved via the policy/v1beta1 API.
    • DaemonSet, Deployment, and ReplicaSet resources will no longer be served from extensions/v1beta1, apps/v1beta1, or apps/v1beta2 in v1.16. Migrate to the apps/v1 API, available since v1.9. Existing persisted data can be retrieved via the apps/v1 API.
    • PriorityClass resources have been promoted to scheduling.k8s.io/v1 with no changes. The scheduling.k8s.io/v1beta1 and scheduling.k8s.io/v1alpha1 versions are now deprecated and will stop being served by default in v1.17. (#73555, #74465, @bsalamat)
    • The export query parameter for list API calls is deprecated and will be removed in v1.18 (#73783, @deads2k)
  • The following features are now GA, and the associated feature gates are deprecated and will be removed in v1.15:

    • CustomPodDNS
    • HugePages
    • MountPropagation
    • PersistentLocalVolumes
  • CoreDNS: The following directives or keywords are deprecated and will be removed in v1.15:

    • upstream option of kubernetes plugin, becoming default behavior in v1.15.
    • proxy plugin replaced by forward plugin

Removed and deprecated metrics

Removed metrics

  • reflector_items_per_list
  • reflector_items_per_watch
  • reflector_last_resource_version
  • reflector_list_duration_seconds
  • reflector_lists_total
  • reflector_short_watches_total
  • reflector_watch_duration_seconds
  • reflector_watches_total

Deprecated metrics

  • rest_client_request_latency_seconds -> rest_client_request_duration_seconds
  • apiserver_proxy_tunnel_sync_latency_secs -> apiserver_proxy_tunnel_sync_duration_seconds
  • scheduler_scheduling_latency_seconds -> scheduler_scheduling_duration_seconds
  • kubelet_pod_worker_latency_microseconds -> kubelet_pod_worker_duration_seconds
  • kubelet_pod_start_latency_microseconds -> kubelet_pod_start_duration_seconds
  • kubelet_cgroup_manager_latency_microseconds -> kubelet_cgroup_manager_duration_seconds
  • kubelet_pod_worker_start_latency_microseconds -> kubelet_pod_worker_start_duration_seconds
  • kubelet_pleg_relist_latency_microseconds -> kubelet_pleg_relist_duration_seconds
  • kubelet_pleg_relist_interval_microseconds -> kubelet_pleg_relist_interval_seconds
  • kubelet_eviction_stats_age_microseconds -> kubelet_eviction_stats_age_seconds
  • kubelet_runtime_operations -> kubelet_runtime_operations_total
  • kubelet_runtime_operations_latency_microseconds -> kubelet_runtime_operations_duration_seconds
  • kubelet_runtime_operations_errors -> kubelet_runtime_operations_errors_total
  • kubelet_device_plugin_registration_count -> kubelet_device_plugin_registration_total
  • kubelet_device_plugin_alloc_latency_microseconds -> kubelet_device_plugin_alloc_duration_seconds
  • docker_operations -> docker_operations_total
  • docker_operations_latency_microseconds -> docker_operations_latency_seconds
  • docker_operations_errors -> docker_operations_errors_total
  • docker_operations_timeout -> docker_operations_timeout_total
  • network_plugin_operations_latency_microseconds -> network_plugin_operations_latency_seconds
  • sync_proxy_rules_latency_microseconds -> sync_proxy_rules_latency_seconds
  • apiserver_request_count -> apiserver_request_total
  • apiserver_request_latencies -> apiserver_request_latency_seconds
  • apiserver_request_latencies_summary -> apiserver_request_latency_seconds
  • apiserver_dropped_requests -> apiserver_dropped_requests_total
  • etcd_helper_cache_hit_count -> etcd_helper_cache_hit_total
  • etcd_helper_cache_miss_count -> etcd_helper_cache_miss_total
  • etcd_helper_cache_entry_count -> etcd_helper_cache_entry_total
  • etcd_request_cache_get_latencies_summary -> etcd_request_cache_get_latency_seconds
  • etcd_request_cache_add_latencies_summary -> etcd_request_cache_add_latency_seconds
  • etcd_request_latencies_summary -> etcd_request_latency_seconds
  • transformation_latencies_microseconds -> transformation_latencies_seconds
  • data_key_generation_latencies_microseconds -> data_key_generation_latencies_seconds

Notable Features

  • Increased the histogram resolution of the API server client certificate to accommodate short-lived (< 6h) client certificates. (#74806, @mxinden)
  • Updated to use golang 1.12 (#74632, @cblecker)
  • The RunAsGroup feature has been promoted to beta and enabled by default. PodSpec and PodSecurityPolicy objects can be used to control the primary GID of containers on supported container runtimes. (#73007, @krmayankk)
  • Added the same information to an init container as a standard container in a pod when using PodPresets. (#71479, @soggiest)
  • kube-conformance image will now run ginkgo with the --dryRun flag if the container is run with the environment variable E2E_DRYRUN set. (#74731, @johnSchnake)
  • Introduced dynamic volume provisioning shim for CSI migration (#73653, @ddebroy)
  • Applied resources from a directory containing kustomization.yaml (#74140, @Liujingfang1)
  • kubeadm: Allowed to download certificate secrets uploaded by init or upload-certs phase, allowing to transfer certificate secrets (certificates and keys) from the cluster to other master machines when creating HA deployments. (#74168, @ereslibre)
  • The --quiet option to kubectl run now suppresses resource deletion messages emitted when the --rm option is specified. (#73266, @awh)
  • Added Custom Resource support to kubectl autoscale (#72678, @rmohr)
  • Cinder volume limit can now be configured from node too (#74542, @gnufied)
  • It is now possible to combine the -f and -l flags in kubectl logs (#67573, @m1kola)
  • New conformance tests added for API Aggregation. (#63947, @jennybuckley)
  • Moved fluentd-elasticsearch addon images to community controlled location (#73819, @coffeepac)
  • Removed local etcd members from the etcd cluster when kubeadm reset (#74112, @pytimer)
  • kubeadm will now not fail preflight checks when running on >= 5.0 Linux kernel (#74355, @brb)
  • Scheduler cache snapshot algorithm has been optimized to improve scheduling throughput. (#74041, @bsalamat)
  • It is now possible to upload certificates required to join a new control-plane to kubeadm-certs secret using the flag --experimental-upload-certs on init or upload-certs phase. (#73907, @yagonobre) @RobertKrawitz)
  • kubectl auth reconcile now outputs details about what changes are being made (#71564, @liggitt)
  • Added Kustomize as a subcommand in kubectl (#73033, @Liujingfang1)
  • Added kubelet_node_name metrics. (#72910, @danielqsj)
  • Updated AWS SDK to v1.16.26 for ECR PrivateLink support (#73435, @micahhausler)
  • Expanded kubectl wait to work with more types of selectors. (#71746, @rctl) (#72832, @MrHohn)
  • Added configuration for AWS endpoint fine control: (#72245, @ampsingram)
  • The CoreDNS configuration now has the forward plugin for proxy in the default configuration instead of the proxy plugin. (#73267, @rajansandeep)
  • Added alpha field storageVersionHash to the discovery document for each resource. Its value must be treated as opaque by clients. Only equality comparison on the value is valid. (#73191, @caesarxuchao)
  • If you are running the cloud-controller-manager and you have the pvlabel.kubernetes.io alpha Initializer enabled, you must now enable PersistentVolume labeling using the PersistentVolumeLabel admission controller instead. You can do this by adding PersistentVolumeLabel in the --enable-admission-plugins kube-apiserver flag. (#73102, @andrewsykim)
  • kubectl supports copying files with wild card (#72641, @dixudx)
  • kubeadm now attempts to detect an installed CRI by its usual domain socket, so that --cri-socket can be omitted from the command line if Docker is not used and there is a single CRI installed. (#69366, @rosti)
  • CSINodeInfo and CSIDriver CRDs have been installed in the local cluster. (#72584, @xing-yang)
  • Node OS/arch labels have been promoted to GA (#73048, @yujuhong)
  • Added support for max attach limit for Cinder (#72980, @gnufied)
  • Enabled mTLS encryption between etcd and kube-apiserver in GCE (#70144, @wenjiaswe)
  • Added ResourceVersion as a precondition for delete in order to ensure a delete fails if an unobserved change happens to an object. (#74040, @ajatprabha)
  • There is now support for collecting pod logs under /var/log/pods/NAMESPACE_NAME_UID to stackdriver with k8s_pod resource type. (#74502, @Random-Liu)
  • Changed CRI pod log directory from /var/log/pods/UID to /var/log/pods/NAMESPACE_NAME_UID. (#74441, @Random-Liu)
  • RuntimeClass has been promoted to beta, and is enabled by default. (#75003, @tallclair)
  • New "dry_run" metric label (indicating the value of the dryRun query parameter) has been added into the metrics:
  • apiserver_request_total
  • apiserver_request_duration_seconds New "APPLY" value for the "verb" metric label which indicates a PATCH with "Content-Type: apply-patch+yaml". This value is experimental and will only be present if the ServerSideApply alpha feature is enabled. (#74997, @jennybuckley)
  • GCE: bumped COS image version to cos-beta-73-11647-64-0 (#75149, @yguo0905)
  • Added alpha support for ephemeral CSI inline volumes that are embedded in pod specs. (#74086, @vladimirvivien)

API Changes

  • [CRI] Added a new field called runtime_handler into PodSandbox and PodSandboxStatus to track the RuntimeClass information of a pod. (#73833, @haiyanmeng)

Detailed Bug Fixes And Changes

API Machinery

  • client-go: PortForwarder.GetPorts() now contain correct local port if no local port was initially specified when setting up the port forwarder (#73676, @martin-helmich)
  • Fixed an issue with missing apiVersion/kind in object data sent to admission webhooks (#74448, @liggitt)
  • Prometheus metrics for crd_autoregister, crd_finalizer and crd_naming_condition_controller are exported. (#71767, @roycaihw)
  • Fixed admission metrics in seconds. (#72343, @danielqsj)
  • When a watch is closed by an HTTP2 load balancer and we are told to go away, skip printing the message to stderr by default.
  • Spedup kubectl by >10 when calling out to kube-apiserver for discovery information. (#73345, @sttts)
  • Fixed watch to not send the same set of events multiple times causing watcher to go back in time (#73845, @wojtek-t) (#73277, @smarterclayton)
  • Fix kube-apiserver not to create default/kubernetes service endpoints before it reports readiness via the /healthz and therefore is ready to serve requests. Also early during startup old endpoints are remove which might be left over from a previously crashed kube-apiserver. (#74668, @sttts)
  • Add a configuration field to shorten the timeout of validating/mutating admission webhook call. The timeout value must be between 1 and 30 seconds. Default to 30 seconds when unspecified. (#74562, @roycaihw)
  • The apiserver, including both the kube-apiserver and apiservers built with the generic apiserver library, will now return 413 RequestEntityTooLarge error if a json patch contains more than 10,000 operations. (#74000, @caesarxuchao)
  • Fixed an error processing watch events when running skewed apiservers (#73482, @liggitt)
  • jsonpath expressions containing [start:end:step] slice are now evaluated correctly (#73149, @liggitt)
  • metadata.deletionTimestamp is no longer moved into the future when issuing repeated DELETE requests against a resource containing a finalizer. (#73138, @liggitt)
  • Fixed kube-apiserver not to create default/kubernetes service endpoints before it reports readiness via the /healthz and therefore is ready to serve requests. Also early during startup old endpoints are remove which might be left over from a previously crashed kube-apiserver. (#74668, @sttts)
  • watch.Until now works for long durations. (#67350, @tnozicka)
  • Added duration metric for CRD webhook converters. (#74376, @mbohlool)
  • Fixed keymutex issues which may crash in some platforms. (#74348, @danielqsj)
  • Considerably reduced the CPU load in kube-apiserver while aggregating OpenAPI specifications from aggregated API servers. (#71223, @sttts)
  • Fixed graceful apiserver shutdown to not drop outgoing bytes before the process terminates. (#72970, @sttts)

Apps

  • Added deleting pods created by DaemonSet assigned to not existing nodes. (#73401, @krzysztof-jastrzebski)
  • Pod eviction now honors graceful deletion by default if no delete options are provided in the eviction request. (#72730, @liggitt)

Auth

  • Added kubectl auth can-i --list option, which allows users to know what actions they can do in specific namespaces. (#64820, @WanLinghao)
  • The rules field in RBAC Role and ClusterRole objects is now correctly reported as optional in the openapi schema. (#73250, @liggitt)
  • system:kube-controller-manager and system:kube-scheduler users are now permitted to perform delegated authentication/authorization checks by default RBAC policy (#72491, @liggitt)
  • Error messages returned in authentication webhook status responses are now correctly included in the apiserver log (#73595, @liggitt)
  • Fixed use of webhook admission plugins with multi-version custom resources (#74154, @mbohlool)

AWS

  • Prevented AWS Network Load Balancer security groups ingress rules to be deleted by ensuring target groups are tagged. (#73594, @masterzen)
  • AWS ELB health checks will now use HTTPS/SSL protocol for HTTPS/SSL backends. (#70309, @2rs2ts)

Azure

CLI

  • Fixed --help flag parsing (#74682, @soltysh)

  • Fixed a bug where kubectl describe cannot obtain the event messages for a static pod (#74156, @gaorong)

  • Fixed panic when performing a set env operation on a --local resource (#65636, @juanvallejo)

  • Missing directories listed in a user's PATH are no longer considered errors and are instead logged by the kubectl plugin list command when listing available plugins. (#73542, @juanvallejo)

  • Now users can get object info like:

      a. kubectl get pod test-pod -o custom-columns=CONTAINER:.spec.containers[0:3].name
      b. kubectl get pod test-pod -o custom-columns=CONTAINER:.spec.containers[-2:].name

(#73063, @WanLinghao)

  • The kubectl api-resources command will no longer fail to display any resources on a single failure (#73035, @juanvallejo)
  • kubectl now loads config file once and uses persistent client config (#71117, @dixudx)
  • Printed SizeLimit of EmptyDir in kubectl describe pod outputs. (#69279, @dtaniwaki)
  • kubectl delete --all-namespaces is now a recognized flag. (#73716, @deads2k)

Cloud Provider

  • Fixed a bug that caused PV allocation on non-English vSphere installations to fail (#73115, @alvaroaleman)

Cluster Lifecycle

  • kubeadm: fixed nil pointer dereference caused by a bug in url parsing (#74454, @bart0sh)
  • CoreDNS adds readinessProbe which prevents loadbalancing to unready pods, and also allows rolling updates to work as expected. (#74137, @rajansandeep)
  • kubeadm no longer allows using v1alpha3 configs for anything else than converting them to v1beta1. (#74025, @rosti)
  • kubeadm: now allows the usage of --kubeconfig-dir and --config flags on kubeadm init (#73998, @yagonobre)
  • kubeadm: all master components are now exclusively relying on the PriorityClassName pod spec for annotating them as cluster critical components. Since scheduler.alpha.kubernetes.io/critical-pod annotation is no longer supported by Kubernetes 1.14 this annotation is no longer added to master components. (#73857, @ereslibre)
  • kubeadm no longer dumps backtrace if it fails to remove the running containers on reset. (#73951, @rosti)
  • kubeadm: fixed a bug in the underlying library for diff related to characters like '%' (#73941, @neolit123)
  • Scale max-inflight now limits together with master VM sizes. (#73268, @wojtek-t)
  • kubeadm reset: fixed a crash caused by the absence of a configuration file (#73636, @bart0sh)
  • CoreDNS is now version 1.3.1 (#73610, @rajansandeep)
  • kubeadm: When certificates are present in joining a new control plane now ensures that they match at least the required SANs (#73093, @ereslibre)
  • kubeadm: added back --cert-dir option for kubeadm init phase certs sa (#73239, @mattkelly)
  • kubeadm: now explicitly waits for etcd to have grown when joining a new control plane (#72984, @ereslibre)
  • kubeadm: now pulls images when joining a new control plane instance (#72870, @MalloZup)
  • Exited kube-proxy when configuration file changes (#59176, @dixudx)
  • kube-addon-manager was updated to v9.0, and now uses kubectl v1.13.2 and prunes workload resources via the apps/v1 API (#72978, @liggitt)
  • kubeadm: Now allows certain certs/keys to be missing on the secret when transferring secrets using --experimental-upload-certs feature (#75415, @ereslibre)

GCP

  • Fixed liveness probe in fluentd-gcp cluster addon (#74522, @Pluies)
  • Reduced GCE log rotation check from 1 hour to every 5 minutes. Rotation policy is unchanged (new day starts, log file size > 100MB). (#72062, @jpbetz)

Network

Node

  • Fixed help message for --container-runtime-endpoint: only unix socket is support on Linux. (#74712, @feiskyer)
  • Image garbage collection no longer fails for images with only one tag but more than one repository associated. (#70647, @corvus-ch)
  • Re-issued Allocate grpc calls before starting a container that requests device-plugin resources if the cached state is missing. (#73824, @jiayingz)
  • [CRI] Added a new field called runtime_handler into PodSandbox and PodSandboxStatus to track the RuntimeClass information of a pod. (#73833, @haiyanmeng)
  • Kubelet now tries to stop containers in unknown state once before restart or remove. (#73802, @Random-Liu)
  • When pleg channel is full, events are now discarded and count is recorded (#72709, @changyaowei)
  • Fixed the unexpected NotReady status when Node's iops is full if the runtime is dockershim. (#74389, @answer1991)
  • Fixed #73264 cpuPeriod was not reset, but used as set via flag, although it was disabled via alpha gate (#73342, @szuecs)
  • Updated kubelet CLI summary documentation and generated webpage (#73256, @deitch)
  • Set a low oom_score_adj for containers in pods with system-critical priorities (#73758, @sjenning)
  • kubelet: Resolved hang/timeout issues when running large numbers of pods with unique ConfigMap/Secret references (#74755, @liggitt)
  • Events reported for container creation, start, and stop now report the container name in the message and are more consistently formatted. (#73892, @smarterclayton)
  • Removed stale OutOfDisk condition from kubelet side (#72507, @dixudx)
  • Fixed the setting of NodeAddresses when using the vSphere CloudProvider and nodes that have multiple IP addresses. (#70805, @danwinship)
  • Fixed dockershim panic issues when deleting docker images. (#75367, @feiskyer)
  • Kubelet no longer watches ConfigMaps and Secrets for terminated pods, in worst scenario causing it to not be able to send other requests to kube-apiserver (#74809, @oxddr)
  • A new TaintNodesByCondition admission plugin taints newly created Node objects as "not ready", to fix a race condition that could cause pods to be scheduled on new nodes before their taints were updated to accurately reflect their reported conditions. This admission plugin is enabled by default if the TaintNodesByCondition feature is enabled. (#73097, @bsalamat)
  • kubelet now accepts pid=<number> in the --system-reserved and --kube-reserved options to ensure that the specified number of process IDs will be reserved for the system as a whole and for Kubernetes system daemons respectively. Please reference Kube Reserved and System Reserved in Reserve Compute Resources for System Daemons in the Kubernetes documentation for general discussion of resource reservation. To utilize this functionality, you must set the feature gate SupportNodePidsLimit=true (#73651

Scheduling

  • Improved fairness of the scheduling queue by placing pods which are attempted recently behind other pods with the same priority. (#73700, @denkensk)
  • Improved scheduler robustness to ensure that unschedulable pods are reconsidered for scheduling when appropriate. (#73700, #72558, @denkensk, #73078, @Huang-Wei)

Storage

  • Fixed scanning of failed iSCSI targets. (#74306, @jsafrane)
  • StorageOS volume plugin updated to fix an issue where volume mount succeeds even if request to mount via StorageOS API fails. (#69782, @darkowlzz)
  • Ensured directories on volumes are group-executable when using fsGroup (#73533, @mxey)
  • Updated CSI version to 1.1 (#75391, @gnufied)
  • Ensured that volumes get provisioned based on the zone information provided in allowedTopologies. (#72731, @skarthiksrinivas)
  • Extended the VolumeSubpathEnvExpansion alpha feature to support environment variable expansion (#71351, @kevtaylor)
  • Fixed a bug that prevented deletion of dynamically provisioned volumes in Quobyte backends. (#68925, @casusbelli)

Testing

  • e2e storage tests now run faster and are easier to read (#72434, @pohly)
  • e2e.test now rejects unknown --provider values instead of merely warning about them. An empty provider name is not accepted anymore and was replaced by skeleton (a provider with no special behavior). (#73402, @pohly)
  • Updated to go1.11.5 (#73326, @ixdy)
  • Updated to use go1.12.1 (#75413, @BenTheElder)
  • e2e tests that require SSH may now be used against clusters that have nodes without external IP addresses by setting the environment variable KUBE_SSH_BASTION to the host:port of a machine that is allowed to SSH to those nodes. The same private key that the test would use is used for the bastion host. The test connects to the bastion and then tunnels another SSH connection to the node. (#72286, @smarterclayton)
  • PidPressure now evicts pods from lowest priority to highest priority (#72844, @dashpole)
  • Split up the mondo kubernetes-test tarball into kubernetes-test-portable and kubernetes-test-{OS}-{ARCH} tarballs. (#74065, @ixdy)

VMware

Windows

Support for Windows nodes and Windows containers went going stable.

Support for Group Managed Service Accounts (GMSA) for Windows containers in Kubernetes. GMSA are a specific type of Active Directory account that provides automatic password management, simplified service principal name (SPN) management, and the ability to delegate the management to other administrators across multiple servers.

  • Fixed smb remount and unmount issues on Windows (#73661, @andyzhangx, #75087, @andyzhangx)
  • Added network stats for Windows nodes and containers (#74788, @feiskyer)
  • The new test [sig-network] DNS should now provide /etc/hosts entries for the cluster [LinuxOnly] [Conformance] will validate the host entries set in the /etc/hosts file (pod's FQDN and hostname), which should be managed by Kubelet. (#72729, @bclau)
  • Allowed the kubelet to pass Windows GMSA credentials down to Docker (#73726, @wk8)
  • Added kube-proxy support for overlay networking and DSR in Windows and new flags for network-name, source-vip, and enable-dsr. (#70896, @ksubrmnn)
  • windows: Ensured graceful termination when being run as windows service (#73292, @steffengy)
  • vSphere cloud provider now correctly retrieves the VM's UUID when running on Windows (#71147, @benmoss)
  • Kubelet: added usageNanoCores from CRI stats provider (#73659, @feiskyer)
  • Introduced support for Windows nodes into the cluster bringup scripts for GCE. (#73442, @pjh)
  • Added network stats for Windows nodes and pods. (#70121, @feiskyer)
  • CoreDNS is only officially supported on Linux at this time. As such, when kubeadm is used to deploy this component into your kubernetes cluster, it will be restricted (using nodeSelectors) to run only on nodes with that operating system. This ensures that in clusters which include Windows nodes, the scheduler will not ever attempt to place CoreDNS pods on these machines, reducing setup latency and enhancing initial cluster stability. (#69940, @MarcPow)

External Dependencies

  • Default etcd server and client have been updated to v3.3.10. (#71615, #70168)

  • The list of validated docker versions has changed. 1.11.1 and 1.12.1 have been removed. The current list is 1.13.1, 17.03, 17.06, 17.09, 18.06, 18.09. (#72823, #72831)

  • The default Go version was updated to 1.12.1. (#75422)

  • CNI has been updated to v0.7.5 (#75455)

  • CSI has been updated to v1.1.0. (#75391)

  • The dashboard add-on has been updated to v1.10.1. (#72495)

  • Cluster Autoscaler has been updated to v1.14.0 (#75480)

  • kube-dns is unchanged at v1.14.13 since Kubernetes 1.12 (#68900)

  • Influxdb is unchanged at v1.3.3 since Kubernetes 1.10 (#53319)

  • Grafana is unchanged at v4.4.3 since Kubernetes 1.10 (#53319)

  • Kibana has been upgraded to v6.6.1. (#71251)

  • CAdvisor has been updated to v0.33.1 (#75140)

  • fluentd-gcp-scaler is unchanged at v0.5.0 since Kubernetes 1.13 (#68837)

  • Fluentd in fluentd-elasticsearch has been upgraded to v1.3.3 (#71180)

  • fluentd-elasticsearch has been updated to v2.4.0 (#71180)

  • The fluent-plugin-kubernetes_metadata_filter plugin in fluentd-elasticsearch has been updated to v2.1.6 (#71180)

  • fluentd-gcp is unchanged at v3.2.0 since Kubernetes 1.13 (#70954)

  • OIDC authentication is unchanged at coreos/go-oidc v2 since Kubernetes 1.10 (#58544)

  • Calico is unchanged at v3.3.1 since Kubernetes 1.13 (#70932)

  • crictl on GCE is unchanged at v1.12.0 since Kubernetes 1.13 (#69033)

  • CoreDNS has been updated to v1.3.1 (#73610)

  • event-exporter has been updated to v0.2.3 (#67691)

  • Es-image has been updated to Elasticsearch 6.6.1 (#71252)

  • metrics-server remains unchanged at v0.3.1 since Kubernetes 1.12 (#68746)

  • GLBC remains unchanged at v1.2.3 since Kubernetes 1.12 (#66793)

  • Ingress-gce remains unchanged at v1.2.3 since Kubernetes 1.12 (#66793)

  • ip-masq-agen remains unchanged at v2.1.1 since Kubernetes 1.12 (#67916)

  • v1.14.0-rc.1

  • v1.14.0-beta.2

  • v1.14.0-beta.1

  • v1.14.0-alpha.3

  • v1.14.0-alpha.2

  • v1.14.0-alpha.1

v1.14.0-rc.1

Documentation

Downloads for v1.14.0-rc.1

filename sha512 hash
kubernetes.tar.gz 5cb5e8b14b301864063fd7531ab3b755fea054f540c55ecce70ac49fb59193488575eb42ba89c8b4a44f6f2d005602ffc50ac286354a16df27637dd2e05f90f0
kubernetes-src.tar.gz 395424c2bcdb5e242995b18e8d6e5c00002ce2cb5a3964c28da0a4a181fada73ffceaccedb1fa9799be9b3c4fb5b451010cba65af4d7385c25c8c8f0298218fc

Client Binaries

filename sha512 hash
kubernetes-client-darwin-386.tar.gz ea304f3f8188db30cddd5423b25dc434b8f05315103f773619a65f83bee872581d83d5498a5f36a3064815e68998746cf661802eab36bfe96217253bac7e751c
kubernetes-client-darwin-amd64.tar.gz 17e106b63067429b9228a4879a7350c01ae98650ef2e6fcc23d00415c2e3a7c340abd5bcfc4b976f3d737d9268159ef5e5e7b08757371daad637e721a2ffb4a9
kubernetes-client-linux-386.tar.gz 406323ea4cbd524807e73b9a2f4eb0a813730b262402c224e5076080b84452137521e5782056e39bf6017bda8ef9e797ed497d51a653ed6822357f43d86bb0ae
kubernetes-client-linux-amd64.tar.gz 293747816e5da30c53ca29f27479fae880404edd5fca413af165e52cc7ea7ff26312bb3916896eceea75af6b232647268324da76d2d30fa2a4a688557427f7c7
kubernetes-client-linux-arm.tar.gz 98b6749c367282048ecd0c5b70ae8b7dafad82c599e359cbbe782a530bbd7bdb84a198577251d6aedbc39fb4e0297bd929d7e988eb557556f523073227375b09
kubernetes-client-linux-arm64.tar.gz 0a650c53946ebd9c38705df36efabbf1fbe3da1acdf418cc4ab881530e6a9089b45ef36bd4a89467106561e2165a00cc2c47f791d9ba422f36544bfa4b1e3b68
kubernetes-client-linux-ppc64le.tar.gz 807263f316f46c9334ec4dad84895cfe2b942ac4ed9aacf3ec8a63193e0d0a6ecaf573d00dfea278d1e552fe91e9474f534ea6798700dbdc84f1d9556ecfda3e
kubernetes-client-linux-s390x.tar.gz 5589562ce0da49f4987388e3e2b6fcf29e92859fae65fb57cdd61bba20ff574c7f3e07f4fa26bb63789f927690ce19710803e0d2e3324bd2c80ddc4925ea973d
kubernetes-client-windows-386.tar.gz 101245cd70221b443150be046e5b5a2c6c83334085bb17f603f59bb68c5960f353d57b6761acf052449f4fb057a5525978cdd7995d06134ab98f8628c23aac0d
kubernetes-client-windows-amd64.tar.gz b1c3802f60cc70ebc1258cb8fc4ffa1154ecee8fda473b033be4f9d1b187354cfd75d085a1ad45a35fcc42ad640c4fecaa2496cb48fad649a774f40a5150825f

Server Binaries

filename sha512 hash
kubernetes-server-linux-amd64.tar.gz 8e09465aab0a1d1ec39afc98af17de9b5de99b763c0c5feb2dc824f2bbae25edb690e9d162fd44f5155bea392bd229f544850fe19e767a8b342050f4bfafc2cc
kubernetes-server-linux-arm.tar.gz 9409c368e1f9f26e633b7df5d6c90435394931d48a21f4ad1548d172b18ac462a859019cc66dba4df69d3b10702820c9a3e1bd5c469646b1db581e52e79e035d
kubernetes-server-linux-arm64.tar.gz 3893290dc11ca71746fee77a44a607ad9e02036bab56b7fc3be247b71b2cf5b3f639fa41317a713abbe9a997abf80c7113ff1155482d0cc04a318ec8beccf869
kubernetes-server-linux-ppc64le.tar.gz 8ac1e70cb39aeb5b1fe92c3ebba428db2036be739f462cc2f876f17dc71a01ba263b5611a15d95e5934e2fc7aa92042bc9b2fe65cf459263be90f5fbb5d83a15
kubernetes-server-linux-s390x.tar.gz 80576cb6cc3a69c4fc0a0358dee5772ecb38437c534a3454c9613426417d4af3c527a0809cce4d46653a7b001c58033b06326c80c498d17387569d22d3ca9b22

Node Binaries

filename sha512 hash
kubernetes-node-linux-amd64.tar.gz 71cd5dd6075a2eea851231a5a855e58b3f479d83358defafd068dd1d09e5b2c426a8a046ee621de91e17d7ecd67465911b93549088bf27a41c6e6b77d692a8c7
kubernetes-node-linux-arm.tar.gz 48babf4a52013c2bc69049167579ba1bc70c769b782a2704c9dfcf44a6a8a72f07e0789af347135ee4797f2bf1a216c348a9a4a26be71855e95e8387bf4e2aac
kubernetes-node-linux-arm64.tar.gz cc19a52beaa7440ca7581e85d1e10137e93c2decdb7d7d7919e7fdcded63e4d94b3434513ca881dae844dd1eb1e2fe98ea5332fad5ffced846f729894ecf0ed1
kubernetes-node-linux-ppc64le.tar.gz 241bf20ae366384efa0fb3798e07e1cdd3d4ea7ba91c146ef7761fb0b93a8514a0dd91f9eb47999ae263d6793e0577c2bbecf88548bcae06cedb437331d6d3bc
kubernetes-node-linux-s390x.tar.gz 37ce3e021073c6c10cf240fb2c3f9a7ab35ece3c0b4a9fecbbdf790eb348b168d179824f3a8eb57d56f962b64f8a6a71925152c087f5bfe43b004cbfae65674d
kubernetes-node-windows-amd64.tar.gz 3248db12c274171f094ef41d6a59523aec35655ba35d151576ff2d2089c269f345e2c0646d585d7c04c440c6b4e7379c499bc8f51fcc8b91388e3dc4d1b6a899

Changelog since v1.14.0-beta.2

Action Required

  • ACTION REQUIRED: Health check (liveness & readiness) probes using an HTTPGetAction will no longer follow redirects to different hostnames from the original probe request. Instead, these non-local redirects will be treated as a Success (the documented behavior). In this case an event with reason "ProbeWarning" will be generated, indicating that the redirect was ignored. If you were previously relying on the redirect to run health checks against different endpoints, you will need to perform the healthcheck logic outside the Kubelet, for instance by proxying the external endpoint rather than redirecting to it. (#75416, @tallclair)

Other notable changes

  • Restores --username and --password flags to kubectl (#75451, @liggitt)
  • fix race condition issue for smb mount on windows (#75371, @andyzhangx)
  • UDP Service conntrack entries for ExternalIPs are now correctly cleared when endpoints are added (#75265, @JacobTanenbaum)
  • kubeadm: the kubeadm init output now provides join control-plane example only when the preconditions for joining a control plane are satisfied (#75420, @fabriziopandini)
  • Fix dockershim panic issues when deleting docker images. (#75367, @feiskyer)
  • kubeadm: Allow certain certs/keys to be missing on the secret when transferring secrets using --experimental-upload-certs feature (#75415, @ereslibre)
  • Update to use go1.12.1 (#75413, @BenTheElder)
  • Update CSI version to 1.1 (#75391, @gnufied)
  • Ensure ownership when deleting a load balancer security group (#74311, @hpedrorodrigues)
  • kubelet: updated logic of verifying a static critical pod. (#75144, @Huang-Wei)
  • Allow disable outbound SNAT when Azure standard load balancer is used together with outbound rules. (#75282, @feiskyer)
  • Add ResourceVersion as a precondition for delete in order to ensure a delete fails if an unobserved change happens to an object. (#74040, @ajatprabha)
  • Services of type=LoadBalancer which have no endpoints will now immediately ICMP reject connections, rather than time out. (#74394, @thockin)
  • Ensure Azure load balancer cleaned up on 404 or 403 when deleting LoadBalancer services. (#75256, @feiskyer)

v1.14.0-beta.2

Documentation

Downloads for v1.14.0-beta.2

filename sha512 hash
kubernetes.tar.gz c1d5f2615c3319fc167c577f40f385abe6652bf4fd3bdb04617b36029dc3000b190c18b4b3a29827da75c680979697d61fffb45b86ba6226f880b98b2f308f4f
kubernetes-src.tar.gz 0a8d8ed208bc0bf424060126c76fcd8dbbd53a9b9695647314a4097f7013f548b76850438933760ff76835867676cddddf65134ad79f977ecdb98632fc2edda3

Client Binaries

filename sha512 hash
kubernetes-client-darwin-386.tar.gz c919d030255c5d3879926d8aaa53939cd5aa37084799748452166ca6668bd1d10edf063d633682cddafaaed43dd1b991f4ad09139c5e4f519bd69f581b3fe0aa
kubernetes-client-darwin-amd64.tar.gz ec14d4a1d720890065211544b099be17315265534cfd20435194dc842cc807c20b5fae78f5b95ba7d05f3d921d522017f50861760d195ce1bf5b1acfdb2dbb29
kubernetes-client-linux-386.tar.gz 6cee12be5b855600ee80f15d1e0511099941b099bd5b252549abdc2a65c077f10ca4d53ab9804a0ce8d51f3b9cbab829cb551733cd4aed37c0d91238b82a8fe4
kubernetes-client-linux-amd64.tar.gz 27d8cd48c1f3259055965b85a6b973ecb5b8a36894f94c232d773f89539e28e6c270bfe35427c70b4ad4800e42c869851981cf88f586b7d488efa538e6c88126
kubernetes-client-linux-arm.tar.gz 7f98230569c61fdf2b141e519f042b2e27ff37aeda746dc30bb7ce226b5d6b0c0bb85c6070b9ffc8d38c2441feb5bd8736c67708a59552e86a2c30cc02ecdece
kubernetes-client-linux-arm64.tar.gz 159da67010af38d87c5318b7ad594120afd6a9b780d11d6e607e7214862cd6514b00da673cce72574771dbc780ab435dbba0a3267f051a20155c05ee0729ded0
kubernetes-client-linux-ppc64le.tar.gz d8ca7871d3d40947db69061284cb31c4d072d4da56fbb11a4485f6853f041835d9605cfc5dcea88d58c7f484cce13dbca485e80891c845291b9b28c574df310b
kubernetes-client-linux-s390x.tar.gz 1c58db90b6e09b8d8f956a00263cb20271b8403f7fb6c5b20d76cca9ed973c35d2f5c910a6d42980ec9159480682d3786a59e9e05ce356a7e3b4181c848ad122
kubernetes-client-windows-386.tar.gz 00fb87dd4899208dd6607c22828f3985ebfd5e1f97cb24e3b2c69c249a4887d5c26c603b3bb4c21f9e2b737c917ddf95a1818d9de5c9ec97d3f5faa0c3dada52
kubernetes-client-windows-amd64.tar.gz 7afdf637d62dde480162ad1521360b2bc78e0d4d20f6e6201e2f19b55b8e9bbd69c1ce8d03101c750ef389c65a1bc0a94dfc9a2d501d6840fd31eacbd3582028

Server Binaries

filename sha512 hash
kubernetes-server-linux-amd64.tar.gz 2ac3c4910cd36f02a62304d78fe144b821edf445c522028e6b57d2dc3bcc7355159a58815d5a6991b3f2c33bb0ef23e07134c8bbf93b34be7452f80c9a9e6edf
kubernetes-server-linux-arm.tar.gz db06b5f1a83ca4ec82428ab771eae2858b188dc23780fb9b146494c06aa6175421090b200c58b670e2d4253a7e0d4b07172a632e0754c35ccdee7e264c636f17
kubernetes-server-linux-arm64.tar.gz b40e1745d1ecfcc95f3a750990244fa128381d6d74246798a62aecb8cec9c77cdcd470e79334eb5c670e1e3a288080b4e26a080c64481ba608e3156c72df474b
kubernetes-server-linux-ppc64le.tar.gz c84297fd2b18b6bdade5a135a3da929e286bbba5c8dd66778091bad4eea1ac4b97a32ab3b146a88f0716bebcce9a4a85a7cd421cb185a3df864dcfa77312b3a9
kubernetes-server-linux-s390x.tar.gz 39c8f6e7f52bec155b11652b4e80c2c52acf8754dbdf80a9d5bab5370d1debf4f4783c1a6968d41822b00ff744c72947df6cbc4623578e7679b9ce9a98f64ff3

Node Binaries

filename sha512 hash
kubernetes-node-linux-amd64.tar.gz bf19ae7140836aea1b6f414532eff886e3b91e0746b9224ce46e60e0b83fa90a8c3df1ff8e01ff340a1e1874ce15da28e98224024ed3139589474e89befa19f2
kubernetes-node-linux-arm.tar.gz 4b8194340a8675107da3969845173fb34ed2b0a38745ec0ac395ebf2116ee84d55be6e22ff84fafbaf4ca60a05f6debf6e95957a2261ac8a587eea32e5803fd5
kubernetes-node-linux-arm64.tar.gz 8494dae5225f3b543afd575003fe0f30eb0f3cf9bc9dfbae72d6bda8f17c5446165433c28842a114af66ac3ae8fca9f92d780d1eb93e9bcb6b5dcc4fe8cd2a7f
kubernetes-node-linux-ppc64le.tar.gz 0bd41d95f0a76c1b057a8913a8b2cffbf6d48c47aef1d9beed0de205b8010e8071e8f527eeaa003730ef97a017083278cb2036cf22a1abfb2f4669b935823cdc
kubernetes-node-linux-s390x.tar.gz 169b16512df42a6cb5b000a3d6d6da5ae48a733c5d11b034eaec6b3816b86ec97b92e4075872900188bc296427037299841224e552ccf079097d5cf333627cbc
kubernetes-node-windows-amd64.tar.gz ccfebde5176cb28529552889250bf706add04df3c3f9aee5b8bdd9ebdb5bce334add8351cdbeebb9bd4b17b31d994b7739d2c494aa4c64bcb3ceba1a6ef53447

Changelog since v1.14.0-beta.1

Action Required

  • ACTION REQUIRED: The node.k8s.io API group and runtimeclasses.node.k8s.io resource have been migrated to a built-in API. If you were using RuntimeClasses (an default-disabled alpha feature, as of Kubernetes v1.12), then you must recreate all RuntimeClasses after upgrading, and the runtimeclasses.node.k8s.io CRD should be manually deleted. RuntimeClasses can no longer be created without a defined handler. (#74433, @tallclair)
    • Introduce a RuntimeClass v1beta1 API. This new beta API renames runtimeHandler to handler, makes it a required field, and cuts out the spec (handler is a top-level field).
  • Transition CSINodeInfo and CSIDriver alpha CRDs to in-tree CSINode and CSIDriver core storage v1beta1 APIs. (#74283, @xing-yang)
    • ACTION REQUIRED: the alpha CRDs are no longer used and drivers will need to be updated to use the beta APIs.
    • The support for _ in the CSI driver name will be dropped as the CSI Spec does not allow that.

Other notable changes

  • Support collecting pod logs under /var/log/pods/NAMESPACE_NAME_UID to stackdriver with k8s_pod resource type. (#74502, @Random-Liu)
  • --make-symlinks for hyperkube was marked hidden for a while, This flag is now deprecated and will be removed in a future release. (#74975, @dims)
  • fix smb unmount issue on Windows (#75087, @andyzhangx)
  • Kubelet no longer watches configmaps and secrets for terminated pods, in worst scenario causing it to not be able to send other requests to kube-apiserver (#74809, @oxddr)
    • Fixes a bug concerning Quobyte volumes where user mappings only worked if the hosts Kubernetes plugin mount was provided via an external configuration using the allow-usermapping-in-volumename option. (#74520, @casusbelli)
  • Change CRI pod log directory from /var/log/pods/UID to /var/log/pods/NAMESPACE_NAME_UID. (#74441, @Random-Liu)
    • It is recommended to drain the node before upgrade, or reboot the node after upgrade.
  • Promote RuntimeClass to beta, and enable by default. (#75003, @tallclair)
  • New "dry_run" metric label (indicating the value of the dryRun query parameter) into the metrics: (#74997, @jennybuckley) * apiserver_request_total * apiserver_request_duration_seconds
    • New "APPLY" value for the "verb" metric label which indicates a PATCH with "Content-Type: apply-patch+yaml". This value is experimental and will only be present if the ServerSideApply alpha feature is enabled.
  • GCE: bump COS image version to cos-beta-73-11647-64-0 (#75149, @yguo0905)
  • Alpha support for ephemeral CSI inline volumes that are embedded in pod specs. (#74086, @vladimirvivien)
  • Add support for node side CSI volume expansion (#74863, @gnufied)
    • Add mechanism for Admission Webhooks to specify which version of AdmissionReview they support (#74998, @mbohlool)
        • Add mechanism for CRD Conversion Webhooks to specify which version of ConversionReview they support
  • Add a new kubelet endpoint for serving first-class resource metrics (#73946, @dashpole)
  • Deprecate AWS, Azure, GCE and Cinder specific volume limit predicates. (#74544, @gnufied)
  • PodReadinessGate feature is now GA. The feature gate will not allow disabling it. (#74434, @freehan)
  • If CSINodeInfo and CSIMigration feature flags are active in the cluster, Kubelet will post NotReady until CSINode is initialized with basic volume plugin mechanism information for well-known drivers (#74835, @davidz627)
  • Add network stats for Windows nodes and containers (#74788, @feiskyer)
  • kubeadm: when calling "reset" on a control-plane node, remove the APIEndpoint information for this node from the ClusterStatus in the kubeadm ConfigMap. (#75082, @neolit123)
  • kube-apiserver now serves OpenAPI specs for registered CRDs with defined (#71192, @roycaihw)
    • validation schemata as an alpha feature, to be enabled via the "CustomResourcePublishOpenAPI" feature gate. Kubectl will validate client-side using those. Note that in
    • future, client-side validation in 1.14 kubectl against a 1.15 cluster will reject
    • unknown fields for CRDs with validation schema defined.
  • Fix kubelet start failure issue on Azure Stack due to InstanceMetadata setting (#74936, @rjaini)
  • add subcommand kubectl create cronjob (#71651, @Pingan2017)
  • The CSIBlockVolume feature gate is now beta, and defaults to enabled. (#74909, @bswartz)
  • Pre-existing log files are now opened with O_APPEND, instead of O_TRUNC. This helps prevent losing logs when components crash-loop, and also enables external log rotation utilities to truncate log files in-place without components extending log files to their pre-truncation sizes on subsequent writes. (#74837, @mtaufen)
  • the test/e2e/e2e.test binary can test arbitrary storage drivers, see the -storage.testdriver parameter (#72836, @pohly)
  • Fix panic in kubectl cp command (#75037, @soltysh)
  • iscsi modules haven't even been loaded /sys/class/iscsi_host directory won't exist (#74787, @jianglingxia)
  • the fluentd addon daemonset will now target all nodes. (#74424, @liggitt) * setting ENABLE_METADATA_CONCEALMENT=true in kube-up will now set a cloud.google.com/metadata-proxy-ready=true label on new nodes. In v1.16, the metadata proxy add-on will switch to using that label as a node selector. * setting KUBE_PROXY_DAEMONSET=true in kube-up will now set a node.kubernetes.io/kube-proxy-ds-ready=true label on new nodes. In v1.16, the kube-proxy daemonset add-on will switch to using that label as a node selector. * In 1.16, the masq-agent daemonset add-on will switch to using node.kubernetes.io/masq-agent-ds-ready as a node selector.
    • Kubelet: replace du and find with a golang implementation (#74675, @dashpole)
        • Kubelet: periodically update machine info to support hot-add/remove
  • kubeadm: add certificate-key and skip-certificate-key-print flags to kubeadm init (#74671, @yagonobre)
  • Admission webhooks rules can now limit scope to only match namespaced, or only cluster-scoped resources with a scope: "Cluster" | "Namespaced" | "*" field. (#74477, @liggitt)
  • The CSIPersistentVolume and KubeletPluginWatcher feature gates cannot be disabled, and will be removed in Kubernetes v1.16 (#74830, @msau42)
  • Kubelet won't evict a static pod with priority system-node-critical upon resource pressure. (#74222, @Huang-Wei)
  • Fixes panic if a kubelet is run against an older kube-apiserver (#74529, @liggitt)
  • The resource group name in Azure providerID is not converted to lower cases. (#74882, @feiskyer)
  • Remove the out-of-tree PersistentVolumeLabel controller because it cannot run without Initializers (removed in v1.14). If you are using AWS EBS, GCE PD, Azure Disk, Cinder Disk or vSphere volumes and rely on zone labels, then enable the PersistentVolumeLabel admission controller in the kube-apiserver in the --enable-admission-plugins flag. (#74615, @andrewsykim)
  • kubeadm: improved RequiredIPVSKernelModulesAvailable warning message (#74033, @bart0sh)
  • Add nullable support to CustomResourceDefinition OpenAPI validation schemata. (#74804, @sttts)
  • Fix kube-apiserver not to create default/kubernetes service endpoints before it reports readiness via the /healthz and therefore is ready to serve requests. Also early during startup old endpoints are remove which might be left over from a previously crashed kube-apiserver. (#74668, @sttts)
  • kubeadm: fix a bug where standard kubeconfig paths were searched even if the user provided /etc/kubernetes/admin.conf explicitly for commands that accept --kubeconfig, like kubeadm token. (#71874, @neolit123)
    • kubeadm: use the default kubeconfig (/etc/kubernetes/admin.conf) for "kubeadm reset" and "kubeadm upgrade" commands.
  • Increase api server client certificate expiration histogram resolution to accommodate short-lived (< 6h) client certificates. (#74806, @mxinden)
  • Default RBAC policy no longer grants access to discovery and permission-checking APIs (used by kubectl auth can-i) to unauthenticated users. Upgraded clusters preserve prior behavior, but cluster administrators wishing to grant unauthenticated users access in new clusters will need to explicitly opt-in to expose the discovery and/or permission-checking APIs: (#73807, @dekkagaijin) * kubectl create clusterrolebinding anonymous-discovery --clusterrole=system:discovery --group=system:unauthenticated * kubectl create clusterrolebinding anonymous-access-review --clusterrole=system:basic-user --group=system:unauthenticated
  • The PersistentLocalVolumes feature is GA. The feature gate cannot be disabled and will be removed in Kubernetes 1.17 (#74769, @msau42)
  • kubelet: resolved hang/timeout issues when running large numbers of pods with unique configmap/secret references by reverting to 1.11 configmap/secret lookup behavior (#74755, @liggitt)
  • Convert latency/latencies in metrics name to duration. (#74418, @danielqsj)
    • The following metrics are changed and mark previous metrics as deprecated:
      • rest_client_request_latency_seconds -> rest_client_request_duration_seconds
      • apiserver_proxy_tunnel_sync_latency_secs -> apiserver_proxy_tunnel_sync_duration_seconds
      • scheduler_scheduling_latency_seconds -> scheduler_scheduling_duration_seconds
  • Fix help message for --container-runtime-endpoint: only unix socket is support on Linux. (#74712, @feiskyer)
  • Update to use golang 1.12 (#74632, @cblecker)
  • The RunAsGroup feature has been promoted to beta and enabled by default. PodSpec and PodSecurityPolicy objects can be used to control the primary GID of containers on supported container runtimes. (#73007, @krmayankk)
  • fix Azure Container Registry anonymous repo image pull error (#74715, @andyzhangx)
  • Adds the same information to an init container as a standard container in a pod when using PodPresets. (#71479, @soggiest)
  • fix the flake in scheduling_queue_test.go (#74611, @denkensk)
  • The kube-apiserver OpenAPI definitions with the prefix "io.k8s.kubernetes.pkg" (deprecated since 1.9) have been removed. (#74596, @sttts)
  • kube-conformance image will now run ginkgo with the --dryRun flag if the container is run with the environment variable E2E_DRYRUN set. (#74731, @johnSchnake)
  • The deprecated MountPropagation feature gate has been removed, and the feature is now unconditionally enabled. (#74720, @bertinatto)
  • Introduce dynamic volume provisioning shim for CSI migration (#73653, @ddebroy)
  • Fix --help flag parsing (#74682, @soltysh)
  • This PR removes the following metrics: (#74636, @logicalhan)
    • reflector_items_per_list
    • reflector_items_per_watch
    • reflector_last_resource_version
    • reflector_list_duration_seconds
    • reflector_lists_total
    • reflector_short_watches_total
    • reflector_watch_duration_seconds
    • reflector_watches_total
    • While this is a backwards-incompatible change, it would have been impossible to setup reliable monitoring around these metrics since the labels were not stable.
  • Add a configuration field to shorten the timeout of validating/mutating admission webhook call. The timeout value must be between 1 and 30 seconds. Default to 30 seconds when unspecified. (#74562, @roycaihw)
  • client-go: PortForwarder.GetPorts() now contain correct local port if no local port was initially specified when setting up the port forwarder (#73676, @martin-helmich)
  • Apply resources from a directory containing kustomization.yaml (#74140, @Liujingfang1)

    •     kubectl apply -k dir
      
    •     # Delete resources from a directory containing kustomization.yaml.
      
    •     kubectl delete -k dir
      
    •     # List resources from a directory containing kustomization.yaml
      
    •     kubectl get -k dir
      
  • kubeadm: Allow to download certificate secrets uploaded by init or upload-certs phase, allowing to transfer certificate secrets (certificates and keys) from the cluster to other master machines when creating HA deployments. (#74168, @ereslibre)
  • Fixes an issue with missing apiVersion/kind in object data sent to admission webhooks (#74448, @liggitt)
  • client-go: the deprecated versionless API group accessors (like clientset.Apps() have been removed). Use an explicit version instead (like clientset.AppsV1()) (#74422, @liggitt)
  • The --quiet option to kubectl run now suppresses resource deletion messages emitted when the --rm option is specified. (#73266, @awh)
  • Add Custom Resource support to "kubectl autoscale" (#72678, @rmohr)
  • Image garbage collection no longer fails for images with only one tag but more than one repository associated. (#70647, @corvus-ch)
    • Fix liveness probe in fluentd-gcp cluster addon (#74522, @Pluies)
  • The new test [sig-network] DNS should provide /etc/hosts entries for the cluster [LinuxOnly] [Conformance] will validate the host entries set in the /etc/hosts file (pod's FQDN and hostname), which should be managed by Kubelet. (#72729, @bclau)
    • The test has the tag [LinuxOnly] because individual files cannot be mounted in Windows Containers, which means that it cannot pass using Windows nodes.

v1.14.0-beta.1

Documentation

Downloads for v1.14.0-beta.1

filename sha512 hash
kubernetes.tar.gz 065014c751635f8c077fbcc105df578594baf8afd8b8339004909198e1bd68d0a7ce3644ed5d54e5964d1306aa650f22a5ce83063415240f4dffd6706c1cc33b
kubernetes-src.tar.gz 244c19d9fad21ae154ee78fc94888dc60bcfcf3ec72bdc28a82e77c572cbc969d2abbf20397ef9564a35585c08dfe179b105fc25efac973e0a13d78ee2ff8f42

Client Binaries

filename sha512 hash
kubernetes-client-darwin-386.tar.gz 78a61a1e922daa39a9f7dd61b8bad87d202b537bda59f90ae8aae941c0ff412e3d328530af9dd9f22462cbd67254e7ce1556defe48bb10bd6a94d4302464fe8e
kubernetes-client-darwin-amd64.tar.gz 5a1d66fd90a5dc07b95b7617b5583595e0a4a664e7005f6281f846c85b21e28692b2e2d55b7c40c7b8cbb96b6b8cd6c832f340c7cc67579641beabac033014ad
kubernetes-client-linux-386.tar.gz 30991ce0776bae7551b98a811e3ccb5104b0859805c41a216db7d5779cddfb36ef3c5e658ea2adeaf67f8e6f181768850b09a0e8320c2983d34664156ea638ac
kubernetes-client-linux-amd64.tar.gz 5501a9c3a95e43f0c691b05043546f2c497d50a6ad88b88219842d61be83d7dcb8871ff9fd2447c02bc842c4c962f298a3ce2e0618fae70e8aa391c9407626e5
kubernetes-client-linux-arm.tar.gz 4bf341a2f943719d006f4cacc26fdf4d021560d37d49c8d9c4620d294142041155a88dc721d9373a8617e1baa904c02b4545f379ffe87c6ab20e5459a5d3c2de
kubernetes-client-linux-arm64.tar.gz 1c0c660618947b053404ab8ef40bdec0a06d54a1f9edc585a7259806f878327d9ae54100bba98e5b7f44f5db4303276d189792e68517603520a49868c07e684c
kubernetes-client-linux-ppc64le.tar.gz 9d2c3f4bfbad03b41859670f9cfda6596f51a89077fd4da2f74490f71b5de10e459954e897d2a1ba3a217c62caaf1be74424e5bf6a5609868dc4f069ac06c94f
kubernetes-client-linux-s390x.tar.gz 9be5259caf39ff3c4d0f024d616bff50ac417d42a87c56b6877fbdf5aaadecee05631ac8ddd6dca20d52ae4ca7e1227a1fd5e882be9821f3711e144b84d41099
kubernetes-client-windows-386.tar.gz 5b1e75f532d9a4d656cd1c5ec48a19d01c4ca731c6c3d986ebbc48a9f1e1d61d6c5603145808fb929117cf2202bc75f31ec7661a50edf24c5af6b3419ebcc0ab
kubernetes-client-windows-amd64.tar.gz a501d3c0be55d5a73214a83d1f48b39a49d4a84f5f9988fa34cf66ef6bc78d3f3e06c40dc3d59538c3bff07aa2128ee814d9825c31f8b9abbf045e1ebb581bed

Server Binaries

filename sha512 hash
kubernetes-server-linux-amd64.tar.gz aed44ec5bc6bdaf41c20824a9841ee541bfd23362966f9326eca2cafbd03eca69325877784b8c9b058963dc5b8ba656b9da446513fd5eeed8133a783769acd16
kubernetes-server-linux-arm.tar.gz e751cb675013183a70a8817dca0b5c456a1ccf075244b411317e813b5164ae710460a53e81191ded9d2ccaf284ae00304bfbe1b3d219b2a8d57761d733293409
kubernetes-server-linux-arm64.tar.gz 3aabb1a9bd07413d0740adfb638b0e5ca4cd4a58eda244c5ebc1ea01780e0b2863806c35792a0590069ef0cccc2665198afed1984d1c49f0726b75d4216609a4
kubernetes-server-linux-ppc64le.tar.gz 3c57e2b47b3b9ddc1039684840886877947141b1e4d31f909793678544fe92e10aa82a207936f0bcb3c657044c7b875f34f41f086ccb7f97a154499d01266f73
kubernetes-server-linux-s390x.tar.gz fe418cfafc63cbccc1898086296e52026df27ff498753089792175ce0d41d889ca50a4eb5104a84b78a0a25d524dcdf5ec5eec8aa213d58178ae38411cfb58af

Node Binaries

filename sha512 hash
kubernetes-node-linux-amd64.tar.gz 83b63d7e9d18fe35564105fd70629af9ba8f20112933b3ace92a48887702862e013ac3b3e144cfd44c8ead7c766df584766749af9d33a9aaa3808e370d3ea359
kubernetes-node-linux-arm.tar.gz e4ac000be25ced9b308ec698da9702bd10a0a7183bfea9648500ab9d0879989c54328cb4fb51be545831d18676067d6b53e55cc49b14f55f35032a66dfa28806
kubernetes-node-linux-arm64.tar.gz 35fa5ec8a7fc001fe33abd89f8a446ce0ea2a011db27dc8ff544c2b199b065b19372afe95e3616f538347243d7599f29fb5cea1a46a3fcfdfe4a0f2a346683b2
kubernetes-node-linux-ppc64le.tar.gz a5b938e9cdc39fc292269af4c3961a17b9bcdaf85b3c58db680f2d1a4fb088eb648efc268dfe3325aed96f6a7cdcff070dc7a3200cc169010ff2b402fae1a26a
kubernetes-node-linux-s390x.tar.gz d8e2b5a945f881ddb0a25576a614d564ccef0ad4e93c84b30cdc57888e81e04932d798415a1a50cdfc6d2f857e1d027e2034ba9c40b5d8ed0009cfbdc8915e0f
kubernetes-node-windows-amd64.tar.gz 426774466800ca11cae16821c7d92917b2ce138fcb3f6dca813ec0f060649d6a50187982aaac31fbb081604f5bd2c317616b6f211beb139f53192269194f012e

Changelog since v1.14.0-alpha.3

Action Required

  • Added cadvisor metric labels pod and container where pod_name and container_name are present to match instrumentation guidelines. (#69099, @ehashman)
    • Action required: any Prometheus queries that match pod_name and container_name labels (e.g. cadvisor or kubelet probe metrics) should be updated to use pod and container instead. pod_name and container_name labels will be present alongside pod and container labels for one transitional release and removed in the future.

Other notable changes

  • fix parse devicePath issue on Azure Disk (#74499, @andyzhangx)
  • fix issue: fail to detach azure disk when there is server side error (#74398, @andyzhangx)
  • Allow Cinder volume limit to be configured from node too (#74542, @gnufied)
  • Fix subnet annotation checking for Azure internal loadbalancer (#74498, @feiskyer)
  • Allow the kubelet to pass Windows GMSA credentials down to Docker (#73726, @wk8)
  • PriorityClass in scheduling.k8s.io/v1beta1 and scheduling.k8s.io/v1alpha1 are deprecated by PriorityClass scheduling.k8s.io/v1 and will not be served starting in v1.17. (#74465, @bsalamat)
  • kubeadm: fixed nil pointer dereference caused by a bug in url parsing (#74454, @bart0sh)
  • Fix the unexpected NotReady status when Node's iops is full if the runtime is dockershim. (#74389, @answer1991)
  • Split up the mondo kubernetes-test tarball into kubernetes-test-portable and kubernetes-test-{OS}-{ARCH} tarballs. (#74065, @ixdy)
  • Move fluentd-elasticsearch addon images to community controlled location (#73819, @coffeepac)
  • The PriorityClass API has been promoted to scheduling.k8s.io/v1 with no changes. The scheduling.k8s.io/v1beta1 version is now deprecated and will stop being served by default in v1.17. (#73555, @bsalamat)
  • fix get azure accounts timeout issue when there is no out-bound IP (#74191, @andyzhangx)
  • fix mixed protocol issue for azure load balancer (#74200, @andyzhangx)
  • Don't update the Pod object after each scheduling attempt by adding a timestamp to the scheduling queue. (#73700, @denkensk)
  • kubeadm: remove local etcd members from the etcd cluster when kubeadm reset (#74112, @pytimer)
  • Fix keymutex issues which may crash in some platforms. (#74348, @danielqsj)
  • Fixed scanning of failed iSCSI targets. (#74306, @jsafrane)
  • kubeadm: Do not fail preflight checks when running on >= 5.0 Linux kernel (#74355, @brb)
  • Reduces the cache TTL for negative responses to 5s minimum. (#74093, @blakebarnett)
  • The Ingress API is now available via networking.k8s.io/v1beta1. extensions/v1beta1 Ingress objects are deprecated and will no longer be served in v1.18. (#74057, @liggitt)
  • kubelet's --containerized flag will no longer be supported and will be removed in a future release (#74267, @dims)
  • Optimize scheduler cache snapshot algorithm to improve scheduling throughput. (#74041, @bsalamat)
  • Extends the VolumeSubpathEnvExpansion alpha feature to support environment variable expansion (#71351, @kevtaylor)
    • Implements subPathExpr field for expanding environment variables into a subPath
    • The fields subPathExpr and subPath are mutually exclusive
    • Note: This is a breaking change from the previous version of this alpha feature
  • Added kube-proxy support for overlay networking and DSR in Windows and new flags for network-name, source-vip, and enable-dsr. (#70896, @ksubrmnn)
  • StorageOS volume plugin updated to fix an issue where volume mount succeeds even if request to mount via StorageOS API fails. (#69782, @darkowlzz)
  • kubeadm: Allow to upload certificates required to join a new control-plane to kubeadm-certs secret using the flag --experimental-upload-certs on init or upload-certs phase. (#73907, @yagonobre)
  • export query parameter is deprecated and will be removed in a future release (#73783, @deads2k)
  • e2e storage tests run faster and are easier to read (#72434, @pohly)
  • kubectl: fix a bug where "describe" cannot obtain the event messages for a static pod (#74156, @gaorong)
  • windows: Ensure graceful termination when being run as windows service (#73292, @steffengy)
  • CoreDNS adds readinessProbe which prevents loadbalancing to unready pods, and also allows rolling updates to work as expected. (#74137, @rajansandeep)
  • Fixes use of webhook admission plugins with multi-version custom resources (#74154, @mbohlool)
  • kubeadm no longer allows using v1alpha3 configs for anything else than converting them to v1beta1. (#74025, @rosti)
  • Change kubelet metrics to conform metrics guidelines. (#72470, @danielqsj)
    • The following metrics are deprecated, and will be removed in a future release:
      • kubelet_pod_worker_latency_microseconds
      • kubelet_pod_start_latency_microseconds
      • kubelet_cgroup_manager_latency_microseconds
      • kubelet_pod_worker_start_latency_microseconds
      • kubelet_pleg_relist_latency_microseconds
      • kubelet_pleg_relist_interval_microseconds
      • kubelet_eviction_stats_age_microseconds
      • kubelet_runtime_operations
      • kubelet_runtime_operations_latency_microseconds
      • kubelet_runtime_operations_errors
      • kubelet_device_plugin_registration_count
      • kubelet_device_plugin_alloc_latency_microseconds
    • Please convert to the following metrics:
      • kubelet_pod_worker_duration_seconds
      • kubelet_pod_start_duration_seconds
      • kubelet_cgroup_manager_duration_seconds
      • kubelet_pod_worker_start_duration_seconds
      • kubelet_pleg_relist_duration_seconds
      • kubelet_pleg_relist_interval_seconds
      • kubelet_eviction_stats_age_seconds
      • kubelet_runtime_operations_total
      • kubelet_runtime_operations_duration_seconds
      • kubelet_runtime_operations_errors_total
      • kubelet_device_plugin_registration_total
      • kubelet_device_plugin_alloc_duration_seconds
  • This change ensures that volumes get provisioned based on the zone information provided in allowedTopologies. (#72731, @skarthiksrinivas)
    • Storage class spec:
    • kind: StorageClass
    • apiVersion: storage.k8s.io/v1
    • metadata:
    • name: fastpolicy1
    • provisioner: kubernetes.io/vsphere-volume
    • parameters:
    • diskformat: zeroedthick
      
    • storagePolicyName: vSAN Default Storage Policy
      
    • allowedTopologies:
      • matchLabelExpressions:
      • key: failure-domain.beta.kubernetes.io/zone
    • values:
      
    • - zone1
      
    • PV creation Logs:
    • I0109 11:17:52.321372 1 vsphere.go:1147] Starting to create a vSphere volume with volumeOptions: &{CapacityKB:1048576 Tags:map[kubernetes.io/created-for/pvc/namespace:default kubernetes.io/created-for/pvc/name:pvcsc-1-policy kubernetes.io/created-for/pv/name:pvc-34650c12-1400-11e9-aef4-005056804cc9] Name:kubernetes-dynamic-pvc-34650c12-1400-11e9-aef4-005056804cc9 DiskFormat:zeroedthick Datastore: VSANStorageProfileData: StoragePolicyName:vSAN Default Storage Policy StoragePolicyID: SCSIControllerType: Zone:[zone1]}
    • ...
    • I0109 11:17:59.430113 1 vsphere.go:1334] The canonical volume path for the newly created vSphere volume is "[vsanDatastore] 98db185c-6683-d8c7-bc55-0200435ec5da/kubernetes-dynamic-pvc-34650c12-1400-11e9-aef4-005056804cc9.vmdk"
    • Ran regression tests (no zone) and they passed.
  • vSphere cloud provider correctly retrieves the VM's UUID when running on Windows (#71147, @benmoss)
  • Re-issue Allocate grpc calls before starting a container that requests device-plugin resources if the cached state is missing. (#73824, @jiayingz)
  • [CRI] Add a new field called runtime_handler into PodSandbox and PodSandboxStatus to track the RuntimeClass information of a pod. (#73833, @haiyanmeng)
  • kubelet: OS and Arch information is now recorded in kubernetes.io/os and kubernetes.io/arch labels on Node objects. The previous labels (beta.kubernetes.io/os and beta.kubernetes.io/arch) are still recorded, but are deprecated and targeted for removal in 1.18. (#73333, @yujuhong)
  • This change applies zone labels to vSphere Volumes automatically. The zone labels are visible on the PV: (#72687, @subramanian-neelakantan)
    • $ kubectl get pv --show-labels
    • NAME CAPACITY ACCESSMODES STATUS CLAIM REASON AGE LABELS
    • pv-abc 5Gi RWO Bound default/claim1 46s failure-domain.beta.kubernetes.io/region=VC1,failure-domain.beta.kubernetes.io/zone=cluster-1
  • fix smb remount issue on Windows (#73661, @andyzhangx)
  • Kubelet now tries to stop containers in unknown state once before restart or remove. (#73802, @Random-Liu)
  • Deprecate --export flag from kubectl get command. (#73787, @soltysh)
  • Breaking changes in client-go: (#72214, @caesarxuchao)
    • The disk-cached discovery client is moved from k8s.io/client-go/discovery to k8s.io/client-go/discovery/cached/disk.
    • The memory-cached discovery client is moved from k8s.io/client-go/discovery/cached to k8s.io/client-go/discovery/cached/memory.
  • kubelet now accepts pid=<number> in the --system-reserved and --kube-reserved options to ensure that the specified number of process IDs will be reserved for the system as a whole and for Kubernetes system daemons respectively. Please reference Kube Reserved and System Reserved in Reserve Compute Resources for System Daemons in the Kubernetes documentation for general discussion of resource reservation. To utilize this functionality, you must set the feature gate SupportNodePidsLimit=true (#73651, @RobertKrawitz)
  • The apiserver, including both the kube-apiserver and apiservers built with the generic apiserver library, will now return 413 RequestEntityTooLarge error if a json patch contains more than 10,000 operations. (#74000, @caesarxuchao)
  • kubeadm: allow the usage of --kubeconfig-dir and --config flags on kubeadm init (#73998, @yagonobre)
  • when pleg channel is full, discard events and record its count (#72709, @changyaowei)
  • Is ->It in line 6 (#73898, @xiezongzhe)
  • Events reported for container creation, start, and stop now report the container name in the message and are more consistently formatted. (#73892, @smarterclayton)
  • kubectl auth reconcile now outputs details about what changes are being made (#71564, @liggitt)
  • kubeadm: fix a bug in the underlying library for diff related to characters like '%' (#73941, @neolit123)
  • kube-apiserver: a request body of a CREATE/UPDATE/PATCH/DELETE resource operation larger than 100 MB will return a 413 "request entity too large" error. (#73805, @caesarxuchao)
    • Custom apiservers built with the latest apiserver library will have the 100MB limit on the body of resource requests as well. The limit can be altered via ServerRunOptions.MaxRequestBodyBytes.
    • The body size limit does not apply to subresources like pods/proxy that proxy request content to another server.
  • Kustomize is developed in its own repo https://github.com/kubernetes-sigs/kustomize (#73033, @Liujingfang1)
    • This PR added a new subcommand kustomize in kubectl.
    • kubectl kustomize has the same effect as kustomize build
    • To build API resources from somedir with a kustomization.yaml file
    • kubectl kustomize
    • This command can be piped to apply or delete
    • kubectl kustomize | kubectl apply -f -
    • kubectl kustomize | kubectl delete -f -
  • kubeadm: all master components are now exclusively relying on the PriorityClassName pod spec for annotating them as cluster critical components. Since scheduler.alpha.kubernetes.io/critical-pod annotation is no longer supported by Kubernetes 1.14 this annotation is no longer added to master components. (#73857, @ereslibre)
  • Speedup kubectl by >10 when calling out to kube-apiserver for discovery information. (#73345, @sttts)
  • kubeadm no longer dumps backtrace if it fails to remove the running containers on reset. (#73951, @rosti)

v1.14.0-alpha.3

Documentation

Downloads for v1.14.0-alpha.3

filename sha512 hash
kubernetes.tar.gz 5060dcf689dad4e19da5029eb8fc3060a4b2bad988fddff438d0703a45c02481bcfbc15f45d2855f4fd5e9eb43847400ebb25dce19e24f0e0e194a7f57176ce5
kubernetes-src.tar.gz 754c948b5d25b01f211866d473257be5fb576b4b97703eb6fc08679d6525e1f53195a450f3f47b77fabb92bf058583b66230959197b5bcf72528e54ccb349c07

Client Binaries

filename sha512 hash
kubernetes-client-darwin-386.tar.gz 5bd74dfc86bacf89d6b05d541e13bf390216039a42cc90fef2b248820acd84f56a445ec66d52497ff77e1af47455f285c993cd1d44cc3050996189bd328ea2be
kubernetes-client-darwin-amd64.tar.gz 34e16661d66d337083583dfb478756ec8cc664d7cfc2dd1817bf1da03cdc380668be9df9f178b5fd5ccab5014e6686f83b9fee6192fbf77d2298d397e872a893
kubernetes-client-linux-386.tar.gz 15f99e85bcc95f7b8e1b4c6ecc23de36e89a54108003db926e97ec2e7253f363f6ed85e39a47305dbccf596f72e88edd7bcda6d528919da9c0b81541f58506d4
kubernetes-client-linux-amd64.tar.gz 2e61cf9b776150c4f1830d068ffee9701cb04979152ed6b62fc1bf53163e6194029a4f75536e7fda71c3dfce1de285f425bde342a4efdd1f7bf973f105750ac4
kubernetes-client-linux-arm.tar.gz 67fb3805bb1b4a77f6603fbde9bd1d26e179de1a594c85618aa7b17be6abc510a9a0cd499ef4fe974574cf73b364da641121f21864c8472d713eec76e4c52bca
kubernetes-client-linux-arm64.tar.gz 28930dc384b51051081a52874bc4d6dafa3c992dfa214b977ef711de2c2bc3f90bdaa6243bded1e750997fec04b8ffb910db21c266e47e09426c4dbaf916a64d
kubernetes-client-linux-ppc64le.tar.gz f59eda797a57961d52fe67ba8b25a3a10267f9ce46029ed2140ef4b02615ba9944bd83d7a6e7874c7268a09a3422858b9b0c31f861941ef8be126c594fc3a7cc
kubernetes-client-linux-s390x.tar.gz c56bfb64e55cf95251157a8229a3e94310b2c46bb1c1250050893873e3112578978c1f8e29fa56fac63e2aa8a6382523ac34baf6dd523fe0919f8d702521a564
kubernetes-client-windows-386.tar.gz e49a00fbe600892dc5eed0bc21bac64806da65280c818ca79b5e8adbed7fd5ecebb6b647cb9b89ac862257995145b2397996122eefb3c8d127d857c89c29c9ae
kubernetes-client-windows-amd64.tar.gz 797e20969ed4935adcbc80ccbcd72ec5aa697e70b0d071eceefc6dbacea69aff9f6660e7eefad6661ace0afb66067c4ffaa4f6bc82e8b081b57811ab0abde218

Server Binaries

filename sha512 hash
kubernetes-server-linux-amd64.tar.gz eddfc9afd7337475c3865443170d1425dcf4a87d981555871a69bcf132e73d99b1ffa08a00490b30c60232f47bbeca4ad6253cf7e1dad44797b4af044dbdbef4
kubernetes-server-linux-arm.tar.gz dc85cd3a039cc0516beb19018c8378f3b7b88fa2edb8fa1476305e89eb7c64fef2d938bd48fd257ea8e690f7d84a69e9784a42aabed35e83ea7362c60773ba67
kubernetes-server-linux-arm64.tar.gz d7c3a72abaa4c3e3243f8b4b3a8adb8be2758e0f883423ea62d2c61b2081464a8976ad43ea0640a7e453aa4d389e3ea2d6d1baedf3b50e1171eca6e49cd087fe
kubernetes-server-linux-ppc64le.tar.gz b268a94eb056eea8bdf4d5739dec430f75a6a6b3c18e30df68d970c3566b3e4a638b3577f6219596ae54eac740628a7ebfecb0772645e6d960f790235e1d62c7
kubernetes-server-linux-s390x.tar.gz f4cfd8d2faacdd1f0065f9e0f4f8d0db7bd8f438f812f70a07f4cb5272ae9bed3ec876b3cbaf2f2a71e65e4de725e1dc0829b43f60f43c9e43656ac928657d5e

Node Binaries

filename sha512 hash
kubernetes-node-linux-amd64.tar.gz 7040ee3c032ec4fe14530c3e47ee53d731acb947b06e2d560cbcd0e7e513142c0f300302059aaef03e24311946a9c59b576948eec9b520e2367f28fc4f80226c
kubernetes-node-linux-arm.tar.gz 3d32e5243d1c65bce573cfb0f60d643ef3fc684a15551dbc8c3d5435e6854ff104c46c77b0b8708d9c661d52f7865a197ea758f0c17e1ed991993674929ea75e
kubernetes-node-linux-arm64.tar.gz d3a17027fa1c057528422b35e32260f5b7c7246400df595f0ebda5d150456d4388129b1ead4229f98f2b461ff9e85382a7da0d682541844a3c06f0aebe0469b6
kubernetes-node-linux-ppc64le.tar.gz 89ed1f5093b49ab9d58d7a70089e881bf388f3316cb2607fa18e3bf072aff3d27aabe99124334774e63decb67349eb82f33ea509b56a72a51e1443c3352b4558
kubernetes-node-linux-s390x.tar.gz 755a60824a9b8c4090a791d332e410692708ecece90e37388f58eb2c7ddddea6b859fefcc5a53ec3d275fee0a355086f4446ae8e85482a668d248cca9f5e503c
kubernetes-node-windows-amd64.tar.gz c71d8055d89e535771f345e0f673da021915a7a82c75951855ba2574a4250c8a57d0636b4ec9bba209edde8edef30098c6dec2f80403cd46139bb88d814c3751

Changelog since v1.14.0-alpha.2

Action Required

  • The --storage-versions flag of kube-apiserver is removed. The storage versions will always be the default value built-in the kube-apiserver binary. (#67678, @caesarxuchao)

Other notable changes

  • fix #73264 cpuPeriod was not reset, but used as set via flag, although it was disabled via alpha gate (#73342, @szuecs)

  • Update kubelet CLI summary documentation and generated Webpage (#73256, @deitch)

  • Considerably reduced the CPU load in kube-apiserver while aggregating OpenAPI specifications from aggregated API servers. (#71223, @sttts)

  • kubeadm: add a preflight check that throws a warning if the cgroup driver for Docker on Linux is not "systemd" as per the k8s.io CRI installation guide. (#73837, @neolit123)

  • Kubelet: add usageNanoCores from CRI stats provider (#73659, @feiskyer)

  • Fix watch to not send the same set of events multiple times causing watcher to go back in time (#73845, @wojtek-t)

  • system:kube-controller-manager and system:kube-scheduler users are now permitted to perform delegated authentication/authorization checks by default RBAC policy (#72491, @liggitt)

  • Prevent AWS Network Load Balancer security groups ingress rules to be deleted by ensuring target groups are tagged. (#73594, @masterzen)

  • Set a low oom_score_adj for containers in pods with system-critical priorities (#73758, @sjenning)

  • Ensure directories on volumes are group-executable when using fsGroup (#73533, @mxey)

  • kube-apiserver now only aggregates openapi schemas from /openapi/v2 endpoints of aggregated API servers. The fallback to aggregate from /swagger.json has been removed. Ensure aggregated API servers provide schema information via /openapi/v2 (available since v1.10). (#73441, @roycaihw)

  • Change docker metrics to conform metrics guidelines and using histogram for better aggregation. (#72323, @danielqsj)

    • The following metrics are deprecated, and will be removed in a future release:
      • docker_operations
      • docker_operations_latency_microseconds
      • docker_operations_errors
      • docker_operations_timeout
      • network_plugin_operations_latency_microseconds
    • Please convert to the following metrics:
      • docker_operations_total
      • docker_operations_latency_seconds
      • docker_operations_errors_total
      • docker_operations_timeout_total
      • network_plugin_operations_latency_seconds
  • kubectl delete --all-namespaces is a recognized flag. (#73716, @deads2k)

  • MAC Address filter has been fixed in vSphere Cloud Provider, it no longer ignores 00:1c:14 and 00:05:69 prefixes (#73721, @frapposelli)

  • Add kubelet_node_name metrics. (#72910, @danielqsj)

  • The HugePages feature gate has graduated to GA, and can no longer be disabled. The feature gate will be removed in v1.16 (#72785, @derekwaynecarr)

  • Fix a bug that aggregated openapi spec may override swagger securityDefinitions and swagger info in kube-apiserver (#73484, @roycaihw)

  • Fixes a bug that prevented deletion of dynamically provisioned volumes in Quobyte backends. (#68925, @casusbelli)

  • error messages returned in authentication webhook status responses are now correctly included in the apiserver log (#73595, @liggitt)

  • kubeadm: kubeadm alpha preflight and kubeadm alpha preflight node are removed; you can now use kubeadm join phase preflight (#73718, @fabriziopandini)

  • kube-apiserver: the deprecated repair-malformed-updates has been removed (#73663, @danielqsj)

  • e2e.test now rejects unknown --provider values instead of merely warning about them. An empty provider name is not accepted anymore and was replaced by "skeleton" (= a provider with no special behavior). (#73402, @pohly)

  • Updated AWS SDK to v1.16.26 for ECR PrivateLink support (#73435, @micahhausler)

  • Expand kubectl wait to work with more types of selectors. (#71746, @rctl)

  • The CustomPodDNS feature gate has graduated to GA, and can no longer be disabled. The feature gate will be removed in v1.16 (#72832, @MrHohn)

  • The rules field in RBAC Role and ClusterRole objects is now correctly reported as optional in the openapi schema. (#73250, @liggitt)

  • AWS ELB health checks will now use HTTPS/SSL protocol for HTTPS/SSL backends. (#70309, @2rs2ts)

  • kubeadm reset: fixed crash caused by absence of a configuration file (#73636, @bart0sh)

  • CoreDNS is now version 1.3.1 (#73610, @rajansandeep)

      • A new k8s_external plugin that allows external zones to point to Kubernetes in-cluster services.
      • CoreDNS now checks if a zone transfer is allowed. Also allow a TTL of 0 to avoid caching in the cache plugin.
      • TTL is also applied to negative responses (NXDOMAIN, etc).
  • Missing directories listed in a user's PATH are no longer considered errors and are instead logged by the "kubectl plugin list" command when listing available plugins. (#73542, @juanvallejo)

  • remove kubelet flag '--experimental-fail-swap-on' (deprecated in v1.8) (#69552, @Pingan2017)

  • Introduced support for Windows nodes into the cluster bringup scripts for GCE. (#73442, @pjh)

  • Now users could get object info like: (#73063, @WanLinghao)

    • a. kubectl get pod test-pod -o custom-columns=CONTAINER:.spec.containers[0:3].name
    • b. kubectl get pod test-pod -o custom-columns=CONTAINER:.spec.containers[-2:].name
  • scheduler: use incremental scheduling cycle in PriorityQueue to put all in-flight unschedulable pods back to active queue if we received move request (#73309, @cofyc)

  • fixes an error processing watch events when running skewed apiservers (#73482, @liggitt)

  • Prometheus metrics for crd_autoregister, crd_finalizer and crd_naming_condition_controller are exported. (#71767, @roycaihw)

  • Adds deleting pods created by DaemonSet assigned to not existing nodes. (#73401, @krzysztof-jastrzebski)

  • Graduate Pod Priority and Preemption to GA. (#73498, @bsalamat)

  • Adds configuration for AWS endpoint fine control: (#72245, @ampsingram)

    • OverrideEndpoints bool Set to true to allow custom endpoints
    • ServiceDelimiter string Delimiter to use to separate overridden services (multiple services) Defaults to "&"
    • ServicenameDelimiter string Delimiter to use to separate servicename from its configuration parameters Defaults "|"
    • OverrideSeparator string Delimiter to use to separate region of occurrence, url and signing region for each override Defaults to ","
    • ServiceOverrides string example: s3|region1, https://s3.foo.bar, some signing_region & ec2|region2, https://ec2.foo.bar, signing_region
  • The CoreDNS configuration now has the forward plugin for proxy in the default configuration instead of the proxy plugin. (#73267, @rajansandeep)

  • Fixed a bug that caused PV allocation on non-English vSphere installations to fail (#73115, @alvaroaleman)

v1.14.0-alpha.2

Documentation

Downloads for v1.14.0-alpha.2

filename sha512 hash
kubernetes.tar.gz 1330e4421b61f6b1e6e4dee276d4742754bd3dd4493508d67ebb4445065277c619c4da8b4835febf0b2cdcf9e75fce96de1c1d99998904bae2bb794a453693f2
kubernetes-src.tar.gz 352c043bebf13a616441c920f3eec80d3f02f111d8488c31aa903e1483bce6d1fbe7472208f64730142960c8f778ab921ef7b654540a3ec09e53bd7e644521bd

Client Binaries

filename sha512 hash
kubernetes-client-darwin-386.tar.gz ee5aba4efce323167e6d897a2ff6962a240e466333bcae9390be2c8521c6da50ac2cb6139510b693aad49d6393b97a2118ed1fe4f999dd08bdca6d875d25f804
kubernetes-client-darwin-amd64.tar.gz 4b5c0b340322956a8d096c595124a765ac318d0eb460d6320218f2470e22d88221a0a9f1f93d5f3075f1c36b18c7041ee2fcb32e0f9c94d9f79bc3fd3005e68e
kubernetes-client-linux-386.tar.gz 7a5bfe68dd58c8478746a410872b615daf8abb9a78754140fb4d014a0c9177a87859ac046f56f5743fb97a9881abc2cf48c3e51aa02c8a86a754bf2cc59edb54
kubernetes-client-linux-amd64.tar.gz c3139f58070241f2da815f701af3c0bd0ea4fdec1fe54bb859bd11237ac9b75ecb01b62ac1c7a459a4dd79696412c6d2f8cbd492fd062a790ceadd3dcc9b07fd
kubernetes-client-linux-arm.tar.gz 9d96d2e1e11aa61e2c3a5f4f27c18866feae9833b6ee70b15f5cdb5f992849dc1f79821af856b467487092a21a447231fb9c4de6ee6f17defed3cfa16d35b4c6
kubernetes-client-linux-arm64.tar.gz 7b4dd825cf9f217c18b28976a3faa94f0bd4868e541e5be7d57cd770e2b163c6daddf12e5f9ad51d92abde794a444f2a20bf582a30f03c39e60186d356030a2d
kubernetes-client-linux-ppc64le.tar.gz 490638e250c24b6bad8b67358fd7890f7a2f6456ae8ffe537c28bb5b3ce7abc591e6fecbddd6744f0f6c0e24b9f44c31f7ca1f7ebfc3c0d17a96fe8cf27b8548
kubernetes-client-linux-s390x.tar.gz 9dd8c3361eda15dd1594066c55b79cb9a34578c225b2b48647cd5b34619cf23106b845ee25b80d979f8b69e8733148842177500dc48989177b6944677f071f1c
kubernetes-client-windows-386.tar.gz d624b8aead053201765b713d337528be82a71328ee3dd569f556868ceeb4904e64584892a016d247608fc4521c00ead7aed5d973b1206caa2d00406532d5b8b4
kubernetes-client-windows-amd64.tar.gz a1cf8c67984dd4eb4610fa05d27fe9e9e4123159f933e3986e9db835b9cf136962168f0003071001e01e2c1831804ba0a366f2495741aa60a41587a69c09cb62

Server Binaries

filename sha512 hash
kubernetes-server-linux-amd64.tar.gz b93982b56371994c540cd11e6bc21808279340617164992c10f30d8e6ae4d5e270e41c1edc0625d3458a18944ec7aa8c273acbbcd718d60b6cacbc24220c42ac
kubernetes-server-linux-arm.tar.gz bfd76c6b26e5927166d776f6110b97ee36c1d63ad39e2d18899f3e428ebb0f9615bb677ac8e9bcc1864c72a40efd71e1314fe6d137f9c6e54f720270929e3f46
kubernetes-server-linux-arm64.tar.gz 6721dec0df9466cd6c056160c73d598296cebb0af9259eb21b693abb8708901bc8bc30e11815e14d00d6eb12b8bb90b699e3119b922da855e2c411bdf229d6e5
kubernetes-server-linux-ppc64le.tar.gz f8cd307db8141d989ae1218dd2b438bc9cee017d533b1451d2345f9689c451fdb080acd1b9b2f535ed04017e44b81a0585072e7d58a9d201a0ec28fd09df0a6f
kubernetes-server-linux-s390x.tar.gz de7514bbd87a1b363e1bc7787f37d5ea10faac4afe7c5163c23c4df16781aa77570ec553bc4f4b6094166c1fcfc3c431f13e51ffa32f7ea2849e76ec0151ea35

Node Binaries

filename sha512 hash
kubernetes-node-linux-amd64.tar.gz 8c37fd2fe6232d2c148e23df021b8b5347136263399932bcdff0c7a0186f3145de9ede4936b14de7484cc6db9241517d79b5306c380ed374396882900b63e912
kubernetes-node-linux-arm.tar.gz 389e4e77ab9e62968a25b8f4e146a2c3fbb3db2e60e051922edf6395c26cc5380e5a77bf67022339d6ebfe9abd714636d77510bbc42924b4265fdb245fae08c9
kubernetes-node-linux-arm64.tar.gz 7efc32dfeefcef7f860913c25431bd891a435e92cb8d5a95f8deca1a82aa899a007d4b19134493694a4bccb5564867488634a780c128f0cf82c61d98afa889f5
kubernetes-node-linux-ppc64le.tar.gz da30c03bca4b81d810a7df006db02333dea87e336d6cdca9c93392e01c7e43bf4902c969efa7fa53e8a70a0e863b403ec26b87bd38226b8b9f98777ddb0051a0
kubernetes-node-linux-s390x.tar.gz cce43b7f0350b9e5a77ea703225adb9714ef022d176db5b99a0327937d19021d7a8e93ef1169389fd53b895bb98725d23c7565ef80afdd17596c26daf41eeeac
kubernetes-node-windows-amd64.tar.gz d3accf522d80cbfb3d03e9eaa60a09767ba11e88a8a5b44a629192a7c6916b1fb3440f022a5ffc4ea78f3595f254a42f028dd428d117360091cd0c747ec39eb5

Changelog since v1.14.0-alpha.1

Action Required

  • Promote ValidateProxyRedirects to Beta, and enable by default. This feature restricts redirect following from the apiserver to same-host redirects. (#72552, @tallclair)
    • ACTION REQUIRED: If nodes are configured to respond to CRI streaming requests on a different host interface than what the apiserver makes requests on (only the case if not using the built-in dockershim & setting the kubelet flag --redirect-container-streaming=true), then these requests will be broken. In that case, the feature can be temporarily disabled until the node configuration is corrected. We suggest setting --redirect-container-streaming=false on the kubelet to avoid issues.

Other notable changes

  • Added alpha field storageVersionHash to the discovery document for each resource. Its value must be treated as opaque by clients. Only equality comparison on the value is valid. (#73191, @caesarxuchao)
  • Fix admission metrics in seconds. (#72343, @danielqsj)
    • Add metrics *_admission_latencies_milliseconds and *_admission_latencies_milliseconds_summary for backward compatible, but will be removed in a future release.
  • Pod eviction now honors graceful deletion by default if no delete options are provided in the eviction request (#72730, @liggitt)
  • Update to go1.11.5 (#73326, @ixdy)
  • Change proxy metrics to conform metrics guidelines. (#72334, @danielqsj)
    • The metrics sync_proxy_rules_latency_microseconds is deprecated, and will be removed in a future release, please convert to metricssync_proxy_rules_latency_seconds.
  • Add network stats for Windows nodes and pods. (#70121, @feiskyer)
  • kubeadm: When certificates are present joining a new control plane make sure that they match at least the required SANs (#73093, @ereslibre)
  • A new TaintNodesByCondition admission plugin taints newly created Node objects as "not ready", to fix a race condition that could cause pods to be scheduled on new nodes before their taints were updated to accurately reflect their reported conditions. This admission plugin is enabled by default if the TaintNodesByCondition feature is enabled. (#73097, @bsalamat)
  • kube-addon-manager was updated to v9.0, and now uses kubectl v1.13.2 and prunes workload resources via the apps/v1 API (#72978, @liggitt)
  • When a watch is closed by an HTTP2 load balancer and we are told to go away, skip printing the message to stderr by default. (#73277, @smarterclayton)
  • If you are running the cloud-controller-manager and you have the pvlabel.kubernetes.io alpha Initializer enabled, you must now enable PersistentVolume labeling using the PersistentVolumeLabel admission controller instead. You can do this by adding PersistentVolumeLabel in the --enable-admission-plugins kube-apiserver flag. (#73102, @andrewsykim)
  • The alpha Initializers feature, admissionregistration.k8s.io/v1alpha1 API version, Initializers admission plugin, and use of the metadata.initializers API field have been removed. Discontinue use of the alpha feature and delete any existing InitializerConfiguration API objects before upgrading. The metadata.initializers field will be removed in a future release. (#72972, @liggitt)
  • Scale max-inflight limits together with master VM sizes. (#73268, @wojtek-t)
  • kubectl supports copying files with wild card (#72641, @dixudx)
  • kubeadm: add back --cert-dir option for kubeadm init phase certs sa (#73239, @mattkelly)
  • Remove deprecated args '--show-all' (#69255, @Pingan2017)
  • As per deprecation policy in https://kubernetes.io/docs/reference/using-api/deprecation-policy/ (#73001, @shivnagarajan)
    • the taints "node.alpha.kubernetes.io/notReady" and "node.alpha.kubernetes.io/unreachable". are no
    • longer supported or adjusted. These uses should be replaced with "node.kubernetes.io/not-ready"
    • and "node.kubernetes.io/unreachable" respectively instead.
  • The /swagger.json and /swagger-2.0.0.pb-v1 schema documents, deprecated since v1.10, have been removed in favor of /openapi/v2 (#73148, @liggitt)
  • CoreDNS is only officially supported on Linux at this time. As such, when kubeadm is used to deploy this component into your kubernetes cluster, it will be restricted (using nodeSelectors) to run only on nodes with that operating system. This ensures that in clusters which include Windows nodes, the scheduler will not ever attempt to place CoreDNS pods on these machines, reducing setup latency and enhancing initial cluster stability. (#69940, @MarcPow)
  • kubeadm now attempts to detect an installed CRI by its usual domain socket, so that --cri-socket can be omitted from the command line if Docker is not used and there is a single CRI installed. (#69366, @rosti)
  • scheduler: makes pod less racing so as to be put back into activeQ properly (#73078, @Huang-Wei)
  • jsonpath expressions containing [start:end:step] slice are now evaluated correctly (#73149, @liggitt)
  • metadata.deletionTimestamp is no longer moved into the future when issuing repeated DELETE requests against a resource containing a finalizer. (#73138, @liggitt)
  • The "kubectl api-resources" command will no longer fail to display any resources on a single failure (#73035, @juanvallejo)
  • e2e tests that require SSH may be used against clusters that have nodes without external IP addresses by setting the environment variable KUBE_SSH_BASTION to the host:port of a machine that is allowed to SSH to those nodes. The same private key that the test would use is used for the bastion host. The test connects to the bastion and then tunnels another SSH connection to the node. (#72286, @smarterclayton)
  • kubeadm: explicitly wait for etcd to have grown when joining a new control plane (#72984, @ereslibre)
  • Install CSINodeInfo and CSIDriver CRDs in the local cluster. (#72584, @xing-yang)
  • kubectl loads config file once and uses persistent client config (#71117, @dixudx)
  • remove stale OutOfDisk condition from kubelet side (#72507, @dixudx)
  • Node OS/arch labels are promoted to GA (#73048, @yujuhong)
  • Fix graceful apiserver shutdown to not drop outgoing bytes before the process terminates. (#72970, @sttts)
  • Change apiserver metrics to conform metrics guidelines. (#72336, @danielqsj)
    • The following metrics are deprecated, and will be removed in a future release:
      • apiserver_request_count
      • apiserver_request_latencies
      • apiserver_request_latencies_summary
      • apiserver_dropped_requests
      • etcd_helper_cache_hit_count
      • etcd_helper_cache_miss_count
      • etcd_helper_cache_entry_count
      • etcd_request_cache_get_latencies_summary
      • etcd_request_cache_add_latencies_summary
      • etcd_request_latencies_summary
      • transformation_latencies_microseconds
      • data_key_generation_latencies_microseconds
    • Please convert to the following metrics:
      • apiserver_request_total
      • apiserver_request_latency_seconds
      • apiserver_dropped_requests_total
      • etcd_helper_cache_hit_total
      • etcd_helper_cache_miss_total
      • etcd_helper_cache_entry_total
      • etcd_request_cache_get_latency_seconds
      • etcd_request_cache_add_latency_seconds
      • etcd_request_latency_seconds
      • transformation_latencies_seconds
      • data_key_generation_latencies_seconds
  • acquire lock before operating unschedulablepodsmap (#73022, @denkensk)
  • Print SizeLimit of EmptyDir in kubectl describe pod outputs. (#69279, @dtaniwaki)
  • add goroutine to move unschedulable pods to activeq if they are not retried for more than 1 minute (#72558, @denkensk)
  • PidPressure evicts pods from lowest priority to highest priority (#72844, @dashpole)
  • Reduce GCE log rotation check from 1 hour to every 5 minutes. Rotation policy is unchanged (new day starts, log file size > 100MB). (#72062, @jpbetz)
  • Add support for max attach limit for Cinder (#72980, @gnufied)
  • Fixes the setting of NodeAddresses when using the vSphere CloudProvider and nodes that have multiple IP addresses. (#70805, @danwinship)
  • kubeadm: pull images when joining a new control plane instance (#72870, @MalloZup)
  • Enable mTLS encription between etcd and kube-apiserver in GCE (#70144, @wenjiaswe)
  • The /swaggerapi/* schema docs, deprecated since 1.7, have been removed in favor of the /openapi/v2 schema docs. (#72924, @liggitt)
  • Allow users to use Docker 18.09 with kubeadm (#72823, @dims)

v1.14.0-alpha.1

Documentation

Downloads for v1.14.0-alpha.1

filename sha512 hash
kubernetes.tar.gz fac80e5674e547d00987516fb2eca6ea9947529307566be6a12932e3c9e430e8ad094afae748f31e9574838d98052423e3634a067f1456f7c13f6b27bfa63bcc
kubernetes-src.tar.gz d1b5b2c15cb0daa076606f4ccf887724b0166dee0320f2a61d16ab4689931ab0cf5dac4c499aea3d434eb96d589d2b3effe0037e2244978d4290bd19b9a3edea

Client Binaries

filename sha512 hash
kubernetes-client-darwin-386.tar.gz 307c426e4abaf81648af393ddd641c225d87b02d8662d1309fe3528f14ed91b2470f6b46dc8ce0459cf196e2cec906f7eb972bf4c9a96cbd570e206f5a059dca
kubernetes-client-darwin-amd64.tar.gz 8daa85f3e8feaea0d55f20f850038dd113f0f08b62eef944b08a9109d4e69f323a8fcf20c12790c78386b454148bcc9a0cdf106ba3393620709d185c291887fa
kubernetes-client-linux-386.tar.gz 28d73c299cb9859fdfeb3e4869a7a9c77f5679309c2613bd2c72d92dafd5faad0653a7377616190edd29cb8fa1aff104daba98f398e72f3447a132f208dde756
kubernetes-client-linux-amd64.tar.gz eb923e13026f80b743a57100d4f94995f322ab6f107c34ffd9aa74b5a6c6a4a410aff8921a4f675ace7db2ff8158a90874b8f56d3142ad2cbe615c11ec2d4535
kubernetes-client-linux-arm.tar.gz 279b0d0c560900021abea4bbfc25aeca7389f0b37d80022dc3335147344663424e7ba6a0abecb2dca1d2facb4163e26080750736a9a1932d67422f88b0940679
kubernetes-client-linux-arm64.tar.gz d69d28361b9c9e16f3e6804ccda92d55ee743e63aba7fded04edf1f7202b1fa96c235e36ab2ca17df99b4aede80b92150790885bdb7f5b4d7956af3c269dd83c
kubernetes-client-linux-ppc64le.tar.gz ca6ebb87df98bf179c94f54a4e8ae2ef2ea534b1bc5014331f937aa9d4c0442d5423651457871ef5c51f481ba8a3f449d69ef7e42e49c1b313f66cff3d44926f
kubernetes-client-linux-s390x.tar.gz 13fa2058ceba66d8da5ba5982aa302cdd1c61d15253183ab97739229584a178f057f7979b49a035cb2355197dbb388d1642939e2c002b10e23263127030022ab
kubernetes-client-windows-386.tar.gz 42ba4bba477e2958aab674a0fbf888bd5401fa5fbc39466b6cad0fc97e249ac949042c513bf176957bcb336a906e612d9c6790215e78c280225351236ec96993
kubernetes-client-windows-amd64.tar.gz d5f339fe4d37c61babc97208446d1859423b7679f34040f72e9138b72a18d982e66732d1f4b4f3443700f9cbe96bfc0e12eaec0a8a373fb903b49efdafcbae04

Server Binaries

filename sha512 hash
kubernetes-server-linux-amd64.tar.gz bcbcbd3ac4419e54e894d1e595f883e61fcf9db0353a30d794a9e5030cde8957abe8124fa5265e8c52fbc93f07cfe79b2493f791dc225468bf927b7ab4694087
kubernetes-server-linux-arm.tar.gz fda4ea9168555f724659601b06737dea6ec95574569df4ef7e4ab6c2cca3327623ef310bf34f792767f00ee8069b9dd83564835d43daf973087be816be40010b
kubernetes-server-linux-arm64.tar.gz c142857711ec698844cd61188e70b5ab185ba2c8828cf5563a2f42958489e2ae4dbb2c1626271d4f5582167bb363e55ed03afb15e7e86cd414e0dc049fe384c0
kubernetes-server-linux-ppc64le.tar.gz 524a40c5717b24c5a3b2491c4c61cf3038ba5ae7f343797a1b56a5906d6a0a3eb57e9ae78590c28ac3d441d9d1bb480a0c264a07e009a4365503ad2357614aa8
kubernetes-server-linux-s390x.tar.gz ef943fe326b05ece57f2e409ab1cc5fe863f5effa591abae17181c84a5eb4061e9f394ffcc8ee6ebb3f5165b183bab747a8cef540cbb1436343e8180cec037e0

Node Binaries

filename sha512 hash
kubernetes-node-linux-amd64.tar.gz 396f7588e9131dd1b99d101c8bb94fb7e67ab067327ee58dab5a6e24887d8fbb6fc78fe50804abb0ab2f626034881d4280b3f678a1fd8b34891762bf2172b268
kubernetes-node-linux-arm.tar.gz b75c1550438da0b66582d6de90436ee3c44e41e67f74947d93ee9a07ed2b7757762f3f2b05bd7b5589d7e1ea2eb3616b2ef4fe59a9fbe9d8e7cb8f0c9d3dd158
kubernetes-node-linux-arm64.tar.gz b6c46f9250b5565fa178ecc99ffedc6724b0bfffb73acc7d3da2c678af71008a264502cc4a48a6e7452bd0a60d77194141bbc2ea9af49176ea66e27d874b77ac
kubernetes-node-linux-ppc64le.tar.gz 8d505c61a59bc9fc53d6f219d6434ddd962ba383654c46e16d413cee0ad6bd26f276a9860ad3680349bcfacb361e75de07fc44f7d14c054c47b6bd0eae63615f
kubernetes-node-linux-s390x.tar.gz 83b6cf0fb348faa93fa40ec2a947b202b3a5a2081c3896ae39618f947a57b431bc774fbe3a5437719f50f002de252438dc16bac6f632c11140f55d5051094ae6
kubernetes-node-windows-amd64.tar.gz 43471680533685c534023787cd40431b67041bab43e93dea457283ee0f08a8fa02ee9ade3737d8e64d1d3255a281af9a107cb61f9e4d9c99dee188c82a075580

Changelog since v1.13.0

Action Required

  • action required (#68753, @johnSchnake)
    • If you are running E2E tests which require SSH keys and you utilize environment variables to override their location, you may need to modify the environment variable set. On all providers the environment variable override can now be either an absolute path to the key or a relative path (relative to ~/.ssh). Specifically the changes are:
      • Created new GCE_SSH_KEY allowing specification of SSH keys for gce, gke, and kubemark.
      • AWS_SSH_KEY, previously assumed to be an absolute path can now be either relative or absolute
      • LOCAL_SSH_KEY (for local and vsphere providers) was previously assumed to be a filename relative to ~/.ssh but can now also be an absolute path
      • KUBE_SSH_KEY (for skeleton provider) was previously assumed to be a filename relative to ~/.ssh but can now also be an absolute path

Other notable changes

  • Connections from Pods to Services with 0 endpoints will now ICMP reject immediately, rather than blackhole and timeout. (#72534, @thockin)
  • Improve efficiency of preemption logic in clusters with many pending pods. (#72895, @bsalamat)
  • Change scheduler metrics to conform metrics guidelines. (#72332, @danielqsj)
    • The following metrics are deprecated, and will be removed in a future release:
      • e2e_scheduling_latency_microseconds
      • scheduling_algorithm_latency_microseconds
      • scheduling_algorithm_predicate_evaluation
      • scheduling_algorithm_priority_evaluation
      • scheduling_algorithm_preemption_evaluation
      • binding_latency_microseconds
    • Please convert to the following metrics:
      • e2e_scheduling_latency_seconds
      • scheduling_algorithm_latency_seconds
      • scheduling_algorithm_predicate_evaluation_seconds
      • scheduling_algorithm_priority_evaluation_seconds
      • scheduling_algorithm_preemption_evaluation_seconds
      • binding_latency_seconds
  • Fix SelectorSpreadPriority scheduler to match all selectors when distributing pods. (#72801, @Ramyak)
  • Add bootstrap service account & cluster roles for node-lifecycle-controller, cloud-node-lifecycle-controller, and cloud-node-controller. (#72764, @andrewsykim)
  • Fixes spurious 0-length API responses. (#72856, @liggitt)
  • Updates Fluentd to 1.3.2 & added filter_parser (#71180, @monotek)
  • The leaderelection package allows the lease holder to release its lease when the calling context is cancelled. This allows (#71490, @smarterclayton)
    • faster handoff when a leader-elected process is gracefully terminated.
  • Make volume binder resilient to races between main schedule loop and async binding operation (#72045, @cofyc)
  • Bump minimum docker API version to 1.26 (1.13.1) (#72831, @yujuhong)
  • If the TokenRequestProjection feature gate is disabled, projected serviceAccountToken volume sources are now dropped at object creation time, or at object update time if the existing object did not have a projected serviceAccountToken volume source. Previously, these would result in validation errors. (#72714, @mourya007)
  • Add metrics-port to kube-proxy cmd flags. (#72682, @whypro)
  • kubectl: fixed an issue with "too old resource version" errors continuously appearing when calling kubectl delete (#72825, @liggitt)
  • [Breaking change, client-go]: The WaitFor function returns, probably an ErrWaitTimeout, when the done channel is closed, even if the WaitFunc doesn't handle the done channel. (#72364, @kdada)
  • removes newline from json output for windows nodes #72657 (#72659, @jsturtevant)
  • The DenyEscalatingExec and DenyExecOnPrivileged admission plugins are deprecated and will be removed in v1.18. Use of PodSecurityPolicy or a custom admission plugin to limit creation of pods is recommended instead. (#72737, @liggitt)
  • Fix describe statefulset not printing number of desired replicas correctly (#72781, @tghartland)
  • Fix kube-proxy PodSecurityPolicy binding on GCE & GKE. This was only an issue when running kube-proxy as a DaemonSet, with PodSecurityPolicy enabled. (#72761, @tallclair)
  • Drops status.Conditions of new PersistentVolume objects if it was not set on the old object during PrepareForUpdate. (#72739, @rajathagasthya)
  • kubelet: fixes cadvisor internal error when "--container-runtime-endpoint" is set to "unix:///var/run/crio/crio.sock". (#72340, @makocchi-git)
  • The spec.SecurityContext.Sysctls field is now dropped during creation of Pod objects unless the Sysctls feature gate is enabled. (#72752, @rajathagasthya)
    • The spec.AllowedUnsafeSysctls and spec.ForbiddenSysctls fields are now dropped during creation of PodSecurityPolicy objects unless the Sysctls feature gate is enabled.
  • kubeadm: fixed storing of front-proxy certificate in secrets required by kube-controller-manager selfhosting pivoting (#72727, @bart0sh)
  • Administrator is able to configure max pids for a pod on a node. (#72076, @derekwaynecarr)
  • Move users of factory.NewConfigFactory to scheduler.New. (#71875, @wgliang)
  • The spec.SecurityContext.ShareProcessNamespace field is now dropped during creation of Pod objects unless the PodShareProcessNamespace feature gate is enabled. (#72698, @rajathagasthya)
  • kube-apiserver: When configuring integration with external KMS Providers, users can supply timeout value (i.e. how long should kube-apiserver wait before giving up on a call to KMS). (@immutableT ) (#72540, @immutableT)
  • The spec.readinessGates field is now dropped during creation of Pod objects unless the PodReadinessGates feature gate is enabled. (#72695, @rajathagasthya)
  • The spec.dataSource field is now dropped during creation of PersistentVolumeClaim objects unless the VolumeSnapshotDataSource feature gate is enabled. (#72666, @rajathagasthya)
  • Stop kubelet logging a warning to override hostname if there's no change detected. (#71560, @KashifSaadat)
  • client-go: fake clients now properly return NotFound errors when attempting to patch non-existent objects (#70886, @bouk)
  • kubectl: fixes a bug determining the correct namespace while running in a pod when the --context flag is explicitly specified, and the referenced context specifies the namespace default (#72529, @liggitt)
  • Fix scheduling starvation of pods in cluster with large number of unschedulable pods. (#72619, @everpeace)
  • If the AppArmor feature gate is disabled, AppArmor-specific annotations in pod and pod templates are dropped when the object is created, and during update of objects that do not already contain AppArmor annotations, rather than triggering a validation error. (#72655, @liggitt)
  • client-go: shortens refresh period for token files to 1 minute to ensure auto-rotated projected service account tokens are read frequently enough. (#72437, @liggitt)
  • Multiple tests which previously failed due to lack of external IP addresses defined on the nodes should now be passable. (#68792, @johnSchnake)
  • kubeadm: fixed incorrect controller manager pod mutations during selfhosting pivoting (#72518, @bart0sh)
  • Increase Azure default maximumLoadBalancerRuleCount to 250. (#72621, @feiskyer)
  • RuntimeClass is now printed with extra RUNTIME-HANDLER column. (#72446, @Huang-Wei)
  • Updates the kubernetes dashboard add-on to v1.10.1. Skipping dashboard login is no longer enabled by default. (#72495, @liggitt)
  • [GCP] Remove confusing error log entry form fluentd scalers. (#72243, @cezarygerard)
  • change azure disk host cache to ReadOnly by default (#72229, @andyzhangx)
  • Nodes deleted in the cloud provider with Ready condition Unknown should also be deleted on the API server. (#72559, @andrewsykim)
  • kubectl apply --prune now uses the apps/v1 API to prune workload resources (#72352, @liggitt)
  • Fixes a bug in HPA controller so HPAs are always updated every resyncPeriod (15 seconds). (#72373, @krzysztof-jastrzebski)
  • IPVS: "ExternalTrafficPolicy: Local" now works with LoadBalancer services using loadBalancerIP (#72432, @lbernail)
  • Fixes issue with cleaning up stale NFS subpath mounts (#71804, @msau42)
  • Modify the scheduling result struct and improve logging for successful binding. (#71926, @wgliang)
  • Run one etcd storage compaction per default interval of 5min. Do not run one for each resource and each CRD. This fixes the compaction log spam and reduces load on etcd. (#68557, @sttts)
  • kube-apiserver: --runtime-config can once again be used to enable/disable serving specific resources in the extensions/v1beta1 API group. Note that specific resource enablement/disablement is only allowed for the extensions/v1beta1 API group for legacy reasons. Attempts to enable/disable individual resources in other API groups will print a warning, and will return an error in future releases. (#72249, @liggitt)
  • kubeadm: fixed storing of etcd certificates in secrets required by kube-apiserver selfhosting pivoting (#72478, @bart0sh)
  • kubeadm: remove the deprecated "--address" flag for controller-manager and scheduler. (#71973, @MalloZup)
  • kube-apiserver: improves performance of requests made with service account token authentication (#71816, @liggitt)
  • Use prometheus conventions for workqueue metrics. (#71300, @danielqsj)
    • It is now deprecated to use the following metrics:
      • {WorkQueueName}_depth
      • {WorkQueueName}_adds
      • {WorkQueueName}_queue_latency
      • {WorkQueueName}_work_duration
      • {WorkQueueName}_unfinished_work_seconds
      • {WorkQueueName}_longest_running_processor_microseconds
      • {WorkQueueName}_retries
    • Please convert to the following metrics:
      • workqueue_depth
      • workqueue_adds_total
      • workqueue_queue_latency_seconds
      • workqueue_work_duration_seconds
      • workqueue_unfinished_work_seconds
      • workqueue_longest_running_processor_seconds
      • workqueue_retries_total
  • Fix inability to use k8s with dockerd having default IPC mode set to private. (#70826, @kolyshkin)
  • Fix a race condition in the scheduler preemption logic that could cause nominatedNodeName of a pod not to be considered in one or more scheduling cycles. (#72259, @bsalamat)
  • Fix registration for scheduling framework plugins with the default plugin set (#72396, @y-taka-23)
  • The GA VolumeScheduling feature gate can no longer be disabled and will be removed in a future release (#72382, @liggitt)
  • Fix race condition introduced by graceful termination which can lead to a deadlock in kube-proxy (#72361, @lbernail)
  • Fixes issue where subpath volume content was deleted during orphaned pod cleanup for Local volumes that are directories (and not mount points) on the root filesystem. (#72291, @msau42)
  • Fixes kubectl create secret docker-registry compatibility (#72344, @liggitt)
  • Add-on manifests now use the apps/v1 API for DaemonSets, Deployments, and ReplicaSets (#72203, @liggitt)
  • "kubectl wait" command now supports the "--all" flag to select all resources in the namespace of the specified resource types. (#70599, @caesarxuchao)
  • deployments/rollback is now passed through validation/admission controllers (#72271, @jhrv)
  • The Lease API type in the coordination.k8s.io API group is promoted to v1 (#72239, @wojtek-t)
  • Move compatibility_test.go to pkg/scheduler/api (#72014, @huynq0911)
  • New Azure cloud provider option 'cloudProviderBackoffMode' has been added to reduce Azure API retries. Candidate values are: (#70866, @feiskyer) * default (or empty string): keep same with before. * v2: only backoff retry with Azure SDK with fixed exponent 2.
  • Set percentage of nodes scored in each cycle dynamically based on the cluster size. (#72140, @wgliang)
  • Fix AAD support for Azure sovereign cloud in kubectl (#72143, @karataliu)
  • Make kube-proxy service abstraction optional. (#71355, @bradhoekstra)
    • Add the 'service.kubernetes.io/service-proxy-name' label to a Service to disable the kube-proxy service proxy implementation.
  • kubectl: -A can now be used as a shortcut for --all-namespaces (#72006, @soltysh)
  • discovery.CachedDiscoveryInterface implementation returned by NewMemCacheClient has changed semantics of Invalidate method -- the cache refresh is now deferred to the first cache lookup. (#70994, @mborsz)
  • Fix device mountable volume names in DSW to prevent races in device mountable plugin, e.g. local. (#71509, @cofyc)
  • Enable customize in kubectl: kubectl will be able to recognize directories with kustomization.YAML (#70875, @Liujingfang1)
  • Stably sort controllerrevisions. This can prevent pods of statefulsets from continually rolling. (#66882, @ryanmcnamara)
  • Update to use go1.11.4. (#72084, @ixdy)
  • fixes an issue deleting pods containing subpath volume mounts with the VolumeSubpath feature disabled (#70490, @liggitt)
  • Clean up old eclass code (#71399, @resouer)
  • Fix a race condition in which kubeadm only waits for the kubelets kubeconfig file when it has performed the TLS bootstrap, but wasn't waiting for certificates to be present in the filesystem (#72030, @ereslibre)
  • In addition to restricting GCE metadata requests to known APIs, the metadata-proxy now restricts query strings to known parameters. (#71094, @dekkagaijin)
  • kubeadm: fix a possible panic when joining a new control plane node in HA scenarios (#72123, @anitgandhi)
  • fix race condition when attach azure disk in vmss (#71992, @andyzhangx)
  • Update to use go1.11.3 with fix for CVE-2018-16875 (#72035, @seemethere)
  • kubeadm: fix a bug when syncing etcd endpoints (#71945, @pytimer)
  • fix kubelet log flushing issue in azure disk (#71990, @andyzhangx)
  • Disable proxy to loopback and linklocal (#71980, @micahhausler)
  • Fix overlapping filenames in diff if multiple resources have the same name. (#71923, @apelisse)
  • fix issue: vm sku restriction policy does not work in azure disk attach/detach (#71941, @andyzhangx)
  • kubeadm: Create /var/lib/etcd with correct permissions (0700) by default. (#71885, @dims)
  • Scheduler only activates unschedulable pods if node's scheduling related properties change. (#71551, @mlmhl)
  • kube-proxy in IPVS mode will stop initiating connections to terminating pods for services with sessionAffinity set. (#71834, @lbernail)
  • kubeadm: improve hostport parsing error messages (#71258, @bart0sh)
  • Support graceful termination with IPVS when deleting a service (#71895, @lbernail)
  • Include CRD for BGPConfigurations, needed for calico 2.x to 3.x upgrade. (#71868, @satyasm)
  • apply: fix detection of non-dry-run enabled servers (#71854, @apelisse)
  • Clear UDP conntrack entry on endpoint changes when using nodeport (#71573, @JacobTanenbaum)
  • Add successful and failed history limits to cronjob describe (#71844, @soltysh)
  • kube-controller-manager: fixed issue display help for the deprecated insecure --port flag (#71601, @liggitt)
  • kubectl: fixes regression in --sort-by behavior (#71805, @liggitt)
  • Fixes pod deletion when cleaning old cronjobs (#71801, @soltysh)
  • kubeadm: use kubeconfig flag instead of kubeconfig-dir on init phase bootstrap-token (#71803, @yagonobre)
  • kube-scheduler: restores ability to run without authentication configuration lookup permissions (#71755, @liggitt)
  • Add aggregator_unavailable_apiservice_{count,gauge} metrics in the kube-aggregator. (#71380, @sttts)
  • Fixes apiserver nil pointer panics when requesting v2beta1 autoscaling object metrics (#71744, @yue9944882)
  • Only use the first IP address got from instance metadata. This is because Azure CNI would set up a list of IP addresses in instance metadata, while only the first one is the Node's IP. (#71736, @feiskyer)
  • client-go: restores behavior of populating the BearerToken field in rest.Config objects constructed from kubeconfig files containing tokenFile config, or from in-cluster configuration. An additional BearerTokenFile field is now populated to enable constructed clients to periodically refresh tokens. (#71713, @liggitt)
  • kubeadm: remove deprecated kubeadm config print-defaults command (#71467, @rosti)
  • hack/local-up-cluster.sh now enables kubelet authentication/authorization by default (they can be disabled with KUBELET_AUTHENTICATION_WEBHOOK=false and KUBELET_AUTHORIZATION_WEBHOOK=false (#71690, @liggitt)
  • Fixes an issue where Azure VMSS instances not existing in Azure were not being deleted by the Cloud Controller Manager. (#71597, @marc-sensenich)
  • kubeadm reset correcty unmounts mount points inside /var/lib/kubelet (#71663, @bart0sh)
  • Upgrade default etcd server to 3.3.10 (#71615, @jpbetz)
  • When creating a service with annotation: service.beta.kubernetes.io/load-balancer-source-ranges containing multiple source ranges and service.beta.kubernetes.io/azure-shared-securityrule: "false", the NSG rules will be collapsed. (#71484, @ritazh)
  • disable node's proxy use of http probe (#68663, @WanLinghao)
  • Bumps version of kubernetes-cni to 0.6.0 (#71629, @mauilion)
  • On GCI, NPD starts to monitor kubelet, docker, containerd crashlooping, read-only filesystem and corrupt docker overlay2 issues. (#71522, @wangzhen127)
  • When a kubelet is using --bootstrap-kubeconfig and certificate rotation, it no longer waits for bootstrap to succeed before launching static pods. (#71174, @smarterclayton)
  • Add an plugin interfaces for "reserve" and "prebind" extension points of the scheduling framework. (#70227, @bsalamat)
  • Fix scheduling starvation of pods in cluster with large number of unschedulable pods. (#71488, @bsalamat)
  • Reduce CSI log and event spam. (#71581, @saad-ali)
  • Add conntrack as a dependency of kubelet and kubeadm when building rpms and debs. Both require conntrack to handle cleanup of connections. (#71540, @mauilion)
  • UDP connections now support graceful termination in IPVS mode (#71515, @lbernail)
  • Log etcd client errors. The verbosity is set with the usual -v flag. (#71318, @sttts)
  • The DefaultFeatureGate package variable now only exposes readonly feature gate methods. Methods for mutating feature gates have moved into a MutableFeatureGate interface and are accessible via the DefaultMutableFeatureGate package variable. Only top-level commands and options setup should access DefaultMutableFeatureGate. (#71302, @liggitt)
  • node.kubernetes.io/pid-pressure toleration is added for DaemonSet pods, and node.kubernetes.io/out-of-disk isn't added any more even if it's a critical pod. (#67036, @Huang-Wei)
  • Update k8s.io/utils to allow for asynchronous process control (#71047, @hoegaarden)
  • Fixes possible panic during volume detach, if corresponding volume plugin became non-attachable (#71471, @mshaverdo)
  • Fix cloud-controller-manager crash when using AWS provider and PersistentVolume initializing controller (#70432, @mvladev)
  • Fixes an issue where Portworx volumes cannot be mounted if 9001 port is already in use on the host and users remap 9001 to another port. (#70392, @harsh-px)
  • Fix SubPath printing of VolumeMounts. (#70127, @dtaniwaki)
  • Fixes incorrect paths (missing first letter) when copying files from pods to (#69885, @clickyotomy)
    • local in `kubectl cp'.
  • Fix AWS NLB security group updates where valid security group ports were incorrectly removed (#68422, @kellycampbell)
    • when updating a service or when node changes occur.