Automatic merge from submit-queue (batch tested with PRs 46235, 44786, 46833, 46756, 46669)
implements StatefulSet update
**What this PR does / why we need it**:
1. Implements rolling update for StatefulSets
2. Implements controller history for StatefulSets.
3. Makes StatefulSet status reporting consistent with DaemonSet and ReplicaSet.
https://github.com/kubernetes/features/issues/188
**Special notes for your reviewer**:
**Release note**:
```release-note
Implements rolling update for StatefulSets. Updates can be performed using the RollingUpdate, Paritioned, or OnDelete strategies. OnDelete implements the manual behavior from 1.6. status now tracks
replicas, readyReplicas, currentReplicas, and updatedReplicas. The semantics of replicas is now consistent with DaemonSet and ReplicaSet, and readyReplicas has the semantics that replicas did prior to this release.
```
Automatic merge from submit-queue (batch tested with PRs 46235, 44786, 46833, 46756, 46669)
Fixed ResourceConsumer.CleanUp to properly clean up non-replication-controller resources and pods
**What this PR does / why we need it**: Without this fix CleanUp does not remove non-replication-controller resources and pods. This leads to pollution that in some cases inadvertently affects what is happening in AfterEachs before the namespace gets deleted.
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 46997, 47021)
Don't parse human-readable output from gcloud in tests
This is the reason `[k8s.io] Services should be able to change the type and ports of a service [Slow]` is currently failing on GKE e2e tests. For GKE jobs we run a prerelease version of gcloud, in which the default command output was changed.
gcloud's default output for commands is human readable, and is subject to change. Anything scripting against gcloud should always pass `--format=json|yaml|value(...)` so you get standardized output.
fixes: #46918
Automatic merge from submit-queue (batch tested with PRs 47083, 44115, 46881, 47082, 46577)
Add an e2e test for server side get
Print a better error from the response. Performs validation to ensure it
does not regress in alpha state.
This is tests and bug fixes for https://github.com/kubernetes/community/pull/363
@kubernetes/sig-api-machinery-pr-reviews
Implements history utilities for ControllerRevision in the controller/history package
StatefulSetStatus now has additional fields for consistency with DaemonSet and Deployment
StatefulSetStatus.Replicas now represents the current number of createdPods and StatefulSetStatus.ReadyReplicas is the current number of ready Pods
Automatic merge from submit-queue (batch tested with PRs 46718, 46828, 46988)
Update docs/ links to point to main site
**What this PR does / why we need it**:
This updates various links to either point to kubernetes.io or to the kubernetes/community repo instead of the legacy docs/ tree in k/k
Pre-requisite for #46813
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
@kubernetes/sig-docs-maintainers @chenopis @ahmetb @thockin
Go's loop-clause variables are allocated once and the items are copied
to that variable while iterating through the loop. This means, these
variables can't escape the scope since closures are bound to loop-clause
variables whose value change during each iteration. Doing so would lead
to undesired behavior. For more on this topic see:
https://github.com/golang/go/wiki/CommonMistakes
So in order to workaround this problem in sync controller e2e tests, we
iterate through the map and copy the map value to a variable inside the
loop before using it in closures.
Fixes issue: #47059
This commit includes the following two changes:
* Move pre-test checks (pods/nodes ready) to BeforeEach() so that it's
clear whether the test has run or not.
* Dumping logs for unready pods.
Automatic merge from submit-queue (batch tested with PRs 46897, 46899, 46864, 46854, 46875)
Wait for cloud-init to finish before starting tests.
This fixes#46889.
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue
Avoid double printing output of gcloud commands in kubemark
Just noticed we were unnecessarily echoing the result again.
/cc @wojtek-t
Automatic merge from submit-queue
Add EmptyDir volume capacity isolation
This PR adds the support for isolating the emptyDir volume use. If user
sets a size limit for emptyDir volume, kubelet's eviction manager monitors its usage
and evict the pod if the usage exceeds the limit.
This feature is part of local storage capacity isolation and described in the proposal kubernetes/community#306
**Release note**:
```release-note
Alpha feature: allows users to set storage limit to isolate EmptyDir volumes. It enforces the limit by evicting pods that exceed their storage limits
```
Automatic merge from submit-queue (batch tested with PRs 44883, 46836, 46765, 46683, 46050)
Added Bind method to Scheduler Extender
- only one extender can support the bind method
- if an extender supports bind, scheduler delegates the pod binding to the extender
**What this PR does / why we need it**:
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes#41235
**Special notes for your reviewer**:
**Release note**:
```release-note
```
Automatic merge from submit-queue (batch tested with PRs 46972, 42829, 46799, 46802, 46844)
Multizone static pv test
**What this PR does / why we need it**:
Adds an e2e test for checking that pods get scheduled to the same zone as statically created PVs. This tests the PersistentVolumeLabel admission controller, which adds zone and region labels when PVs are created. As part of this, I also had to make changes to volume test utility code to pass in a zone parameter for creating PDs, and also had to add an argument to the e2e test program to accept a list of zones.
Fixes#46995
**Special notes for your reviewer**:
It's probably easier to review each commit separately.
**Release note**:
NONE
Automatic merge from submit-queue (batch tested with PRs 46550, 46663, 46816, 46820, 46460)
[GCE] Support internal load balancers
**What this PR does / why we need it**:
Allows users to expose K8s services externally of the K8s cluster but within their GCP network.
Fixes#33483
**Important User Notes:**
- This is a beta feature. ILB could be enabled differently in the future.
- Requires nodes having version 1.7.0+ (ILB requires health checking and a health check endpoint on kube-proxy has just been exposed)
- This cannot be used for intra-cluster communication. Do not call the load balancer IP from a K8s node/pod.
- There is no reservation system for private IPs. You can specify a RFC 1918 address in `loadBalancerIP` field, but it could be lost to another VM or LB if service settings are modified.
- If you're running an ingress, your existing loadbalancer backend service must be using BalancingMode type `RATE` - not `UTILIZATION`.
- Option 1: With a 1.5.8+ or 1.6.4+ version master, delete all your ingresses, and re-create them.
- Option 2: Migrate to a new cluster running 1.7.0. Considering ILB requires nodes with 1.7.0, this isn't a bad idea.
- Option 3: Possible migration opportunity, but use at your own risk. More to come later.
**Reviewer Notes**:
Several files were renamed, so github thinks ~2k lines have changed. Review commits one-by-one to see the actual changes.
**Release note**:
```release-note
Support creation of GCP Internal Load Balancers from Service objects
```
Handle failure cases on startup gracefully to avoid causing cascading
errors and poor initialization in other components. Initial errors from
config load cause the initializer to pause and hold requests. Return
typed errors to better communicate failures to clients.
Add code to handle two specific cases - admin wants to bypass
initialization defaulting, and mirror pods (which want to bypass
initialization because the kubelet owns their lifecycle).
This PR adds two features:
1. add support for isolating the emptyDir volume use. If user
sets a size limit for emptyDir volume, kubelet's eviction manager
monitors its usage
and evict the pod if the usage exceeds the limit.
2. add support for isolating the local storage for container overlay. If
the container's overly usage exceeds the limit defined in container
spec, eviction manager will evict the pod.
An annotation in the pod spec of the form:
podpreset.admission.kubernetes.io/exclude: "true"
Will cause the admission controller to skip manipulating the pod spec,
no matter the labelling.
The annotation for a podpreset acting on a pod has also been slightly
modified to contain a podpreset prefix:
podpreset.admission.kubernetes.io/podpreset-{name} = resource version
Fixes#44161
Automatic merge from submit-queue (batch tested with PRs 45871, 46498, 46729, 46144, 46804)
PD e2e test: Ready node check now uses the most up-to-date node count.
Follow-up to PR #46746
<!-- Steps to write your release note:
1. Use the release-note-* labels to set the release note state (if you have access)
2. Enter your extended release note in the below block; leaving it blank means using the PR title as the release note. If no release note is required, just write `NONE`.
Automatic merge from submit-queue
Respect PDBs during node upgrades and add test coverage to the ServiceTest upgrade test.
This is still a WIP... needs to be squashed at least, and I don't think it's currently passing until I increase the scale of the RC, but please have a look at the general outline. Thanks!
Fixes#38336
@kow3ns @bdbauer @krousey @erictune @maisem @davidopp
```
On GCE, node upgrades will now respect PodDisruptionBudgets, if present.
```
Automatic merge from submit-queue (batch tested with PRs 46681, 46786, 46264, 46680, 46805)
Enable Dialer on the Aggregator
Centralize the creation of the dialer during startup.
Have the dialer then passed in to both APIServer and Aggregator.
Aggregator the uses the dialer as its Transport base.
**What this PR does / why we need it**:Enables the Aggregator to use the Dialer/SSHTunneler to connect to the user-apiserver.
**Which issue this PR fixes** : fixes ##46679
**Special notes for your reviewer**:
**Release note**: None
Automatic merge from submit-queue
Implement Daemonset history
~Depends on #45867 (the 1st commit, ignore it when reviewing)~ (already merged)
Ref https://github.com/kubernetes/community/pull/527/ and https://github.com/kubernetes/community/pull/594
@kubernetes/sig-apps-api-reviews @kubernetes/sig-apps-pr-reviews @erictune @kow3ns @lukaszo @kargakis
---
TODOs:
- [x] API changes
- [x] (maybe) Remove rollback subresource if we decide to do client-side rollback
- [x] deployment controller
- [x] controller revision
- [x] owner ref (claim & adoption)
- [x] history reconstruct (put revision number, hash collision avoidance)
- [x] de-dup history and relabel pods
- [x] compare ds template with history
- [x] hash labels (put it in controller revision, pods, and maybe deployment)
- [x] clean up old history
- [x] Rename status.uniquifier when we reach consensus in #44774
- [x] e2e tests
- [x] unit tests
- [x] daemoncontroller_test.go
- [x] update_test.go
- [x] ~(maybe) storage_test.go // if we do server side rollback~
kubectl part is in #46144
---
**Release note**:
```release-note
```
Automatic merge from submit-queue (batch tested with PRs 46620, 46732, 46773, 46772, 46725)
Fix AppArmor test for docker 1.13
... & better debugging.
The issue is that we run the pod containers in a shared PID namespace with docker 1.13, so PID 1 is no longer the container's root process. Since it's messy to get the container's root process, I switched to using `/proc/self` to read the apparmor profile. While this wouldn't catch a regression that caused only the init process to run with the wrong profile, I think it's a good approximation.
/cc @aulanov @Amey-D
Automatic merge from submit-queue (batch tested with PRs 46620, 46732, 46773, 46772, 46725)
Added missing documentation to NodeInstanceGroup.
**What this PR does / why we need it**:
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
```
Automatic merge from submit-queue
Add initializer support to admission and uninitialized filtering to rest storage
Initializers are the opposite of finalizers - they allow API clients to react to object creation and populate fields prior to other clients seeing them.
High level description:
1. Add `metadata.initializers` field to all objects
2. By default, filter objects with > 0 initializers from LIST and WATCH to preserve legacy client behavior (known as partially-initialized objects)
3. Add an admission controller that populates .initializer values per type, and denies mutation of initializers except by certain privilege levels (you must have the `initialize` verb on a resource)
4. Allow partially-initialized objects to be viewed via LIST and WATCH for initializer types
5. When creating objects, the object is "held" by the server until the initializers list is empty
6. Allow some creators to bypass initialization (set initializers to `[]`), or to have the result returned immediately when the object is created.
The code here should be backwards compatible for all clients because they do not see partially initialized objects unless they GET the resource directly. The watch cache makes checking for partially initialized objects cheap. Some reflectors may need to change to ask for partially-initialized objects.
```release-note
Kubernetes resources, when the `Initializers` admission controller is enabled, can be initialized (defaulting or other additive functions) by other agents in the system prior to those resources being visible to other clients. An initialized resource is not visible to clients unless they request (for get, list, or watch) to see uninitialized resources with the `?includeUninitialized=true` query parameter. Once the initializers have completed the resource is then visible. Clients must have the the ability to perform the `initialize` action on a resource in order to modify it prior to initialization being completed.
```
Automatic merge from submit-queue
Add local storage (scratch space) allocatable support
This PR adds the support for allocatable local storage (scratch space).
This feature is only for root file system which is shared by kubernetes
componenets, users' containers and/or images. User could use
--kube-reserved flag to reserve the storage for kube system components.
If the allocatable storage for user's pods is used up, some pods will be
evicted to free the storage resource.
This feature is part of local storage capacity isolation and described in the proposal https://github.com/kubernetes/community/pull/306
**Release note**:
```release-note
This feature exposes local storage capacity for the primary partitions, and supports & enforces storage reservation in Node Allocatable
```
Automatic merge from submit-queue (batch tested with PRs 46239, 46627, 46346, 46388, 46524)
Dynamic webhook admission control plugin
Unit tests pass.
Needs plumbing:
* [ ] service resolver (depends on @wfender PR)
* [x] client cert (depends on ????)
* [ ] hook source (depends on @caesarxuchao PR)
Also at least one thing will need to be renamed after Chao's PR merges.
```release-note
Allow remote admission controllers to be dynamically added and removed by administrators. External admission controllers make an HTTP POST containing details of the requested action which the service can approve or reject.
```
Automatic merge from submit-queue (batch tested with PRs 46239, 46627, 46346, 46388, 46524)
move labels to components which own the APIs
During the apimachinery split in 1.6, we accidentally moved several label APIs into apimachinery. They don't belong there, since the individual APIs are not general machinery concerns, but instead are the concern of particular components: most commonly the kubelet. This pull moves the labels into their owning components and out of API machinery.
@kubernetes/sig-api-machinery-misc @kubernetes/api-reviewers @kubernetes/api-approvers
@derekwaynecarr since most of these are related to the kubelet
Add support for creating resources that are not immediately visible to
naive clients, but must first be initialized by one or more privileged
cluster agents. These controllers can mark the object as initialized,
allowing others to see them.
Permission to override initialization defaults or modify an initializing
object is limited per resource to a virtual subresource "RESOURCE/initialize"
via RBAC.
Initialization is currently alpha.
Automatic merge from submit-queue
[Flaky PR Test] Fix summary test
fixes issue: #46797
As we can see in the [example failure build log](https://storage.googleapis.com/kubernetes-jenkins/logs/ci-kubernetes-node-kubelet/4319/build-log.txt), the summary containers are pinging google 100s of times a second. This causes the summary container to be killed occasionally, and fail the test. The summary containers are only supposed to ping every 10 seconds according to the current test. As it turns out, we were missing a semicolon, and were not sleeping between pings. For background, we ping google to generate network traffic, so that the summary test can validate network metrics.
This PR adds the semicolon to make the container sleep between calls, and decreases the sleep time from 10 seconds to 1 second, as 1 call / 10 seconds did not produce enough activity.
cc @kubernetes/kubernetes-build-cops @dchen1107
Automatic merge from submit-queue (batch tested with PRs 46648, 46500, 46238, 46668, 46557)
Add an e2e test for AdvancedAuditing
Enable a simple "advanced auditing" setup for e2e tests running on GCE, and add an e2e test that creates & deletes a pod, a secret, and verifies that they're audited.
Includes https://github.com/kubernetes/kubernetes/pull/46548
For https://github.com/kubernetes/features/issues/22
/cc @ericchiang @sttts @soltysh @ihmccreery
Automatic merge from submit-queue (batch tested with PRs 46648, 46500, 46238, 46668, 46557)
Support validating package versions in node conformance test
**What this PR does / why we need it**:
This PR adds a package validator in node conformance test for checking whether the locally installed packages meet the image spec.
**Special notes for your reviewer**:
The image spec for GKE (which has the package spec) will be in a separate PR. Then we will publish a new node conformance test image for GKE whose name should use the convention in https://github.com/kubernetes/kubernetes/issues/45760 and have `gke` in it.
**Release note**:
```
NONE
```
Respect PDBs during node upgrades and add test coverage to the
ServiceTest upgrade test. Modified that test so that we include pod
anti-affinity constraints and a PDB.
This PR adds the check for local storage request when admitting pods. If
the local storage request exceeds the available resource, pod will be
rejected.
This PR adds the support for allocatable local storage (scratch space).
This feature is only for root file system which is shared by kubernetes
componenets, users' containers and/or images. User could use
--kube-reserved flag to reserve the storage for kube system components.
If the allocatable storage for user's pods is used up, some pods will be
evicted to free the storage resource.
Automatic merge from submit-queue (batch tested with PRs 43505, 45168, 46439, 46677, 46623)
Add TPR to CRD migration helper.
This is a helper for migrating TPR data to CustomResource. It's rather hacky because it requires crossing apiserver boundaries, but doing it this way keeps the mess contained to the TPR code, which is scheduled for deletion anyway.
It's also not completely hands-free because making it resilient enough to be completely automated is too involved to be worth it for an alpha-to-beta migration, and would require investing significant effort to fix up soon-to-be-deleted TPR code. Instead, this feature will be documented as a best-effort helper whose results should be verified by hand.
The intended benefit of this over a totally manual process is that it should be possible to copy TPR data into a CRD without having to tear everything down in the middle. The process would look like this:
1. Upgrade to k8s 1.7. Nothing happens to your TPRs.
1. Create CRD with group/version and resource names that match the TPR. Still nothing happens to your TPRs, as the CRD is hidden by the overlapping TPR.
1. Delete the TPR. The TPR data is converted to CustomResource data, and the CRD begins serving at the same REST path.
Note that the old TPR data is left behind by this process, so watchers should not receive DELETE events. This also means the user can revert to the pre-migration state by recreating the TPR definition.
Ref. https://github.com/kubernetes/kubernetes/issues/45728
Centralize the creation of the dialer during startup.
Have the dialer then passed in to both APIServer and Aggregator.
Aggregator the sets the dialer on its Transport base.
This should allow the SSTunnel to be used but also allow the Aggregation
Auth to work with it.
Depending on Environment InsecureSkipTLSVerify *may* need to be set to
true.
Fixed as few tests to call CreateDialer as part of start-up.
Automatic merge from submit-queue (batch tested with PRs 46076, 43879, 44897, 46556, 46654)
Local storage plugin
**What this PR does / why we need it**:
Volume plugin implementation for local persistent volumes. Scheduler predicate will direct already-bound PVCs to the node that the local PV is at. PVC binding still happens independently.
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*:
Part of #43640
**Release note**:
```
Alpha feature: Local volume plugin allows local directories to be created and consumed as a Persistent Volume. These volumes have node affinity and pods will only be scheduled to the node that the volume is at.
```
Automatic merge from submit-queue
Node authorizer
This PR implements the authorization portion of https://github.com/kubernetes/community/blob/master/contributors/design-proposals/kubelet-authorizer.md and kubernetes/features#279:
* Adds a new authorization mode (`Node`) that authorizes requests from nodes based on a graph of related pods,secrets,configmaps,pvcs, and pvs:
* Watches pods, adds edges (secret -> pod, configmap -> pod, pvc -> pod, pod -> node)
* Watches pvs, adds edges (secret -> pv, pv -> pvc)
* When both Node and RBAC authorization modes are enabled, the default RBAC binding that grants the `system:node` role to the `system:nodes` group is not automatically created.
* Tightens the `NodeRestriction` admission plugin to require identifiable nodes for requests from users in the `system:nodes` group.
This authorization mode is intended to be used in combination with the `NodeRestriction` admission plugin, which limits the pods and nodes a node may modify. To enable in combination with RBAC authorization and the NodeRestriction admission plugin:
* start the API server with `--authorization-mode=Node,RBAC --admission-control=...,NodeRestriction,...`
* start kubelets with TLS boostrapping or with client credentials that place them in the `system:nodes` group with a username of `system:node:<nodeName>`
```release-note
kube-apiserver: a new authorization mode (`--authorization-mode=Node`) authorizes nodes to access secrets, configmaps, persistent volume claims and persistent volumes related to their pods.
* Nodes must use client credentials that place them in the `system:nodes` group with a username of `system:node:<nodeName>` in order to be authorized by the node authorizer (the credentials obtained by the kubelet via TLS bootstrapping satisfy these requirements)
* When used in combination with the `RBAC` authorization mode (`--authorization-mode=Node,RBAC`), the `system:node` role is no longer automatically granted to the `system:nodes` group.
```
```release-note
RBAC: the automatic binding of the `system:node` role to the `system:nodes` group is deprecated and will not be created in future releases. It is recommended that nodes be authorized using the new `Node` authorization mode instead. Installations that wish to continue giving all members of the `system:nodes` group the `system:node` role (which grants broad read access, including all secrets and configmaps) must create an installation-specific ClusterRoleBinding.
```
Follow-up:
- [ ] enable e2e CI environment with admission and authorizer enabled (blocked by kubelet TLS bootstrapping enablement in https://github.com/kubernetes/kubernetes/pull/40760)
- [ ] optionally enable this authorizer and admission plugin in kubeadm
- [ ] optionally enable this authorizer and admission plugin in kube-up
Automatic merge from submit-queue
Switch gcloud compute copy-files to scp
gcloud is deprecating `gcloud compute copy-files` and switching to `gcloud compute scp`. Make the change before things start to break.
https://cloud.google.com/sdk/gcloud/reference/compute/copy-files
Warnings we get: `W0529 10:28:59.097] WARNING: `gcloud compute copy-files` is deprecated. Please use `gcloud compute scp` instead. Note that `gcloud compute scp` does not have recursive copy on by default. To turn on recursion, use the `--recurse` flag.`
/cc @jlowdermilk
Automatic merge from submit-queue
Support grabbing test suite metrics
**What this PR does / why we need it**:
Add support for grabbing metrics that cover the entire test suite's execution.
Update the "interesting" controller-manager metrics to match the
current names for the garbage collector, and add namespace controller
metrics to the list.
If you enable `--gather-suite-metrics-at-teardown`, the metrics file is written to a file with a name such as `MetricsForE2ESuite_2017-05-25T20:25:57Z.json` in the `--report-dir`. If you don't specify `--report-dir`, the metrics are written to the test log output.
I'd like to enable this for some of the `pull-*` CI jobs, which will require a separate PR to test-infra.
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
@kubernetes/sig-testing-pr-reviews @smarterclayton @wojtek-t @gmarek @derekwaynecarr @timothysc
Automatic merge from submit-queue (batch tested with PRs 45534, 37212, 46613, 46350)
[e2e]Fix define redundant parameter
When timeout to reach HTTP service, redundant parameter make the
error is nil.
Automatic merge from submit-queue
no-snat test
This test checks that Pods can communicate with each other in the same cluster without SNAT.
I intend to create a job that runs this in small clusters (\~3 nodes) at a low frequency (\~once per day) so that we have a signal as we work on allowing multiple non-masquerade CIDRs to be configured (see [kubernetes-incubator/ip-masq-agent](https://github.com/kubernetes-incubator/ip-masq-agent), for example).
/cc @dnardo
