Automatic merge from submit-queue (batch tested with PRs 42637, 42648)
Support multiple --feature-gates flags in the command line
Fixes the issue in https://github.com/kubernetes/kubernetes/pull/42647.
Before this change the whole set of gates was replaced with new values. Now values are overridden one by one.
Automatic merge from submit-queue (batch tested with PRs 41890, 42593, 42633, 42626, 42609)
Remove everything that is not new from batch/v2alpha1
Fixes#37166.
@lavalamp you've asked for it
@erictune this is a prereq for moving CronJobs to beta. I initially planned to put all in one PR, but after I did that I figured out it'll be easier to review separately. ptal
@kubernetes/api-approvers @kubernetes/sig-api-machinery-pr-reviews ptal
Automatic merge from submit-queue (batch tested with PRs 42506, 42585, 42596, 42584)
Preserve custom etcd prefix compatibility for etcd3
Fixes#42505
```release-note
restored normalization of custom `--etcd-prefix` when `--storage-backend` is set to etcd3
```
Automatic merge from submit-queue (batch tested with PRs 42080, 41653, 42598, 42555)
Support whitespace in command path for gcp auth plugin
```
External command option on gcp client auth plugin supports whitespace in command path.
```
Splitting on whitespace to get cmd+args breaks when the path the executable contains spaces. Resolve by adding a new "cmd-args" field to config to allow the full string of "cmd-path" to be interpreted as path to executable.
This change is backwards compatible with existing behavior.
Automatic merge from submit-queue (batch tested with PRs 42080, 41653, 42598, 42555)
StatefulSet: Respect ControllerRef
**What this PR does / why we need it**:
This is part of the completion of the [ControllerRef](https://github.com/kubernetes/community/blob/master/contributors/design-proposals/controller-ref.md) proposal. It brings StatefulSet into full compliance with ControllerRef. See the individual commit messages for details.
**Which issue this PR fixes**:
Fixes#36859
**Special notes for your reviewer**:
**Release note**:
```release-note
StatefulSet now respects ControllerRef to avoid fighting over Pods. At the time of upgrade, **you must not have StatefulSets with selectors that overlap** with any other controllers (such as ReplicaSets), or else [ownership of Pods may change](https://github.com/kubernetes/community/blob/master/contributors/design-proposals/controller-ref.md#upgrading).
```
cc @erictune @kubernetes/sig-apps-pr-reviews
Specific use case is GKE users running gcloud/kubectl on Windows
with a cloud sdk installation path containing spaces. Also improving
test coverage using trick borrowed from exec_test.go
Automatic merge from submit-queue (batch tested with PRs 31783, 41988, 42535, 42572, 41870)
Clean user agent to reduce metrics cardinality
**What this PR does / why we need it**:
This PR is an example implementation for my issue #31781.
``` release-note
```
This commit cleans common browser user-agents to reduce the metrics
cardinality in exported prometheus metrics.
Resolveskubernetes/kubernetes#31781
The StatefulSet Listers still use selectors, because this is the
behavior expected by callers. This clarifies the meaning of the
returned list. Some callers may need to switch to using
GetControllerOf() instead, but that is a separate, case-by-case issue.
Automatic merge from submit-queue
make the system:authenticated group adder smarter
Fixes#42437
This prevents the group adder from adding the system:authenticated group when:
1. it's already in the list
2. the user is system:anonymous
3. system:unauthenticated is in the list
Smaller alternative to https://github.com/kubernetes/kubernetes/pull/42421 for 1.6.
@kubernetes/sig-auth-pr-reviews @enj @liggitt
Automatic merge from submit-queue (batch tested with PRs 42443, 38924, 42367, 42391, 42310)
Apply custom defaults to init containers
Adds overridden defaults to init containers. They were not being defaulted the same way normal containers were.
Automatic merge from submit-queue (batch tested with PRs 41306, 42187, 41666, 42275, 42266)
discovery restmapping should always prefer /v1
The core kube API, empty group, version==v1 should always be the most preferred group and resource from a rest mapper. This special cases that. All the others should be based on discovery order as we previously agreed.
@kubernetes/sig-cli-pr-reviews @kubernetes/sig-api-machinery-pr-reviews
@enj
Automatic merge from submit-queue (batch tested with PRs 41306, 42187, 41666, 42275, 42266)
Server timeout returns an incorrect error
Not a valid Status object in JSON
Part of #42163
Automatic merge from submit-queue (batch tested with PRs 41984, 41682, 41924, 41928)
RC/RS: Fully Respect ControllerRef
**What this PR does / why we need it**:
This is part of the completion of the [ControllerRef](https://github.com/kubernetes/community/blob/master/contributors/design-proposals/controller-ref.md) proposal. It brings ReplicaSet and ReplicationController into full compliance with ControllerRef. See the individual commit messages for details.
**Which issue this PR fixes**:
Although RC/RS had partially implemented ControllerRef, they didn't use it to determine which controller to sync, or to update expectations. This could lead to instability or controllers getting stuck.
Ref: https://github.com/kubernetes/kubernetes/issues/24433
**Special notes for your reviewer**:
**Release note**:
```release-note
```
cc @erictune @kubernetes/sig-apps-pr-reviews
Automatic merge from submit-queue
Only set a groupresource on errors if a resource was specified
@deads2k this gets rid of the weird .meta.k8s.io error on unauthorized errors
Automatic merge from submit-queue (batch tested with PRs 42128, 42064, 42253, 42309, 42322)
Fix panic on nil invalid field error
bug fix for validation panic
if a field.Invalid is constructed with a nil badvalue, the Error() method panics, since reflect.TypeOf() returns nil
Automatic merge from submit-queue (batch tested with PRs 42128, 42064, 42253, 42309, 42322)
Add storage.k8s.io/v1 API
This is combined version of reverted #40088 (first 4 commits) and #41646. The difference is that all controllers and tests use old `storage.k8s.io/v1beta1` API so in theory all tests can pass on GKE.
Release note:
```release-note
StorageClassName attribute has been added to PersistentVolume and PersistentVolumeClaim objects and should be used instead of annotation `volume.beta.kubernetes.io/storage-class`. The beta annotation is still working in this release, however it will be removed in a future release.
```
Automatic merge from submit-queue (batch tested with PRs 41672, 42084, 42233, 42165, 42273)
Keep entries of apimachinery in client-go's Godeps.json
1. copy.sh now copies staging/ to a temp dir, creates snapshots for them, and then adds the temp dir to GOPATH before running `godep save`. This way, we don't need the `go list` check added in #41987 so that's removed @sttts.
2. copy.sh fills the entries for k8s.io/apimachinery with dummy SHA1 in client-go's Godeps.json. The publish robot will later fill the actually published one.
3. copy.sh removes the entire `_vendor` from client-go. We cannot keep `_vendor` because it makes hard to use the `git filter-branch` magic for client-go, because the filtered commits might contain changes to `_vendor` while client-go only has `vendor/`.
Some notes on point 3: the robot will restore the vendor/ later, but exclude k8s.io/* and glog, so that users retrieving client-go with `go get` won't end up with multiple copies of them. I think this should be a ***general rule*** for published k8s.io/* repos: every published repos should have complete Godeps.json with latest commit hashes of other k8s.io/* repos, but vendor/k8s.io/* and vendor/glog will be removed from all repos.
@sttts @lavalamp @deads2k
Automatic merge from submit-queue (batch tested with PRs 42126, 42130, 42232, 42245, 41932)
apiserver/pkg/server: include scheme in insecure self client config
Noticed this during bootkube development: https://github.com/kubernetes-incubator/bootkube/issues/325
In Go 1.8's `url.Parse` became more strict, and `url.Parse("127.0.0.1:8080")` now fails.
https://beta.golang.org/doc/go1.8#net_urlhttps://play.golang.org/p/dw_cPeotG4
Accidentally compiled bootkube with 1.8 and tracked a panic down to the loopback client config. Though we're still using the old "genericapiserver", this seems to be translated from the old one.
The actual panic we observed was
```
E0222 19:40:11.364949 5 server.go:254] Failed to create clientset: parse 127.0.0.1:8080: first path segment in URL cannot contain colon
panic: parse 127.0.0.1:8080: first path segment in URL cannot contain colon
goroutine 35 [running]:
github.com/kubernetes-incubator/bootkube/vendor/k8s.io/kubernetes/pkg/client/clientset_generated/internalclientset/typed/core/internalversion.NewForConfigOrDie(0xc420728ea0, 0xc420738e30)
/home/eric/src/github.com/kubernetes-incubator/bootkube/vendor/k8s.io/kubernetes/pkg/client/clientset_generated/internalclientset/typed/core/internalversion/core_client.go:132 +0x62
github.com/kubernetes-incubator/bootkube/vendor/k8s.io/kubernetes/pkg/master.(*Config).Complete(0xc420739db8, 0x0)
/home/eric/src/github.com/kubernetes-incubator/bootkube/vendor/k8s.io/kubernetes/pkg/master/master.go:180 +0x40a
github.com/kubernetes-incubator/bootkube/vendor/k8s.io/kubernetes/cmd/kube-apiserver/app.Run(0xc42007a3c0, 0x0, 0x0)
/home/eric/src/github.com/kubernetes-incubator/bootkube/vendor/k8s.io/kubernetes/cmd/kube-apiserver/app/server.go:347 +0x1e8d
github.com/kubernetes-incubator/bootkube/pkg/bootkube.(*bootkube).Run.func1(0xc4206b01e0, 0xc420164300)
/home/eric/src/github.com/kubernetes-incubator/bootkube/pkg/bootkube/bootkube.go:124 +0x2f
created by github.com/kubernetes-incubator/bootkube/pkg/bootkube.(*bootkube).Run
/home/eric/src/github.com/kubernetes-incubator/bootkube/pkg/bootkube/bootkube.go:124 +0xb0
```
I don't actually know if this is the correct fix or if there should be changes to `NewForConfigOrDie`. Am looking for comments more than anything.
edit: @abourget pointed out over on bootkube that the actual panic was fixed in the internalclient by https://github.com/kubernetes/kubernetes/pull/38519.
cc @deads2k @sttts @kubernetes/sig-api-machinery-pr-reviews
Automatic merge from submit-queue (batch tested with PRs 41931, 39821, 41841, 42197, 42195)
Admission Controller: Add Pod Preset
Based off the proposal in https://github.com/kubernetes/community/pull/254
cc @pmorie @pwittrock
TODO:
- [ ] tests
**What this PR does / why we need it**: Implements the Pod Injection Policy admission controller
**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
Added new Api `PodPreset` to enable defining cross-cutting injection of Volumes and Environment into Pods.
```
Automatic merge from submit-queue
Add RBAC roles for bootstrap controllers
Supercedes https://github.com/kubernetes/kubernetes/pull/42221
When locking down controllers to individual RBAC roles we need to make sure that the bootstrap controllers have the right permissions.
This adds the roles and bindings at the correct namespace scopes for the bootstrap-signer and token-cleaner controllers.
@liggitt ptal
@jbeda @luxas you got a good way to test this? It must not be covered in normal e2e or we'd've seen the issue before.
The RC/RS Listers still use selectors, because this is the behavior
expected by callers. This clarifies the meaning of the returned list.
Some callers may need to switch to using GetControllerOf() instead,
but that is a separate, case-by-case issue.
Automatic merge from submit-queue
Re-writing of the resolv.conf file generated by docker
Fixes#17406
Docker 1.12 will contain feature "The option --dns and --net=host should not be mutually exclusive" (docker/docker#22408)
This patch adds optional support for this ability in kubelet (for now in case of "hostNetwork: true" set all dns settings are ignored if any).
To enable feature use newly added kubelet flag: --allow-dns-for-hostnet=true
Introduced chages:
1. Re-writing of the resolv.conf file generated by docker.
Cluster dns settings aren't passed anymore to docker api in all cases, not only for pods with host network:
the resolver conf will be overwritten after infra-container creation to override docker's behaviour.
2. Added new one dnsPolicy - 'ClusterFirstWithHostNet', so now there are:
- ClusterFirstWithHostNet - use dns settings in all cases, i.e. with hostNet=true as well
- ClusterFirst - use dns settings unless hostNetwork is true
- Default
Fixes#17406
Automatic merge from submit-queue (batch tested with PRs 41597, 42185, 42075, 42178, 41705)
stop spamming logs on restart of api server
**What this PR does / why we need it**:
This PR reduces an excessive amount of log spam on startup of any API server with larger amounts of content. We have experienced server restart times ~5 minutes that is a consequence of us tracing each directory in etcd during our initial decoding. This spams the journald process, and results in much of those messages getting rate limited and dropped, and making actual problems much harder to debug. This stops us logging discrete keys in favor of just knowing that an operation was slow for the particular object type (which is enough information!)
this PR also makes it so we know what trace operation was ended in our log messages!
sample log now:
```
Trace "decodeNodeList *[]api.PolicyBinding" (started 2017-02-20 22:20:33.98771112 +0000 UTC)
Decoded 10718 nodes
decodeNodeList *[]api.PolicyBinding [500ms] END
```
Automatic merge from submit-queue (batch tested with PRs 41597, 42185, 42075, 42178, 41705)
auto discovery CA for extension API servers
This is what the smaller pulls were leading to. Only the last commit is unique and I expect I'll still tweak some pod definitions, but this is where I was going.
@sttts @liggitt
