Automatic merge from submit-queue (batch tested with PRs 45573, 46354, 46376, 46162, 46366)
GCE - Retrieve subnetwork name/url from gce.conf
**What this PR does / why we need it**:
Features like ILB require specifying the subnetwork if the network is type manual.
**Notes:**
The network URL can be [constructed](68e7e18698/pkg/cloudprovider/providers/gce/gce.go (L211-L217)) by fetching instance metadata; however, the subnetwork is not provided through this feature. Users must specify the subnetwork name/url through the gce.conf.
Although multiple subnets can exist in the same region for a network, the cloud provider will only use one subnet url for creating LBs.
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 45573, 46354, 46376, 46162, 46366)
Subresources are not included in apiserver prometheus metrics
Subresources are very often completely different code paths and errors
generated on those code paths are important to distinguish.
@kubernetes/sig-api-machinery-pr-reviews
```release-note
The Prometheus metrics for the kube-apiserver for tracking incoming API requests and latencies now return the `subresource` label for correctly attributing the type of API call.
```
Automatic merge from submit-queue (batch tested with PRs 45913, 46065, 46352, 46363, 46373)
Fix CheckPodsCondition to print out the correct podName
From a couple CIs (https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/ci-kubernetes-e2e-gce-serial/1114, https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/ci-kubernetes-e2e-gce-gci-qa-serial-master/2246, https://storage.googleapis.com/kubernetes-jenkins/logs/ci-kubernetes-e2e-gci-gke-pre-release/2187), all indicate we print out the wrong pod name in CheckPodsCondition for _"Pod XXX failed to be running and ready, or succeeded."_:
```
I0524 02:09:50.173] May 24 02:09:50.173: INFO: Waiting for pod heapster-v1.3.0-3806988011-kzkg6 in namespace 'kube-system' status to be 'running and ready, or succeeded'(found phase: "Running", readiness: false) (4m55.033881993s elapsed)
I0524 02:09:52.178] May 24 02:09:52.178: INFO: Waiting for pod heapster-v1.3.0-3806988011-kzkg6 in namespace 'kube-system' status to be 'running and ready, or succeeded'(found phase: "Running", readiness: false) (4m57.03848264s elapsed)
I0524 02:09:54.183] May 24 02:09:54.182: INFO: Waiting for pod heapster-v1.3.0-3806988011-kzkg6 in namespace 'kube-system' status to be 'running and ready, or succeeded'(found phase: "Running", readiness: false) (4m59.043463323s elapsed)
I0524 02:09:56.183] May 24 02:09:56.183: INFO: Pod fluentd-gcp-v2.0-6wf67 failed to be running and ready, or succeeded.
I0524 02:09:56.184] May 24 02:09:56.183: INFO: Wanted all 23 pods to be running and ready, or succeeded. Result: false. Pods: [heapster-v1.3.0-3806988011-kzkg6 kube-proxy-bootstrap-e2e-minion-group-bbwn rescheduler-v0.3.0-bootstrap-e2e-master monitoring-influxdb-grafana-v4-1q59k l7-default-backend-1044750973-zgxsc etcd-server-events-bootstrap-e2e-master kube-apiserver-bootstrap-e2e-master kube-proxy-bootstrap-e2e-minion-group-6nqb kube-proxy-bootstrap-e2e-minion-group-mzbz fluentd-gcp-v2.0-chd2x kube-dns-806549836-f8p46 fluentd-gcp-v2.0-44x97 kube-dns-autoscaler-2528518105-vlg8t fluentd-gcp-v2.0-p1h4b kube-controller-manager-bootstrap-e2e-master l7-lb-controller-v0.9.3-bootstrap-e2e-master kubernetes-dashboard-2917854236-tn3nx kube-dns-806549836-fq2fp kube-scheduler-bootstrap-e2e-master etcd-empty-dir-cleanup-bootstrap-e2e-master kube-addon-manager-bootstrap-e2e-master etcd-server-bootstrap-e2e-master fluentd-gcp-v2.0-6wf67]
I0524 02:09:56.184] May 24 02:09:56.183: INFO: At least one pod wasn't running and ready or succeeded at test start.
I0524 02:09:56.184] [AfterEach] [k8s.io] Restart [Disruptive]
```
Check the codes and found we always print out the last pod name, which is random. Pass the pod name into channel to fix.
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 46149, 45897, 46293, 46296, 46194)
Chaosmonkey - Signal stop to tests and wait for done when disruption fails
**What this PR does / why we need it**:
Prevents tests from leaking resources because their Teardown was never called when test disruption fails.
**Which issue this PR fixes**
First problem of #45842
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue
Enable "kick the tires" support for Nvidia GPUs in COS
This PR provides an installation daemonset that will install Nvidia CUDA drivers on Google Container Optimized OS (COS).
User space libraries and debug utilities from the Nvidia driver installation are made available on the host in a special directory on the host -
* `/home/kubernetes/bin/nvidia/lib` for libraries
* `/home/kubernetes/bin/nvidia/bin` for debug utilities
Containers that run CUDA applications on COS are expected to consume the libraries and debug utilities (if necessary) from the host directories using `HostPath` volumes.
Note: This solution requires updating Pod Spec across distros. This is a known issue and will be addressed in the future. Until then CUDA workloads will not be portable.
This PR updates the COS base image version to m59. This is coupled with this PR for the following reasons:
1. Driver installation requires disabling a kernel feature in COS.
2. The kernel API for disabling this interface changed across COS versions
3. If the COS image update is not handled in this PR, then a subsequent COS image update will break GPU integration and will require an update to the installation scripts in this PR.
4. Instead of having to post `3` PRs, one each for adding the basic installer, updating COS to m59, and then updating the installer again, this PR combines all the changes to reduce review overhead and latency, and additional noise that will be created when GPU tests break.
**Try out this PR**
1. Get Quota for GPUs in any region
2. `export `KUBE_GCE_ZONE=<zone-with-gpus>` KUBE_NODE_OS_DISTRIBUTION=gci`
3. `NODE_ACCELERATORS="type=nvidia-tesla-k80,count=1" cluster/kube-up.sh`
4. `kubectl create -f cluster/gce/gci/nvidia-gpus/cos-installer-daemonset.yaml`
5. Run your CUDA app in a pod.
**Another option is to run a e2e manually to try out this PR**
1. Get Quota for GPUs in any region
2. export `KUBE_GCE_ZONE=<zone-with-gpus>` KUBE_NODE_OS_DISTRIBUTION=gci
3. `NODE_ACCELERATORS="type=nvidia-tesla-k80,count=1"`
4. `go run hack/e2e.go -- --up`
5. `hack/ginkgo-e2e.sh --ginkgo.focus="\[Feature:GPU\]"`
The e2e will install the drivers automatically using the daemonset and then run test workloads to validate driver integration.
TODO:
- [x] Update COS image version to m59 release.
- [x] Remove sleep from the install script and add it to the daemonset
- [x] Add an e2e that will run the daemonset and run a sample CUDA app on COS clusters.
- [x] Setup a test project with necessary quota to run GPU tests against HEAD to start with https://github.com/kubernetes/test-infra/pull/2759
- [x] Update node e2e serial configs to install nvidia drivers on COS by default
Automatic merge from submit-queue (batch tested with PRs 38990, 45781, 46225, 44899, 43663)
Support parallel scaling on StatefulSets
Fixes#41255
```release-note
StatefulSets now include an alpha scaling feature accessible by setting the `spec.podManagementPolicy` field to `Parallel`. The controller will not wait for pods to be ready before adding the other pods, and will replace deleted pods as needed. Since parallel scaling creates pods out of order, you cannot depend on predictable membership changes within your set.
```
Packaged the script as a docker container stored in gcr.io/google-containers
A daemonset deployment is included to make it easy to consume the installer
A cluster e2e has been added to test the installation daemonset along with verifying installation
by using a sample CUDA application.
Node e2e for GPUs updated to avoid running on nodes without GPU devices.
Signed-off-by: Vishnu kannan <vishnuk@google.com>
Automatic merge from submit-queue (batch tested with PRs 46033, 46122, 46053, 46018, 45981)
Log age of stats used for evictions during eviction tests
I recently added prometheus metrics for the age of the metrics used for evictions #43031. It would be nice to surface these during eviction tests, so I can better assess how old stats are, and whether or not the age of stats causes extra evictions.
This isnt super-high priority, and can be done after code-freeze, since it is a testing improvement. Feel free to take a look whenever either of you has time.
/assign @mtaufen
/assign @Random-Liu
Automatic merge from submit-queue (batch tested with PRs 45996, 46121, 45707, 46011, 45564)
Fix waitForNPods in restart.go
From https://github.com/kubernetes/kubernetes/issues/45991#issuecomment-302292404.
Don't redefine `pods` so we can return real pod names instead of empty array.
/assign @dchen1107 @bowei
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 45990, 45544, 45745, 45742, 45678)
Add explicit image tag to cockroachdb example and test
@gyliu513
```release-note
NONE
```
Module remotecommand originally part of kubernetes/pkg/client/unversioned was moved
to client-go/tools, and will be used as authoritative in kubectl, e2e and other places.
Module remotecommand relies on util/exec module which will be copied to client-go/pkg/util
Automatic merge from submit-queue
Reorganize kubelet tree so apis can be independently versioned
@yujuhong @lavalamp @thockin @bgrant0607
This is an example of how we might reorganize `pkg/kubelet` so the apis it exposes can be independently versioned. This would also provide a logical place to put the `KubeletConfiguration` type, which currently lives in `pkg/apis/componentconfig`; it could live in e.g. `pkg/kubelet/apis/config` instead.
Take a look when you have a chance and let me know what you think. The most significant change in this PR is reorganizing `pkg/kubelet/api` to `pkg/kubelet/apis`, the rest is pretty much updating import paths and `BUILD` files.
Automatic merge from submit-queue (batch tested with PRs 45623, 45241, 45460, 41162)
Promotes Source IP preservation for Virtual IPs from Beta to GA
Fixes#33625. Feature issue: kubernetes/features#27.
Bullet points:
- Declare 2 fields (ExternalTraffic and HealthCheckNodePort) that mirror the ESIPP annotations.
- ESIPP alpha annotations will be ignored.
- Existing ESIPP beta annotations will still be fully supported.
- Allow promoting beta annotations to first class fields or reversely.
- Disallow setting invalid ExternalTraffic and HealthCheckNodePort on services. Default ExternalTraffic field for nodePort or loadBalancer type service to "Global" if not set.
**Release note**:
```release-note
Promotes Source IP preservation for Virtual IPs to GA.
Two api fields are defined correspondingly:
- Service.Spec.ExternalTrafficPolicy <- 'service.beta.kubernetes.io/external-traffic' annotation.
- Service.Spec.HealthCheckNodePort <- 'service.beta.kubernetes.io/healthcheck-nodeport' annotation.
```
Automatic merge from submit-queue (batch tested with PRs 45623, 45241, 45460, 41162)
Revert #45101 Mark PersistentVolumes as [Feature:Volumes]
**What this PR does / why we need it**:
Reverts #45101
`Feature` tag should only be used when a test/suite has dependencies not met by core CI. That is not the case for NFS backed PV tests.
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 45685, 45572, 45624, 45723, 45733)
Remove a test utility function that is redundant and kinda broken
Framework.WaitForAnEndpoint() has no timeout, so if something goes wrong and the endpoint doesn't get created, the test will hang forever. (This is happening for some reason sometimes in OpenShift right now, and when the CI system eventually times out and kills the VM, it loses the logs that would explain what failed.)
There's already another nearly-identical WaitForEndpoint() method that *does* take a timeout, so people can just use that instead.
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 45653, 45719, 45729, 45730, 44250)
Print pod startup latency metric as perfdata
Follows #45657
This should print pod startup latency in same format as api calls latencies.
cc @wojtek-t @gmarek
Automatic merge from submit-queue
move from daemon_restart.go to framework/util.go
**What this PR does / why we need it**:
Moves the func `nodeExec` from daemon_restart.go to framework/util.go. This is the correct file for this func and is a more intuitive pkg for other callers to use. This is a small step of the larger effort of restructuring e2e tests to be more logically structured and easier for newcomers to understand.
```release-note
NONE
```
cc @timothysc @copejon
