note that we don't change the behaviour of kubectl.
For example it won't scale new resources. That's the end goal.
The first step is to retrofit existing code to use the generic scaler.
When running golint on test/e2e/e2e.go, the following erros
were faced:
$ golint e2e.go
e2e.go:329:2: var metricsJson should be metricsJSON
e2e.go:342:1: comment on exported function RunE2ETests should be of the
form "RunE2ETests ..."
This PR fixes them.
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Allow Ceph server some time to start
Ceph server needs to create our "foo" volume on startup. It keeps the image small, however it makes the server container start slow.
Add sleep before the server is usable. Without this PR, all pods that use Ceph fail to start for couple of seconds with cryptic "image foo not found" error and it clutters logs and pod logs and makes it harder to spot real errors.
```release-note
NONE
```
/sig storage
/sig testing
/assign @jeffvance @copejon
Automatic merge from submit-queue (batch tested with PRs 58025, 57112, 57879, 57571, 58062). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Let mutating webhook defaults the object
...after applying the patch sent back by the webhook
This should be treated as a bug and cherrypicked to 1.9.
Fixes#57982
Automatic merge from submit-queue (batch tested with PRs 57733, 57613, 57953). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Explicit kubelet flags
The Kubelet was using the global flag set.
Libraries also often use the global flag set.
There are flags in the Kubelet's usage statement for which the Kubelet does not control registration.
This is bad, we must have full control of our command-line API.
This PR puts a stop to it.
I worked through the Kubelet's usage printout and tracked down the flags.
In the below list, flags with checkboxes are one of:
- explicitly **registered** in this PR
- thought about and outright **rejected**
- thought about and **registered-deprecated** for legacy reasons
For the others we need to figure out whether they should be **registered** or **rejected** or **registered-deprecated**.
### cadvisor:
grep commands, for reference:
```
# flag definitions in cadvisor
git grep -E "\"application_metrics_count_limit\"|\"boot_id_file\"|\"container_hints\"|\"containerd\"|\"docker\"|\"docker_env_metadata_whitelist\"|\"docker_only\"|\"docker_root\"|\"docker-tls\"|\"docker-tls-ca\"|\"docker-tls-cert\"|\"docker-tls-key\"|\"enable_load_reader\"|\"event_storage_age_limit\"|\"event_storage_event_limit\"|\"global_housekeeping_interval\"|\"housekeeping_interval\"|\"log_cadvisor_usage\"|\"machine_id_file\"|\"storage_driver_buffer_duration\"|\"storage_driver_db\"|\"storage_driver_host\"|\"storage_driver_password\"|\"storage_driver_secure\"|\"storage_driver_table\"|\"storage_driver_user\"" -- vendor/github.com/google/cadvisor
# flag invocations
git grep -E "\--application_metrics_count_limit|--boot_id_file|--container_hints|--containerd|--docker|--docker_env_metadata_whitelist|--docker_only|--docker_root|--docker_tls|--docker_tls_ca|--docker_tls_cert|--docker_tls_key|--enable_load_reader|--event_storage_age_limit|--event_storage_event_limit|--global_housekeeping_interval|--housekeeping_interval|--log_cadvisor_usage|--machine_id_file|--storage_driver_buffer_duration|--storage_driver_db|--storage_driver_host|--storage_driver_password|--storage_driver_secure|--storage_driver_table|--storage_driver_user" -- ':(exclude)pkg/generated/bindata.go' ':(exclude)Godeps' ':(exclude)CHANGELOG*' ':(exclude)vendor'
# normalized flag invocations
git grep -E "\--application-metrics-count-limit|--boot-id-file|--container-hints|--containerd|--docker|--docker-env-metadata-whitelist|--docker-only|--docker-root|--docker-tls|--docker-tls-ca|--docker-tls-cert|--docker-tls-key|--enable-load-reader|--event-storage-age-limit|--event-storage-event-limit|--global-housekeeping-interval|--housekeeping-interval|--log-cadvisor-usage|--machine-id-file|--storage-driver-buffer-duration|--storage-driver-db|--storage-driver-host|--storage-driver-password|--storage-driver-secure|--storage-driver-table|--storage-driver-user" -- ':(exclude)pkg/generated/bindata.go' ':(exclude)Godeps' ':(exclude)CHANGELOG*' ':(exclude)vendor'
# flag names
git grep -E "application_metrics_count_limit|boot_id_file|container_hints|containerd|docker|docker_env_metadata_whitelist|docker_only|docker_root|docker_tls|docker_tls_ca|docker_tls_cert|docker_tls_key|enable_load_reader|event_storage_age_limit|event_storage_event_limit|global_housekeeping_interval|housekeeping_interval|log_cadvisor_usage|machine_id_file|storage_driver_buffer_duration|storage_driver_db|storage_driver_host|storage_driver_password|storage_driver_secure|storage_driver_table|storage_driver_user" ':(exclude)pkg/generated/bindata.go' ':(exclude)Godeps' ':(exclude)CHANGELOG*' ':(exclude)vendor'
# normalized flag names
git grep -E "application-metrics-count-limit|boot-id-file|container-hints|containerd|docker|docker-env-metadata-whitelist|docker-only|docker-root|docker-tls|docker-tls-ca|docker-tls-cert|docker-tls-key|enable-load-reader|event-storage-age-limit|event-storage-event-limit|global-housekeeping-interval|housekeeping-interval|log-cadvisor-usage|machine-id-file|storage-driver-buffer-duration|storage-driver-db|storage-driver-host|storage-driver-password|storage-driver-secure|storage-driver-table|storage-driver-user" ':(exclude)pkg/generated/bindata.go' ':(exclude)Godeps' ':(exclude)CHANGELOG*' ':(exclude)vendor'
# normalized and underscore names combined
git grep -E "application_metrics_count_limit|boot_id_file|container_hints|containerd|docker|docker_env_metadata_whitelist|docker_only|docker_root|docker_tls|docker_tls_ca|docker_tls_cert|docker_tls_key|enable_load_reader|event_storage_age_limit|event_storage_event_limit|global_housekeeping_interval|housekeeping_interval|log_cadvisor_usage|machine_id_file|storage_driver_buffer_duration|storage_driver_db|storage_driver_host|storage_driver_password|storage_driver_secure|storage_driver_table|storage_driver_user|application-metrics-count-limit|boot-id-file|container-hints|containerd|docker-env-metadata-whitelist|docker-only|docker-root|docker-tls|docker-tls-ca|docker-tls-cert|docker-tls-key|enable-load-reader|event-storage-age-limit|event-storage-event-limit|global-housekeeping-interval|housekeeping-interval|log-cadvisor-usage|machine-id-file|storage-driver-buffer-duration|storage-driver-db|storage-driver-host|storage-driver-password|storage-driver-secure|storage-driver-table|storage-driver-user" ':(exclude)pkg/generated/bindata.go' ':(exclude)Godeps' ':(exclude)CHANGELOG*' ':(exclude)vendor'
```
- [x] --docker-root (**registered** - this is used in `cluster/saltbase/salt/kubelet/default`)
- [x] --housekeeping-interval (**registered** - e2e node tests rely on this (`test/e2e_node/resource_collector.go`))
- [x] --application-metrics-count-limit (**registered-deprecated** - temporarily register for legacy)
- [x] --boot-id-file (**registered-deprecated** - temporarily register for legacy)
- [x] --container-hints (**registered-deprecated** - temporarily register for legacy)
- [x] --containerd (**registered-deprecated** - temporarily register for legacy)
- [x] --docker (**registered-deprecated** - temporarily register for legacy)
- [x] --docker-env-metadata-whitelist (**registered-deprecated** - temporarily register for legacy)
- [x] --docker-only (**registered-deprecated** - temporarily register for legacy)
- [x] --docker-tls (**registered-deprecated** - temporarily register for legacy)
- [x] --docker-tls-ca (**registered-deprecated** - temporarily register for legacy)
- [x] --docker-tls-cert (**registered-deprecated** - temporarily register for legacy)
- [x] --docker-tls-key (**registered-deprecated** - temporarily register for legacy)
- [x] --enable-load-reader (**registered-deprecated** - temporarily register for legacy)
- [x] --event-storage-age-limit (**registered-deprecated** - the Kubelet overrides the default via the global flagset (`pkg/kubelet/cadvisor/cadvisor_linux.go`), but nothing else in core repo provides)
- [x] --event-storage-event-limit (**registered-deprecated** - the Kubelet overrides the default via the global flagset (`pkg/kubelet/cadvisor/cadvisor_linux.go`), but nothing else in core repo provides)
- [x] --global-housekeeping-interval (**registered-deprecated** - temporarily register for legacy)
- [x] --log-cadvisor-usage (**registered-deprecated** - temporarily register for legacy)
- [x] --machine-id-file (**registered-deprecated** - temporarily register for legacy)
- [x] --storage-driver-user (**registered-deprecated** - temporarily register for legacy)
- [x] --storage-driver-password (**registered-deprecated** - temporarily register for legacy)
- [x] --storage-driver-host (**registered-deprecated** - temporarily register for legacy)
- [x] --storage-driver-db (**registered-deprecated** - temporarily register for legacy)
- [x] --storage-driver-table (**registered-deprecated** - temporarily register for legacy)
- [x] --storage-driver-secure (**registered-deprecated** - temporarily register for legacy)
- [x] --storage-driver-buffer-duration (**registered-deprecated** - temporarily register for legacy)
### pkg/apiserver/util/logs:
- [x] --log-flush-frequency (**registered** - like the glog flags, this is probably useful)
### pkg/credentialprovider/azure/azure_credentials.go:
- [x] --azure-container-registry-config (**registered** - This isn't quite as straightforward as --google-json-key, because the file it points to isn't static. For now we will just register, and we will deprecate it when there is an alternative. See below comments.)
### pkg/credentialprovider/gcp/jwt.go:
- [x] --google-json-key (**registered-deprecated** - This is really old legacy stuff to allow kubelets to authenticate with gcr (see: d5e0054eb0). See @liggit's below comment for what should be used instead.)
### pkg/cloudprovider/providers/gce/gce_loadbalancer.go:
- [x] --cloud-provider-gce-lb-src-cidrs (**rejected** - Kubelet doesn't need to know about the cidrs that were opened in the firewall for the load balancer)
### glog:
I registered all of these, since this logging library is used pretty much everywhere in the Kubelet, and all of its toggles are probably useful.
- [x] --logtostderr (**registered**)
- [x] --alsologtostderr (**registered**)
- [x] -v, --v (**registered**)
- [x] --stderrthreshold (**registered**)
- [x] --vmodule (**registered**)
- [x] --log-backtrace-at (**registered**)
- [x] --log-dir (**registered**)
### verflag:
This is how you get the Kubelet's version, absolutely necessary to register this.
- [x] --version (**registered**)
```release-note
The Kubelet now explicitly registers all of its command-line flags with an internal flagset, which prevents flags from third party libraries from unintentionally leaking into the Kubelet's command-line API. Many unintentionally leaked flags are now marked deprecated, so that users have a chance to migrate away from them before they are removed. One previously leaked flag, --cloud-provider-gce-lb-src-cidrs, was entirely removed from the Kubelet's command-line API, because it is irrelevant to Kubelet operation.
```
Thanks to @liggitt for being surprised that we didn't already do this.
Thanks to @dashpole for realizing cadvisor flags were leaked after seeing #55863.
Thanks to @tallclair who recognized this problem a long time ago in #19432.
This explicitly registers Kubelet flags from libraries that were
registering flags globally, and stops parsing the global flag set.
In general, we should always be explicit about flags we register
and parse, so that we maintain control over our command-line API.
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Move scheduler out of plugin directory
**What this PR does / why we need it**:
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
This is but one step toward resolving the referenced issue.
/ref #57579
**Special notes for your reviewer**:
**Release note**:
```release-note
Default scheduler code is moved out of the plugin directory.
plugin/pkg/scheduler -> pkg/scheduler
plugin/cmd/kube-scheduler -> cmd/kube-scheduler
```
/sig scheduling
Automatic merge from submit-queue (batch tested with PRs 57906, 57425, 56939, 57317, 57762). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Fix local e2e test with changed error message
**What this PR does / why we need it**:
Node affinity error message got changed as part of log message refactoring. Updating e2e test.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#57662
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
test: e2e: support NFS test on overlayfs
TL;DR; NFS is not supported over overlayfs. In CRI-O we default to
overlayfs as the storage driver, therefore running the NFS test always
fails to export the directory. This patch adds a volume mount from the
host so /exports isn't on overlayfs anymore.
Signed-off-by: Antonio Murdaca <runcom@redhat.com>
@Random-Liu @mrunalp PTAL - we got this after enabling all the e2e tests as discussed at kubecon (thanks!)
```release-note
NONE
```
This moves plugin/pkg/scheduler to pkg/scheduler and
plugin/cmd/kube-scheduler to cmd/kube-scheduler.
Bulk of the work was done with gomvpkg, except for kube-scheduler main
package.
Prepatory work fpr removing cloud provider dependency from node
controller running in Kube Controller Manager. Splitting the node
controller into its two major pieces life-cycle and CIDR/IP
management. Both pieces currently need the the cloud system to do their work.
Removing lifecycles dependency on cloud will be fixed ina followup PR.
Moved node scheduler code to live with node lifecycle controller.
Got the IPAM/Lifecycle split completed. Still need to rename pieces.
Made changes to the utils and tests so they would be in the appropriate
package.
Moved the node based ipam code to nodeipam.
Made the relevant tests pass.
Moved common node controller util code to nodeutil.
Removed unneeded pod informer sync from node ipam controller.
Fixed linter issues.
Factored in feedback from @gmarek.
Factored in feedback from @mtaufen.
Undoing unneeded change.
Automatic merge from submit-queue (batch tested with PRs 56382, 57549). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
e2e: Only create PSP if RBAC is enabled
**What this PR does / why we need it**: Creating privileged PSPs during e2e tests depends on RBAC being enabled in the target cluster. However it is possible to use PSPs without having RBAC enabled thus only enable creating PSPs during e2e tests if both PSP and RBAC is enabled.
Fix#57840
**Special notes for your reviewer**:
**Release note**:
```release-note
Only create Privileged PSP binding during e2e tests if RBAC is enabled.
```
Automatic merge from submit-queue (batch tested with PRs 49856, 56257, 57027, 57695, 57432). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Update multizone pd tests to only allow gce and gke
**What this PR does / why we need it**:
Restricts multizone pd tests to only gce and gke environments.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
e2e: test containers projected volume updates should not exit
The mounttest container should be running until the test determines to
give up (i.e., time out) and kill it. It should not exit prematurely by
itself. Bump the `--retry-timeout` to a much higher value.
The local-volume bootstrapper/provisioner configuration format changed in
https://github.com/kubernetes-incubator/external-storage/pull/352
This format is exposed in v2.0.0 of the provisioner and boostratpper images.
This PR updates the e2e test config, so that the existing tests continue
to work.
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Add custom metrics e2e test with two metrics.
**What this PR does / why we need it**:
Adds an e2e test for HPA, where a deployment is scaled based on two custom metrics.
**Release note**:
```
NONE
```
TL;DR; NFS is not supported over overlayfs. In CRI-O we default to
overlayfs as the storage driver, therefore running the NFS test always
fails to export the directory. This patch adds a volume mount from the
host so /exports isn't on overlayfs anymore.
Signed-off-by: Antonio Murdaca <runcom@redhat.com>
Automatic merge from submit-queue (batch tested with PRs 56600, 56814). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Fix volume size allocation in gcd_pd
**What this PR does / why we need it**:
GCE PDs are allocated in chunks of GBs not GiB but CreateVolume function incorrectly creates volume in chunks of GiB.
1 GiB = 1024 * 1024 * 1024 Bytes
1 GB = 1000 * 1000 * 1000 Bytes
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#56081
**Special notes for your reviewer**:
**Release note**:
```release-note
Fixed dynamic provisioning of GCE PDs to round to the next GB instead of GiB
```
utils/image/manifest has an additional `arch` parameter, which determines
whether an image ends in `-$ARCH` (like `-amd64`).
All locations that previously had gcr.io urls referenced in costants or inline
have been updated to refere test/utils/image.
This e2e test tests the CSI volume plugin in kubernetes with
a CSI hostPath driver. It is also setup to be able to be
tested with more drivers in the future.
Automatic merge from submit-queue (batch tested with PRs 56108, 56811, 57335, 57331, 54530). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
E2E autoscaling test scenario "should add node to the particular mig" fix
Resize targeted MIG to one to fix flakes when initial size of it is zero. Scale-up from zero doesn't make sense in this scenario, and the test assumes there will be always at least one node.
Automatic merge from submit-queue (batch tested with PRs 56108, 56811, 57335, 57331, 54530). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Fix Stackdriver Logging e2e tests
This PR fixes the problems in the SD Logging tests:
- Because of Docker problem, trimming works unpredictably on some OSs, so instead of assuming exact trimming mechanism, just assume trimming
- Increases the throughput of reading logs through Pub/Sub by increasing the parallelism level
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 54379, 56593, 56685, 54174, 57309). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Use k8s.gcr.io vanity domain for container images
Related issue: https://github.com/kubernetes/release/issues/281
```release-note
Use "k8s.gcr.io" for container images rather than "gcr.io/google_containers". This is just a redirect, for now, so should not impact anyone materially.
Documentation and tools should all convert to the new name. Users should take note of this in case they see this new name in the system.
```
