- Instead of using cm.capacity field to communicate device plugin resource
capacity, this PR changes to use an explicit cm.GetDevicePluginResourceCapacity()
function that returns device plugin resource capacity as well as any inactive
device plugin resource. Kubelet syncNodeStatus call this function during its
periodic run to update node status capacity and allocatable. After this call,
device plugin can remove the inactive device plugin resource from its allDevices
field as the update is already pushed to API server.
- Extends device plugin checkpoint data to record registered resources
so that we can finish resource removing even upon kubelet restarts.
- Passes sourcesReady from kubelet to device plugin to avoid removing
inactive pods during grace period of kubelet restart.
Automatic merge from submit-queue (batch tested with PRs 55112, 56029, 55740, 56095, 55845). 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>.
Updating vsphere cloud provider to support k8s cluster spread across multiple vCenters
**What this PR does / why we need it**:
vSphere cloud provider in Kubernetes 1.8 was designed to work only if all the nodes of the cluster are in one single datacenter folder. This is a hard restriction that makes the cluster not span across different folders/datacenter/vCenters. Users have use-cases to span the cluster across datacenters/vCenters.
**Which issue(s) this PR fixes**
Fixes # https://github.com/vmware/kubernetes/issues/255
**Special notes for your reviewer**:
This is a change purely in vsphere cloud provider and no changes in kubernetes core are needed.
**Release note**:
```release-note
With this change
- User should be able to create k8s cluster which spans across multiple ESXi clusters, datacenters or even vCenters.
- vSphere cloud provider (VCP) uses OS hostname and not vSphere Inventory VM Name.
That means, now VCP can handle cases where user changes VM inventory name.
- VCP can handle cases where VM migrates to other ESXi cluster or datacenter or vCenter.
The only requirement is the shared storage. VCP needs shared storage on all Node VMs.
```
Internally tested and reviewed the code.
@tthole, @shaominchen, @abrarshivani
Automatic merge from submit-queue (batch tested with PRs 55112, 56029, 55740, 56095, 55845). 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>.
[Azure] Load Balancer general fixes
**What this PR does / why we need it**:
Fixes Azure Load Balancer functions, documented in #55627
1. Resolve leaking Azure resource issue
2. Enable Azure LB creation for every availability set
3. Auto assign service to LB based on service annotation
4. General improvement in testing
5. Consolidate NSG rules
Change made by @rjtsdl, @rite2nikhil, and @itowlson
**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#55627#54961
**Special notes for your reviewer**:
Ready to review
**Release note**:
```release-note
NONE
```
/sig azure
Automatic merge from submit-queue (batch tested with PRs 54824, 55911, 55730, 55979, 55961). 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 kubeletconfig round trip test
I noticed we were missing one of these.
fixes#55959
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 54824, 55911, 55730, 55979, 55961). 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>.
Device plugin API merge of handler and manager
**What this PR does / why we need it**: We are trying different approaches to make the device plugin implementation simpler and more robust. One option is to merge the notion of the `device_plugin_handler` into the `device_manager`. This is for several reasons:
1) Some calls go directly from handler to manager, adding little value.
2) The separation of concern is not clear between the two components. They have a 1:1 relationship.
3) The separation and abstractions needed are at a different level. Code that can be refactored will most likely live in abstractions which hide details around lock acquisition and check pointing.
In this PR, we will **just** merge the two interfaces. After this, there is several opportunities for simplifying and cleaning up the device plugin.
Fixes#55180
**Special notes for your reviewer**: This is a WIP. May very well get dropped, but keeping up for the sake of early sharing and showing the progress of the code move.
**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>.
fix cadvisor.New signature for cross build
**What this PR does / why we need it**: fixes the `pkg/kubelet/cadvisor.New` signature on non-linux platforms to match the new one on linux. This should fix the cross build
**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#56002
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
/sig release
Automatic merge from submit-queue (batch tested with PRs 55839, 54495, 55884, 55983, 56069). 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>.
seccomp is an alpha feature and not feature gated
Move SeccompProfileRoot to KubeletFlags and document flag as alpha.
wrt https://github.com/kubernetes/kubernetes/pull/53833#issuecomment-345396575, seccomp is an alpha feature, but this isn't clearly documented anywhere (the annotation just has the word "alpha" in it, and that's your signal that it's alpha).
Since seccomp was around before feature gates, it doesn't have one.
Thus SeccompProfileRoot should not be part of KubeletConfiguration, and this PR moves it to KubeletFlags, and amends the help text to note the alpha state of the feature.
fixes: #56087
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 55839, 54495, 55884, 55983, 56069). 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>.
deviceplugin: fix race when multiple plugins are registered
**What this PR does / why we need it**:
When registering multiple device plugins to Kubelet concurrently, there exists a race that crashes the Kubelet.
Consider two plugins: D1 and D2. The call order method is roughly
D1 -> manager.go:register -> endpoint.go:listAndWatch -> device_plugin_handler.go:(*D1).callback
D2 -> manager.go:register -> endpoint.go:listAndWatch -> device_plugin_handler.go:(*D2).callback
The callback function accesses HandlerImpl's allDevices map that maps (resourceName -> DeviceID). If both plugins reach these accesses at the same time, Kubelet crashes with "fatal error: concurrent map read and map write".
This can be solved by making sure handler is locked when allDevices are being updated. The functionality is needed to avoid Kubelet crashes when multiple device plugins are trying to register with Kubelet at the same moment. Occurs frequently when single binary tries to register itself as multiple plugins.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
**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>.
Stop using VersionedObject in resource.Builder, use unstructured
Remove the need for kubectl callers to distinguish between unstructured and versioned / type aware builders. The factory should create a single builder than can be set to return unstructured objects. Callers can then use one of the new helpers on `resource.Info` to convert the objects into the desired form - `Internal()` for printers, `Typed()` for external versions, and `Unstructured()` to ensure the object is in the right state. Leverages the new scheme support for unstructured conversion so that higher level callers can perform best effort conversion (get typed versions if you have them, otherwise use default behavior).
`get.go` demonstrates this by removing the previous logic that depended on the underlying scheme. Other commands are updated to be consistent.
Includes #55650 and #55647.
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 of unreachable taint key out of alpha
**What this PR does / why we need it**:
Move of unreachable taint key out of alpha, which already happened in community doc.
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes#54198
**Special notes for your reviewer**:
Please see #54198 for the context of this inconsistency.
**Release note**:
```release-note
Move unreachable taint key out of alpha.
Please note the existing pods with the alpha toleration should be updated by user himself to tolerate the GA taint.
```
Automatic merge from submit-queue (batch tested with PRs 55974, 54701). 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>.
Refactor `reconcileAutoscaler` method in hpa
**What this PR does / why we need it**:
There have been a couple of recent bugs in the "normalizing" part of the
`reconcileAutoscaler` method. This part of the code base is responsible
for, among other things, taking the suggested desired replicas based on
the metrics, ensuring it conforms to certain conditions, and updating it
if it does not. Isolate the part that converts the desired replicas
based on a given set of rules into its own function.
We are refactoring this part of the code base to make the logic simpler
and to make it easier to write unit tests.
**Which issue this PR fixes**:
Fixes#53728
**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>.
Allow constructing spdy executor from existing transports
If you already have an existing transport, it is not always possible to reconstruct a client config from it.
Allow constructing a spdy executor, given a connection/upgrade transport
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 55963, 55790, 55670, 55931). 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 GRS, RAGRS storage account type support for azure disk
**What this PR does / why we need it**:
add Standard_GRS, Standard_RAGRS storage account type support for azure disk, and Standard_ZRS, Premium_GRS are not supported for azure disk now
**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#55774
**Special notes for your reviewer**:
**Release note**:
```
add Standard_GRS, Standard_RAGRS support for azure disk
```
/sig azure
@rootfs @brendanburns
Automatic merge from submit-queue (batch tested with PRs 55963, 55790, 55670, 55931). 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>.
include error message for volumeManager during attach and mount
**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)*:
Fixes#55609
**Special notes for your reviewer**:
/assign @jsafrane @jingxu97
**Release note**:
```release-note
None
```
Automatic merge from submit-queue (batch tested with PRs 55963, 55790, 55670, 55931). 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>.
admission/webhook: move webhook initializer into plugin
Follow-up of https://github.com/kubernetes/kubernetes/pull/55132.
Non-generic plugin intitializers should go into the admission plugin itself. No need leak that into the generic apiserver. Keeping them contained in the webhook package makes it easier (it was already possible) for extension apiservers to provide one, but we don't need to leak it into the generic plugin initializer.
Automatic merge from submit-queue (batch tested with PRs 55217, 54260). 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>.
Unit tests for Azure service session affinity
**What this PR does / why we need it**: We added session affinity support in the Azure load balancer in commit 8b50b83067. This PR adds unit tests for this behaviour.
**Which issue this PR fixes**: None
**Special notes for your reviewer**: None
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 55217, 54260). 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>.
Checking Cinder disk when tickers are delivered
Cinder's WaitForAttach() runs probeAttachedVolume() 2 times per
second(L243 and L247), that triggers SCSI rescan and generates many udev "change"
events.
**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>.
Fix representation of the pv's capacity which provisioned by glusterfs
**What this PR does / why we need it**:
This PR fixes representation of the pv's capacity which provisioned by glusterfs.
Gluster's volume size is calculated in GB, and than this value is setted as GiB for pv's storage capacity.
**Which issue(s) this PR fixes**:
Fixes#55937
**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>.
Using ipset doing SNAT and packet filter in IPVS kube-proxy
**What this PR does / why we need it**:
Try ipset in ipvs proxy mode.
**Which issue this PR fixes**:
fixes#54203
xref: #53393, #53775
**Special notes for your reviewer**:
**Release note**:
```release-note
Using ipset doing SNAT and packet filtering in IPVS kube-proxy
```
/sig network
/area kube-proxy
Callers must take a dependency on one or the other set of conversions
and default client behavior. Future changes may add a Versioned() type,
but this is an accurate reflection of current code state.
The unified RESTMapper and Typer follow the new rules, but on error will
fallback to the legacy path (while still supporting Unstructured
objects). This allows callers to handle the appropriate distinction
themselves if necessary.
Add a LocalParam() method to the resource.Builder that DRYs up a large
chunk of complicated code in set commands.
Also remove error messages that depended on ObjectKinds() - future
changes will potentially remove this interface and the replacements here
are equivalent.
