Automatic merge from submit-queue (batch tested with PRs 57511, 57978). 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>.
Renews cached NodeInfo with new vSphere connection
**What this PR does / why we need it**:
This PR modifies two public functions of nodemanager.go- GetNodeInfo and GetNodeDetails. For both these functions NodeInfo object is renewed with new GoVmomiClient and new vclib VirtualMachine and Datacenter.
**Which issue(s) this PR fixes** :
Fixes vmware#404
**Special notes for your reviewer**:
Code has been structured to minimize impact on existing 1.9 release code and any side-effects due to NodeInfo modification. This is a quick solution for vSphere connection renewal problem. A more enhanced solution is target for upcoming major release.
Testing:
- [x] Successfully tried out pod creation, deletion with dynamic volume.
- [x] Successfully ran e2e tests.
**Release note**:
```release-note
Fixes authentication problem faced during various vSphere operations.
```
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 NLB icmp permission duplication
**What this PR does / why we need it**:
Fixes an issue with the ICMP rule for MTU during the creation of a NLB
**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#56703
Automatic merge from submit-queue (batch tested with PRs 57991, 57789). 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 exists status for azure GetLoadBalancer
**What this PR does / why we need it**:
We see a lot of log indicating load balancer not found in azure:
```
E0109 07:00:31.126306 1 service_controller.go:776] Failed to process service kube-system/heapster. Retrying in 5m0s: error getting LB for service kube-system/heapster: Service(kube-system/heapster) - Loadbalancer not found
I0109 07:00:31.126384 1 event.go:218] Event(v1.ObjectReference{Kind:"Service", Namespace:"kube-system", Name:"heapster", UID:"400266e7-f507-11e7-bbc2-000d3af86f66", APIVersion:"v1", ResourceVersion:"450", FieldPath:""}): type: 'Warning' reason: 'CreatingLoadBalancerFailed' Error creating load balancer (will retry): error getting LB for service kube-system/heapster: Service(kube-system/heapster) - Loadbalancer not found
I0109 07:00:31.158858 1 azure_backoff.go:177] LoadBalancerClient.List(name) - backoff: success
E0109 07:00:31.158930 1 service_controller.go:776] Failed to process service kube-system/kubernetes-dashboard. Retrying in 5m0s: error getting LB for service kube-system/kubernetes-dashboard: Service(kube-system/kubernetes-dashboard) - Loadbalancer not found
I0109 07:00:31.158988 1 event.go:218] Event(v1.ObjectReference{Kind:"Service", Namespace:"kube-system", Name:"kubernetes-dashboard", UID:"4052f12b-f507-11e7-bbc2-000d3af86f66", APIVersion:"v1", ResourceVersion:"498", FieldPath:""}): type: 'Warning' reason: 'CreatingLoadBalancerFailed' Error creating load balancer (will retry): error getting LB for service kube-system/kubernetes-dashboard: Service(kube-system/kubernetes-dashboard) - Loadbalancer not found
```
It's interesting that those service does not need loadbalancer, and caller is just checking whether one loadbalancer exists.
009701f181/pkg/controller/service/service_controller.go (L287)
And in we can see when err is not nil, it will not check exists value. Thus we should not return error when exists=false.
This was changed in:
edfb2ad552 (diff-c901394068476b4ccb003a6c6efad57cR63)
The PR removes the error when exists=false.
**Which issue(s) this PR 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>.
Do not set BaseURI twice
**What this PR does / why we need it**:
Do not set BaseURI again. BaseURI has been set by NewAccountsClientWithBaseURI and NewDisksClientWithBaseURI method.
**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#57951
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
/assign @karataliu
add getNodeNameByID and use volume.AttachedDevice as devicepath
use uppercase functionname
do not delete automatically nodes if node is shutdowned in openstack
do not delete node
fix gofmt
fix cinder detach if instance is not in active state
fix gofmt
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>.
Ensure Azure LB removable when VMSS is enabled
**What this PR does / why we need it**:
When VMSS enabled, Azure LB not removed after all LoadBalancer services deleted.
**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#57826
**Special notes for your reviewer**:
This PR upgrades Azure GO SDK to latest release and adds a workaround to fix the problem.
**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>.
Add generic interface for azure clients
**What this PR does / why we need it**:
Continue of #43287. Moving remaining clients to generic interfaces.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Continue of #43287.
**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>.
Minor commenting fixes for Azure Disk Controllers from CR
**What this PR does / why we need it**:
Minor commenting fixes for Azure Disk Controllers from code review.
**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>.
Uniform Azure VM api calls
**What this PR does / why we need it**:
There is still a call to 'VirtualMachinesClient.Get' directly in azure_backoff, which does not go through the cache approach.
This PR uniforms all calls for getting azure vm to use 'getVirtualMachine'. Also refine unused 'exists' return value.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Related #57031
Follow-up #57432
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 57696, 57821, 56317). 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 DefaultMaxEBSVolumes constant into scheduler
**What this PR does / why we need it**:
A constant only used by the scheduler lives in the aws cloudprovider package. Moving the constant into the only package where it is used reduces import bloat. Testing with the dockerized build environment, the kube-scheduler binary went from 61748499 bytes to 47339144 bytes on amd64 with this change.
**Release note**:
```release-note
NONE
```
A constant only used by the scheduler lives in the aws cloudprovider
package. Moving the constant into the only package where it is used
reduces import bloat.
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>.
Add cache for VirtualMachinesClient.Get in azure cloud provider
**What this PR does / why we need it**:
Add a timed cache for 'VirtualMachinesClient.Get'
Currently cloud provider will send several get calls to same URL in short period, which is not necessary.
**Which issue(s) this PR fixes**:
Fixes#57031
**Special notes for your reviewer**:
**Release note**:
NONE
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>.
Fix vmss listing for Azure cloud provider
**What this PR does / why we need it**:
Fix a stupid bug of vmss listing: if there is only one instance, listScaleSetsWithRetry and listScaleSetVMsWithRetry will return empty list.
This PR also adds more verbose logs.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Related of #43287.
**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>.
Fixing vSphere Cloud Provider to use "vsphere-cloud-provider" to create ClientBuilder
**What this PR does / why we need it**:
vSphere cloud Provider is not using lower case naming while creating clientBuilder.
With this fix, ClientBuilder is created using lowercase naming.
With mixed upper-lower case name, controller manager is crashing.
**Which issue(s) this PR fixes**
Fixes # https://github.com/kubernetes/kubernetes/issues/57279
**Special notes for your reviewer**:
None
**Release note**:
```release-note
This fixes controller manager crash in certain vSphere cloud provider environment.
```
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>.
Remove useInstanceMetadata param from Azure cloud provider
**What this PR does / why we need it**:
With out-of-tree Azure cloud provider (#50752), metadata won't work any more (kubelet won't call those metadata functions any more).
This PR removes the parameter useInstanceMetadata from Azure cloud provider.
**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#57646.
**Special notes for your reviewer**:
**Release note**:
```release-note
Remove useInstanceMetadata parameter from Azure cloud provider.
```
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>.
Reduce VirtualMachineScaleSetsClient#List calls for Azure cloud provider
**What this PR does / why we need it**:
For master nodes not managed by VMSS, current cloud provider would updateCaches each time when finding master nodes info. This could result in call limits of `VirtualMachineScaleSetsClient#List`.
This PR adds a caches to those nodes which reduces the cache updating significantly.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Continue of #43287.
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
