Automatic merge from submit-queue.

Automated cherry pick of #62670: Fix kubelet flags.

Cherry pick of #62670 on release-1.10.

#62670: Fix kubelet flags.

Signed-off-by: Lantao Liu <lantaol@google.com>

Automatic merge from submit-queue.

Automated cherry pick of #63303: Return attach error to A/D controller.

Cherry pick of #63303 on release-1.10.

#63303: Return attach error to A/D controller.

```release-note
Fixed error reporting of CSI volumes attachment.
```

Automatic merge from submit-queue.

Automated cherry pick of #62505: update godeps to use latest pflag

Cherry pick of #62505 on release-1.10.

#62505: update godeps to use latest pflag

```release-note
Show help for deprecated Kubelet flags
```

Automatic merge from submit-queue.

Automated cherry pick of #64383: fix azure file size grow issue

Cherry pick of #64383 on release-1.10.

#64383: fix azure file size grow issue

Automatic merge from submit-queue.

Automated cherry pick of #63295: Fix CSI volume detach when the volume is already detached.

Cherry pick of #63295 on release-1.10.

#63295: Fix CSI volume detach when the volume is already detached.

```release-note
Fixed detach of already detached CSI volumes.
```

Automatic merge from submit-queue.

properly skip cadvisor proxy test

xref https://github.com/kubernetes/kubernetes/issues/64158

per https://github.com/kubernetes/kubernetes/issues/64158#issuecomment-393229314

/assign @liggitt @AishSundar

Automatic merge from submit-queue.

Automated cherry pick of #63918: new event exporter config with support for new stackdriver

Cherry pick of #63918 on release-1.10.

#63918: new event exporter config with support for new stackdriver

Automatic merge from submit-queue.

Wait for the job to be removed

**What this PR does / why we need it**:
In master we've dropped reapers (#63979) which means the old client does not wait long enough for the resource to be gone (since it's being removed on the server along with its dependents). To fix our e2e ([failure here](https://k8s-gubernator.appspot.com/build/kubernetes-jenkins/logs/ci-kubernetes-e2e-gce-new-master-gci-kubectl-skew/7980)) we need to backport parts of the aforementioned PR which is updating tests to wait for the resource to be removed. 

**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 #64362

/assign @MaciekPytel 
/assign @deads2k 

**Release note**:
```release-note
NONE
```

Automatic merge from submit-queue.

Automated cherry pick of #64349: Fix nodeport repair for ESIPP services

Cherry pick of #64349 on release-1.10.

#64349: Fix nodeport repair for ESIPP services

Merge pull request #64478 from jiayingz/automated-cherry-pick-of-#64296-#64340-#64404-upstream-release-1.10

Automatic merge from submit-queue.

Automated cherry pick of #64296: Update nvidia-gpu-device-plugin to apps/v1 and use #64340: Fix DsFromManifest() after we switch from extensions/v1beta1 #64404: DaemonSet internals are still in extensions

Cherry pick of #64296 #64340 #64404 on release-1.10.

#64296: Update nvidia-gpu-device-plugin to apps/v1 and use
#64340: Fix DsFromManifest() after we switch from extensions/v1beta1
#64404: DaemonSet internals are still in extensions

```release-note
Fix nvidia-gpu-device-plugin DaemonSet config to ensure correct device plugin upgrade
```

Automatic merge from submit-queue.

Automated cherry pick of #63720: log error for os.NewComputeV2

Cherry pick of #63720 on release-1.10.

#63720: log error for os.NewComputeV2

Fix DsFromManifest() after we switch from extensions/v1beta1 to apps/v1 in cluster/addons/device-plugins/nvidia-gpu/daemonset.yaml.

Even though RollingUpdate is the default updateStrategy, we need to
specify it explicitly here because otherwise updating from
extensions/v1beta1 to apps/v1 doesn't change the updateStrategy.

Automatic merge from submit-queue.

Automated cherry pick of #63806: kubeadm - do not generate etcd ca/certs for external etcd

Cherry pick of #63806 on release-1.10.

#63806: kubeadm - do not generate etcd ca/certs for external etcd

Automatic merge from submit-queue.

Automated cherry pick of #64150 upstream release 1.10

Cherry pick of #64150 on release-1.10.
#64150 : Fix incorrectly set resource version in List

Automatic merge from submit-queue.

Automated cherry pick of #64394 upstream release 1.10 

Cherry pick of #64394 on release-1.10.
#64394 : Fix GKE Regional Clusters upgrade tests

"NotFound" error should be treated as successful detach.

WaitForAttach runs on nodes, not in A/D controller.

Automatic merge from submit-queue.

Add retry to AssertCleanup

**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 addresses the other two problems mentioned in https://github.com/kubernetes/kubernetes/issues/64362 by cherry-picking https://github.com/kubernetes/kubernetes/pull/61565 and https://github.com/kubernetes/kubernetes/pull/62155 on to 1.10 release. 

**Release note**:
```release-note
NONE
```

Automatic merge from submit-queue.

Automated cherry pick of #64180:  Check SessionAffinity for Azure load balancer

Cherry pick of #64180 on release-1.10.

#64180: Check SessionAffinity for Azure load balancer

Automatic merge from submit-queue.

Automated cherry pick of #61926: Update libcontainer to include PRs with fixes to systemd

Cherry pick of #61926 on release-1.10.

#61926: Update libcontainer to include PRs with fixes to systemd

Automatic merge from submit-queue.

Prevent 1.10 e2es testing deprecated CAdvisorPort in 1.11

**What this PR does / why we need it**:
The public cadvisor port by default is disabled in #63881, targeted for v1.11.

But 1.10 e2e tests get run against v1.11+ masters during upgrade tests.

https://k8s-testgrid.appspot.com/sig-release-master-upgrade#gce-1.10-master-upgrade-cluster-parallel

**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 #64158

**Special notes for your reviewer**:
/cc luxas BenTheElder krzyzacy
 
**Release note**:

```release-note
Prevent 1.10 e2es testing deprecated CAdvisorPort in v1.11
```

Automatic merge from submit-queue.

Automated cherry pick of #63696: gce: Prefer MASTER_ADVERTISE_ADDRESS in apiserver setup

Cherry pick of #63696 on release-1.10.

#63696: gce: Prefer MASTER_ADVERTISE_ADDRESS in apiserver setup

```release-note
NONE
```

Automatic merge from submit-queue.

Automated cherry pick of #61459: etcd client add dial timeout

Cherry pick of #61459 on release-1.10.

Fixes https://github.com/kubernetes/kubernetes/issues/62538

#61459: etcd client add dial timeout

```release-note
kube-apiserver now sets a dial timeout when connecting to etcd, to allow it to tolerate/retry connections to a dead etcd server
```

Automatic merge from submit-queue.

Don't panic is admission options is nil

Backport https://github.com/kubernetes/kubernetes/pull/63460

cc: @sttts @deads2k 

```release-note
nil admission options are now consistent with other nil options.
```

Automatic merge from submit-queue.

Automated cherry pick of #63961: Add a way to pass extra arguments to etcd.

Cherry pick of #63961 on release-1.10.

#63961: Add a way to pass extra arguments to etcd.

Only generate the etcd CA and certificates if not configured for external
etcd.