Commit Graph

47170 Commits

Author SHA1 Message Date
Cao Shufeng
939b962975 refactor code from restful namer
Some codes are replaced because they will never run.
2017-04-20 17:34:08 +08:00
Kubernetes Submit Queue
94a5074bd6 Merge pull request #44565 from vmware/kubernetes-e2e-v7
Automatic merge from submit-queue

adding test for volume fstype validation

**What this PR does / why we need it**:
This PR is adding a test for volume fstype validation. Test verifies fstype specified in storage-class is being honored after volume creation.

Steps:

1. Create StorageClass with fstype set to valid type (default case included). 
2. Create PVC which uses the StorageClass created in step 1.
3. Wait for PV to be provisioned.
4. Wait for PVC's status to become Bound.
5. Create pod using PVC on specific node.
6. Wait for Disk to be attached to the node.
7. Execute command in the pod to get fstype.
8. Delete pod and Wait for Volume Disk to be detached from the Node.
9. Delete PVC, PV and Storage Class.



**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #

**Special notes for your reviewer**:

**Release note**:

```release-note
None
```

cc: @jeffvance @tusharnt
2017-04-20 01:25:20 -07:00
Kubernetes Submit Queue
33bdec3f22 Merge pull request #44696 from justinsb/fix_44695
Automatic merge from submit-queue (batch tested with PRs 42272, 44696)

e2e test fix: Wait longer when first creating ELB

On any cloud (GCE or AWS), a lag between creating the LoadBalancer and
having it actually start serving traffic is expected.  On AWS the lag is
larger, and we weren't correctly using the longer wait on our first
request.

Use a longer wait period on our first request.

Fix #44695

```release-note
NONE
```
2017-04-20 00:33:01 -07:00
xilabao
c1197924cd fix kubeadm init when certdir changed 2017-04-20 15:23:45 +08:00
caiyixiang
31829850b9 delete unuse const 2017-04-20 15:17:36 +08:00
Kubernetes Submit Queue
afc01d92d2 Merge pull request #42272 from marun/apiserver-fail-fast
Automatic merge from submit-queue

apiserver: Update genericapiserver to panic on listener error

Previously runServer would try to listen again if a listener error occurred.  This commit changes the response to a panic to allow a process manager (systemd/kubelet/etc) to react to the failure.

**Release note**:

```release-note
The Kubernetes API server now exits if it encounters a networking failure (e.g. the networking interface hosting its address goes away) to allow a process manager (systemd/kubelet/etc) to react to the problem.  Previously the server would log the failure and try again to bind to its configured address:port.
```

cc: @liggitt @sttts @deads2k @derekwaynecarr
2017-04-19 23:51:34 -07:00
Wojciech Tyczynski
2f250435fd Don't rebuild endpoints map in iptables kube-proxy all the time. 2017-04-20 08:34:46 +02:00
Kubernetes Submit Queue
a40fef5db7 Merge pull request #40890 from marun/e2e-use-kubeconfig-host
Automatic merge from submit-queue

e2e: Prefer kubeconfig host to default

Previously it was necessary to pass ``-host`` to ``e2e.test`` even if ``-kubeconfig`` was specified since otherwise a localhost default would be used.  This change ensures that the default is only used when kubeconfig is not set. 

cc: @jayunit100
2017-04-19 20:26:00 -07:00
FengyunPan
1bad02cb22 Use const value for pod's conditions 2017-04-20 10:43:24 +08:00
Justin Santa Barbara
854900c4a1 e2e test fix: Wait longer when first creating ELB
On any cloud (GCE or AWS), a lag between creating the LoadBalancer and
having it actually start serving traffic is expected.  On AWS the lag is
larger, and we weren't correctly using the longer wait on our first
request.

Use a longer wait period on our first request.

Fix #44695
2017-04-19 22:12:59 -04:00
Kubernetes Submit Queue
fba605ce05 Merge pull request #44661 from xiangpengzhao/fix-vsphere-panic
Automatic merge from submit-queue (batch tested with PRs 44687, 44689, 44661)

Fix panic when using `kubeadm init` with vsphere cloud-provider

**What this PR does / why we need it**:
Check if the reference is nil when finding machine reference by UUID.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #44603

**Special notes for your reviewer**:
This is just a quick fix for the panic.

**Release note**:

```release-note
NONE
```
2017-04-19 18:52:59 -07:00
Kubernetes Submit Queue
6d67cdd4d9 Merge pull request #44689 from rmmh/avoid-get-kube-dl-flake
Automatic merge from submit-queue (batch tested with PRs 44687, 44689, 44661)

Retry in get-kube.sh to avoid download flakes.

GCS has up to 2% 5xx rates, so retrying is critical.

This is currently failing about 8 times per day [according to the dashboard](https://storage.googleapis.com/k8s-gubernator/triage/index.html?test=Extract#be2f33fb1e6dd2389d12). It could be backported to reduce the flake rate.

Relase note:
```release-note
NONE
```
2017-04-19 18:52:58 -07:00
Kubernetes Submit Queue
18cd0f48ba Merge pull request #44687 from cblecker/hacklib-tar
Automatic merge from submit-queue

Add hack/lib to kubernetes release tarball

**What this PR does / why we need it**:
Add hack/lib to kubernetes release tarball, to fix an issue with https://get.k8s.io/ script introduced in #42748.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*:
Fixes https://github.com/kubernetes/kubernetes/pull/42748#issuecomment-295412268

**Special notes for your reviewer**:
I'm new to bazel, so hopefully I'm not off-base here :)

**Release note**:

```release-note
NONE
```

cc: @ixdy @dcbw @smarterclayton
2017-04-19 18:09:26 -07:00
Mike Danese
0800ab92fb e2e test client creation using the certificates API 2017-04-19 17:35:52 -07:00
Kubernetes Submit Queue
36c5d12cf4 Merge pull request #44452 from gnufied/fix-aws-device-failure-reuse
Automatic merge from submit-queue

Implement LRU for AWS device allocator

On failure to attach do not use device from pool
    
In AWS environment when attach fails on the node
lets not use device from the pool. This makes sure
that a bigger pool of devices is available.
2017-04-19 16:38:13 -07:00
Daniel Nardo
2320fde49c Update kubelet to use the network-plugin-dir if the cni-bin-dir flag
is not set.
2017-04-19 16:33:06 -07:00
Kubernetes Submit Queue
d1f4664f5b Merge pull request #44677 from wwwtyro/rye/kube-proxy-master-flag
Automatic merge from submit-queue

select one api endpoint at random when deploying kubernetes-core charm

**What this PR does / why we need it**: Fixes a bug in the kubernetes-worker Juju charm code that attempted to give kube-proxy more than one api endpoint.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #

**Special notes for your reviewer**: https://github.com/juju-solutions/bundle-canonical-kubernetes/issues/255

**Release note**:

```release-note
Fixes a bug in the kubernetes-worker Juju charm code that attempted to give kube-proxy more than one api endpoint.
```
2017-04-19 15:45:35 -07:00
Anthony Yeh
be1fe95534 CronJob: Use PATCH to adopt Jobs. 2017-04-19 15:42:34 -07:00
Anthony Yeh
33d7788793 CronJob: Add e2e test for adoption.
Currently, an e2e test is the only way to ensure we have the proper RBAC
permissions to adopt Jobs.
2017-04-19 15:42:34 -07:00
Anthony Yeh
1e14323ac2 CronJob: List children before parents.
This prevents a race with the GC while it orphans dependents.
2017-04-19 15:42:34 -07:00
Anthony Yeh
4e1b07d9c2 CronJob: Check ControllerRef Name and UID in unit test. 2017-04-19 15:42:34 -07:00
Anthony Yeh
5cef455a0d CronJob: Use synchronous deletion for CronJobs in e2e.
This is needed now that the default is OrphanDependents.
2017-04-19 15:42:34 -07:00
Christoph Blecker
89500f149a
Add hack/lib to kubernetes release tarball 2017-04-19 15:33:43 -07:00
Ryan Hitchman
65cd30dbbe Retry in get-kube.sh to avoid download flakes. 2017-04-19 15:30:28 -07:00
Anthony Yeh
01a4ec75d4 Update CHANGELOG.md for v1.6.2. 2017-04-19 15:09:08 -07:00
Maru Newby
9413071ce8 e2e: Prefer kubeconfig host to default 2017-04-19 14:58:43 -07:00
Kubernetes Submit Queue
3bede2fb4c Merge pull request #44679 from ixdy/porter-cert
Automatic merge from submit-queue

Update gcr.io/google_containers/porter image to 4524579c0e

**What this PR does / why we need it**: updates the porter image to one built at 4524579c0e using go1.8.1.

This incorporates #44638, which has a new dummy certificate that is compliant with go1.8+.

Image has already been pushed.

**Release note**:

```release-note
NONE
```

/assign @liggitt
/cc @luxas @lavalamp
2017-04-19 14:07:27 -07:00
Anthony Yeh
d5b86bbae4 Job: Add e2e test for Pod adopt/release.
An e2e test is currently the only way to ensure we have the correct RBAC
permissions to PATCH Pods.
2017-04-19 14:03:36 -07:00
Anthony Yeh
aa7bc25812 Job: Add PATCH Pods permission.
This is needed to update ControllerRef during adopt/release.
2017-04-19 14:03:36 -07:00
Anthony Yeh
f502ab6a8b Job: Recheck DeletionTimestamp before adoption.
This is necessary to avoid racing with the GC when it orphans
dependents.
2017-04-19 14:03:36 -07:00
Anthony Yeh
fd8dd26d09 Job: Check ControllerRef Name and UID in unit test. 2017-04-19 14:03:36 -07:00
Anthony Yeh
06536cb357 Job: Check that ControllerRef UID matches. 2017-04-19 14:03:36 -07:00
Anthony Yeh
e207f6c767 Job: Fix CronJob e2e test for async Job deletion.
Now that the default delete option for Job is OrphanDependents,
Job deletion is asynchronous.
2017-04-19 14:03:36 -07:00
Anthony Yeh
c82b537bee Job: Always set BlockOwnerDeletion in ControllerRef. 2017-04-19 14:03:36 -07:00
Anthony Yeh
ad026026e0 Job: Update Lister documentation for ControllerRef.
The Job Listers still use selectors, because this is the
behavior expected by callers. This clarifies the meaning of the
returned list. Some callers may need to switch to using
GetControllerOf() instead, but that is a separate, case-by-case issue.
2017-04-19 14:03:36 -07:00
Anthony Yeh
067a8ff3c2 Job: Use ControllerRef to route watch events.
This is part of the completion of ControllerRef, as described here:

https://github.com/kubernetes/community/blob/master/contributors/design-proposals/controller-ref.md#watches
2017-04-19 14:03:35 -07:00
Hemant Kumar
a16ee2f514 Implement LRU for AWS device allocator
In AWS environment when attach fails on the node
lets not use device from the pool. This makes sure we
don't reuse recently freed devices
2017-04-19 16:52:57 -04:00
Kubernetes Submit Queue
dd3085cdab Merge pull request #44678 from wwwtyro/rye/fix-service-removal
Automatic merge from submit-queue

Fixes a missing comma in a list of strings.

**What this PR does / why we need it**: Fixes a missing comma in a list of strings in the kubernetes-worker Juju charm.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #

**Special notes for your reviewer**:

**Release note**:

```release-note
Fixes a missing comma in a list of strings.
```
2017-04-19 13:21:40 -07:00
Jeff Grafton
df4e205460 Update gcr.io/google_containers/porter image to 4524579c0e 2017-04-19 11:50:41 -07:00
Rye Terrell
39818e16d4 fix missing comma in list of strings 2017-04-19 13:33:08 -05:00
Rye Terrell
6002481524 select one api endpoint at random instead of erroneously using all of them 2017-04-19 13:19:35 -05:00
Kubernetes Submit Queue
4524579c0e Merge pull request #44674 from madhusudancs/fed-dump-logs-pod-pods
Automatic merge from submit-queue (batch tested with PRs 44499, 44674)

Strip "pods/" prefix from the pod names returned by kubectl get pods.

Note that the result returned by kubectl get -o name has plural prefixes "pods". We were already trying to remove the prefix "pod", but that's not how the results are returned unfortunately.

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

cc @perotinus @kubernetes/sig-federation-pr-reviews
2017-04-19 09:51:58 -07:00
Kubernetes Submit Queue
505ec43dab Merge pull request #44499 from wojtek-t/edge_based_services_in_proxy
Automatic merge from submit-queue

Edge based services in proxy

This is sibling effort to what I did for endpoints in KubeProxy.
This PR is first one (changing config & iptables) - userspace will follow.
2017-04-19 09:43:04 -07:00
Kubernetes Submit Queue
20a6c4caf9 Merge pull request #44673 from gmarek/multiple_files
Automatic merge from submit-queue (batch tested with PRs 44667, 44673)

Allow summaries to be printed out to ReportDir instead of stdout

Fix #44003

cc @shyamjvs
2017-04-19 08:55:59 -07:00
p0lyn0mial
4a3dcff3bf bazel update 2017-04-19 17:22:44 +02:00
Kubernetes Submit Queue
d46983f2e5 Merge pull request #44667 from gmarek/fix_get_traces
Automatic merge from submit-queue

Fix traces for get requests

Currently traces for GETs are not useful at all, as they're measuring the wrong thing.
2017-04-19 08:06:41 -07:00
Madhusudan.C.S
c2a8f0bae9 Remove "pods/" prefix from the pod names returned by kubectl get pods.
Note that the result returned by kubectl get -o name has plural prefixes "pods".
2017-04-19 07:39:56 -07:00
gmarek
be987ac247 Allow summaries to be printed out to ReportDir instead of stdout 2017-04-19 16:17:36 +02:00
Kubernetes Submit Queue
daeac53d51 Merge pull request #44635 from Cynerva/gkk/ceph-secret-type
Automatic merge from submit-queue

Fix ceph-secret type to kubernetes.io/rbd in kubernetes-master charm

**What this PR does / why we need it**:

This fixes the type of the ceph-secret secret that's created by the kubernetes-master charm.

Without the `kubernetes.io/rbd` type, automatic provisioning of PVCs doesn't work.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #

**Special notes for your reviewer**:

**Release note**:

```release-note
Fix ceph-secret type to kubernetes.io/rbd in kubernetes-master charm
```
2017-04-19 07:15:57 -07:00
Klaus Ma
47dbb8bbc9 Used Pods interface for binding. 2017-04-19 21:32:38 +08:00