Commit Graph

39388 Commits

Author SHA1 Message Date
Kubernetes Submit Queue
8feec28d6e Merge pull request #37248 from mtaufen/serial-port-output
Automatic merge from submit-queue

Include serial port output in GCP log-dump

This should help us debug reboot failures e.g. https://github.com/kubernetes/kubernetes/issues/33882
Also useful for OOM-kill investigation.

/cc @dchen1107
2016-11-21 18:21:27 -08:00
Kubernetes Submit Queue
a85d352de9 Merge pull request #37234 from liggitt/revert-flex_node_conditions
Automatic merge from submit-queue

Revert "Avoid hard-coding list of Node Conditions"

* we don't know how other API consumers are using node conditions (there was no prior expectation that the scheduler would block on custom conditions)
* not all conditions map directly to schedulability (e.g. `MemoryPressure`/`DiskPressure`)
* not all conditions use True to mean "unschedulable" (e.g. `Ready`)

This reverts commit 511b2ecaa8 to avoid breaking existing API users and to avoid constraining future uses of the node conditions API
2016-11-21 17:41:59 -08:00
Kubernetes Submit Queue
88a809470d Merge pull request #34454 from dshulyak/current_enter_running
Automatic merge from submit-queue

Test case for scalling down before scale up finished

Verifies that current pet (one which was created last by scale up action)
will enter running before scale down will take place

related: #30082
2016-11-21 17:01:33 -08:00
Kubernetes Submit Queue
53caa61c20 Merge pull request #35767 from xiangpengzhao/fix-api-doc
Automatic merge from submit-queue

Fix a link in api doc

Remove an extra symbol `, it breaks a link in api doc.
cc @caesarxuchao , thanks!
2016-11-21 15:57:07 -08:00
Kubernetes Submit Queue
7bd518efc8 Merge pull request #34617 from VojtechVitek/patch-1
Automatic merge from submit-queue

Update Docker for Mac installation link

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

Fixes broken obsolete link about Docker for Mac installation.
2016-11-21 15:20:10 -08:00
Kubernetes Submit Queue
0fd70a4990 Merge pull request #36444 from soltysh/issue31481
Automatic merge from submit-queue

Read all resources for finalization and gc, not just preferred

Fixes #31481. 

Currently when starting namespace controller or garbage collector we only gather preferred version resources, which in case of multiple versions (you guessed it `batch/v2alpha1.CronJobs` again) isn't sufficient. This PR adds additional method which actually retrieves all resources from all versions and works on them.

@kubernetes/sig-api-machinery ptal
2016-11-21 15:19:59 -08:00
Michael Taufen
9f547084c3 Include serial port output in GCP log-dump 2016-11-21 14:41:02 -08:00
Kubernetes Submit Queue
67dc87395b Merge pull request #36869 from madhusudancs/kubefed-join-context-secret-name-flags
Automatic merge from submit-queue

[Federation][join-flags] Add flags for cluster context and secret names while joining clusters to federation.

Vast majority of cluster contexts are not RFC 1123 subdomains. Since
cluster and secret names for the API objects are derived from the
cluster context name, there is no way for users to join clusters
with such context names to federation, unless they modify the context
name in their kubeconfigs itself. That's a lot of inconvenience and
entirely goes against the goal and beats the purpose of the `kubefed`
tool. So we are providing these flags to allow users to override these
values.

Also, since users register their clusters with federation, it is makes
sense in terms of user experience to make the cluster name a positional
argument because that feels more natural. Also, specifying cluster name
in the join command as a mandatory positional argument make
`kubefed join` consistent with `kubefed unjoin`. This also means
`--cluster-context` is now made a flag and defaults to cluster name if
unspecified.

`--secret-name` also defaults to the cluster name if unspecified.

Fixes: Issue #35954

cc @kubernetes/sig-cluster-federation @quinton-hoole @irfanurrehman
2016-11-21 13:02:03 -08:00
Kubernetes Submit Queue
a47614dd15 Merge pull request #37122 from childsb/revert_gid
Automatic merge from submit-queue

Revert "Use Gid when provisioning Gluster Volumes."

On further inspection the design in #35460 was not secure enough.  This PR reverts the change. 

This reverts commit 7a0d219d12.
2016-11-21 12:11:31 -08:00
Jordan Liggitt
c214abb5ff
Revert "Avoid hard-coding list of Node Conditions"
This reverts commit 511b2ecaa8.
2016-11-21 14:55:11 -05:00
Kubernetes Submit Queue
435268ac19 Merge pull request #32868 from dshulyak/scale_up_down_test_case
Automatic merge from submit-queue

Validate pet set scale up/down order

This change implements test cases to validate that:
- scale up will be done in order
- scale down in reverse order
- if any of the pets will be unhealthy, scale up/down will halt

related: https://github.com/kubernetes/kubernetes/issues/30082
2016-11-21 11:36:01 -08:00
Kubernetes Submit Queue
ab20072189 Merge pull request #37203 from Random-Liu/fix-restart-test
Automatic merge from submit-queue

Filter out non-RestartAlways mirror pod in restart test.

Fixes #37202.

> A quick fix is to filter out non-RestartAlways pods. Because either RestartNever and RestartOnFailure pods could succeed, and we can not deal with terminated mirror pods very well now.

@yujuhong @gmarek 
/cc @kubernetes/sig-node
2016-11-21 10:58:02 -08:00
Kubernetes Submit Queue
27a0a86282 Merge pull request #37174 from mbohlool/optional
Automatic merge from submit-queue

Fields with omitempty tag should still be considered as optional

We've added an "+optional" tag while ago for optional fields. Before that OpenAPI spec generated assumed all fields with "omitempty" in their json tags are optional. This should be still the case (as well as +optional tag) until these two things happen:

- We update all documentation asking developers to use +optional (My bad, I should have added this after the +optional PR)
- We fix swagger 1.2 spec generator to use +optional instead of omitempty.

Fixes #37149
2016-11-21 10:57:52 -08:00
Kubernetes Submit Queue
9b2c9681c9 Merge pull request #37223 from gmarek/hollow-node
Automatic merge from submit-queue

Increase hollow-node qps limit

@saad-ali - test only change. I really need tests to be faster...
2016-11-21 09:54:50 -08:00
Kubernetes Submit Queue
dff670aac8 Merge pull request #37123 from Crassirostris/kube-proxy-logs
Automatic merge from submit-queue

Add kube-proxy logs to fluentd configs

Related to https://github.com/kubernetes/kubernetes/issues/37107

Makes fluentd collect logs from kube-proxy. It's completely backward-compatible change that does not cause problems currently, so I suggest not to bump version.

cc @piosz
2016-11-21 09:54:39 -08:00
Kubernetes Submit Queue
2af762b67a Merge pull request #37142 from AdoHe/fix_kubectl_get_all
Automatic merge from submit-queue

kubectl resource filter convert unstructured obj before apply filter …

Fix #37128 

@pwittrock @saad-ali ptal.
2016-11-21 08:28:02 -08:00
Kubernetes Submit Queue
6b8c6e169d Merge pull request #37221 from gmarek/master
Automatic merge from submit-queue

Update start-kubemark-master to correctly work in case of master-comp…

@saad-ali - test only change.
2016-11-21 07:52:35 -08:00
Kubernetes Submit Queue
c9ed5262c7 Merge pull request #35947 from brendandburns/tpr
Automatic merge from submit-queue

make groupVersionResource listing dynamic for namespace controller

@derekwaynecarr @kubernetes/sig-api-machinery 

```release-note
Third party resources are now deleted when a namespace is deleted.
```

Fixes https://github.com/kubernetes/kubernetes/issues/32306
2016-11-21 07:52:25 -08:00
gmarek
70f631a13f Increase hollow-node qps limit 2016-11-21 16:46:25 +01:00
Mik Vyatskov
b5b47620f5 Add kube-proxy logs to fluentd configs 2016-11-21 15:43:28 +01:00
gmarek
15471488ba Update start-kubemark-master to correctly work in case of master-components restart 2016-11-21 15:40:58 +01:00
Wojciech Tyczynski
75108a3757 Merge pull request #37217 from kubernetes/revert-37095-kubemark_logrotation
Revert "Add log rotation to kubemark"
2016-11-21 15:38:18 +01:00
Kubernetes Submit Queue
907d132087 Merge pull request #37069 from zreigz/fix-etcd-for-ubuntu
Automatic merge from submit-queue

Fix etcd unavailable error when performing kube-up.sh for Ubuntu prov…

**What this PR does / why we need it**:
This PR fixes 'etcd unavailable error' when performing kube-up.sh for Ubuntu provider

**Which issue this PR fixes** 
fixes: https://github.com/kubernetes/kubernetes/issues/36340
2016-11-21 06:20:27 -08:00
Wojciech Tyczynski
9ccddb9b7d Revert "Add log rotation to kubemark" 2016-11-21 14:55:52 +01:00
Kubernetes Submit Queue
21c3b3028f Merge pull request #34414 from yarntime/add_defaults_test
Automatic merge from submit-queue

add test file for autoscaling defaults

add test file `pkg/apis/autoscaling/v1/defaults_test.go`
2016-11-21 05:23:13 -08:00
Kubernetes Submit Queue
47cf64e9c5 Merge pull request #37198 from gmarek/test-command
Automatic merge from submit-queue

Add commented local testing command in kubemark/run-e2e-tests.sh
2016-11-21 02:06:10 -08:00
Dmitry Shulyak
c09e969e08 Test case for scalling down before scale up finished
Verifies that current pet (one which was created last by scale up action)
will enter running before scale down will take place

Change-Id: Ib47644c22b3b097bc466f68c5cac46c78dd44552
2016-11-21 11:46:34 +02:00
Random-Liu
d9d148f4d6 Filter out non-RestartAlways mirror pod in restart test. 2016-11-21 01:43:14 -08:00
Dmitry Shulyak
36d1af6145 Validate pet set scale up/down order
This change implements test cases to validate that:

    scale up will be done in order
    scale down in reverse order
    if any of the pets will be unhealthy, scale up/down will halt

Change-Id: I4487a7c9823d94a2995bbb06accdfd8f3001354c
2016-11-21 11:21:32 +02:00
gmarek
3d9f002ebd Add commented local testing command in kubemark/run-e2e-tests.sh 2016-11-21 10:00:54 +01:00
Kubernetes Submit Queue
da007483bd Merge pull request #37193 from gmarek/event-disk
Automatic merge from submit-queue

Pass GCLOUD_COMMON_ARGS do disk attach in Kubemark
2016-11-21 00:51:43 -08:00
gmarek
87f87206bc Pass GCLOUD_COMMON_ARGS do disk attach in Kubemark 2016-11-21 09:04:04 +01:00
Kubernetes Submit Queue
ab184e1b1f Merge pull request #36399 from mwielgus/fedrc1
Automatic merge from submit-queue

Make desired objects in fed controller reconcilation function fully writable

When creating desired objects for sub-cluster in reconciliate functions we don't make full copies. This may result in unexpected race condition in the future when someone actually adds some object modifications (not needed in most cases but anyway - the cost of extra copy will be relatively small comparing to inter-cluster traffic). In case of ReplicaSet and Deployment the spec copy was a shallow one - ok for setting just replicas but also error-prone in the future.

cc: @quinton-hoole @nikhiljindal @madhusudancs
2016-11-21 00:00:26 -08:00
Brendan Burns
ef6529bf2f make groupVersionResource listing dynamic when third party resources are
enabled.
2016-11-20 20:48:57 -08:00
Kubernetes Submit Queue
0042ce5684 Merge pull request #36892 from gmarek/nodecontroller
Automatic merge from submit-queue

Add logs near force deletions of Pods

We should always log something when control plane force deletes the Pod.

@davidopp I think that logging force deletions is enough, or do you think we should log soft deletions as well?

cc @deads2k
2016-11-20 16:00:10 -08:00
mbohlool
55d7fbd52c Update generated OpenAPI spec 2016-11-20 00:23:40 -08:00
mbohlool
71c07650ea All omitempty fields should be optional 2016-11-20 00:16:33 -08:00
Kubernetes Submit Queue
238459b092 Merge pull request #35968 from resouer/fix-predicates
Automatic merge from submit-queue

Fix invalid predicates describe

Just fix some nits when working on predicate cache.

Anyone can review and approve this.
2016-11-19 13:19:41 -08:00
Kubernetes Submit Queue
2b86ed3f90 Merge pull request #37077 from soltysh/issue34585
Automatic merge from submit-queue

Retry job update after failure to prevent modification conflict

This fixes #34585 flake.

@janetkuo || @kubernetes/sig-apps  ptal

I've been getting too many emails recently wrt to that issue, so I wanted to "clean" my inbox a bit 😉
2016-11-19 12:43:14 -08:00
Maciej Szulik
b253c20d80 Retry job update after failure to prevent modification conflict 2016-11-19 19:40:08 +01:00
Harry Zhang
5554dbf907 Fix invalid predicates describe 2016-11-19 22:30:15 +08:00
Kubernetes Submit Queue
076051b9f2 Merge pull request #37113 from kubernetes/revert-35697-upgrade-tests
Automatic merge from submit-queue

Revert "Add more test cases to k8s e2e upgrade tests"

Reverts kubernetes/kubernetes#35697
2016-11-19 01:39:44 -08:00
Kubernetes Submit Queue
b9d2d74a94 Merge pull request #37038 from ymqytw/retry_old_patch_after_new_patch_fail
Automatic merge from submit-queue

Fix kubectl Stratigic Merge Patch compatibility

As @smarterclayton pointed out in [comment1](https://github.com/kubernetes/kubernetes/pull/35647#pullrequestreview-8290820) and [comment2](https://github.com/kubernetes/kubernetes/pull/35647#pullrequestreview-8290847) in PR #35647,
we cannot assume the API servers publish version and they shares the same version.

This PR removes all the calls of GetServerSupportedSMPatchVersion().
Change the behavior of `apply` and `edit` to:
Retrying with the old patch version, if the new version fails.
Default other usage of SMPatch to the new version, since they don't update list of primitives.

fixes #36916

cc: @pwittrock @smarterclayton
2016-11-19 01:02:47 -08:00
Kubernetes Submit Queue
290dff2d9d Merge pull request #36911 from dashpole/eviction_defaults
Automatic merge from submit-queue

Eviction Thresholds Update

Sets the defaults for the eviction-hard threshold for GCE based on what we were using during testing: "memory.available<250Mi,nodefs.available<10%,nodefs.inodesFree<5%".
Sets flags for e2e tests to use eviction-minimum-reclaim: "nodefs.available<5%,nodefs.inodesFree<5%"

this fixes #32537
2016-11-19 00:21:19 -08:00
Kubernetes Submit Queue
75affa4a43 Merge pull request #37059 from mbohlool/dev2
Automatic merge from submit-queue

OpenAPI Bugfix: []byte should be treated as integer array

data field of v1.Secret is a map of string to byte array. Generated spec should generate a map of string to (type="string", format="byte" that means map of base64 string) however current code converts it to an array of integer that is wrong.

fixes #37126
2016-11-18 23:44:20 -08:00
Kubernetes Submit Queue
d725b3e3cd Merge pull request #36775 from bowei/kube-dns-config-map
Automatic merge from submit-queue

Add limited config-map support to kube-dns

This is an integration bugfix for https://github.com/kubernetes/kubernetes/issues/36194

```release-note
kube-dns

Added --config-map and --config-map-namespace command line options. 
If --config-map is set, kube-dns will load dynamic configuration from the config map 
referenced by --config-map-namespace, --config-map. The config-map supports
the following properties: "federations".

--federations flag is now deprecated. Prefer to set federations via the config-map.
Federations can be configured by settings the "federations" field to the value currently 
set in the command line.

Example:

  kind: ConfigMap
  apiVersion: v1
  metadata:
    name: kube-dns
    namespace: kube-system
  data:
    federations: abc=def
```
2016-11-18 23:03:54 -08:00
Kubernetes Submit Queue
8588571eca Merge pull request #36841 from colemickens/colemickens-pr-azure-multiple-ipconfig
Automatic merge from submit-queue

azure: support nics with multiple ipconfigs

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

When I initially wrote the cloudprovider, the ipconfig primary field either wasn't present or wasn't populated. Now it is and we have someone trying to use kubelet on a node with a nic with multiple ipconfigs and they ran into this.

**Which issue this PR fixes**: n/a no issue filed.

**Special notes for your reviewer**:

**Release note**:
```release-note
azure: support multiple ipconfigs on a NIC
```

If we can get this backported to 1.4.x, that would be great.
2016-11-18 22:29:01 -08:00
AdoHe
67f6754e0f kubectl resource filter convert unstructured obj before apply filter func 2016-11-19 13:03:39 +08:00
Kubernetes Submit Queue
1de096bd52 Merge pull request #37135 from feiskyer/seccomp2
Automatic merge from submit-queue

CRI: address knows issues of seccomp

For https://github.com/kubernetes/kubernetes/issues/36997.

To move forward, we decided to list seccomp profile path problem as a know issue in CRI (see [here](https://github.com/kubernetes/kubernetes/issues/36997#issuecomment-261673120)).

cc/ @yujuhong @euank @kubernetes/sig-node
2016-11-18 20:45:12 -08:00
Kubernetes Submit Queue
fa0eb19bcc Merge pull request #37114 from mikedanese/fix-bzl
Automatic merge from submit-queue

	revert revert of test/e2e/generated/BUILD
2016-11-18 17:03:38 -08:00