Commit Graph

40976 Commits (def802272904c04f0f3f736c8616bc68fb1edf8f)

Author SHA1 Message Date
Kubernetes Submit Queue def8022729 Merge pull request #37100 from apatil/master
Automatic merge from submit-queue

Adding NAMESPACE to elasticsearch pod environment from downward API

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

The elasticsearch cluster addon references the `NAMESPACE` environmental variable here: https://github.com/kubernetes/kubernetes/blob/master/cluster/addons/fluentd-elasticsearch/es-image/elasticsearch_logging_discovery.go#L56 but, on k8s 1.3.7 at least, it is not automatically set. This PR gets it from the downward API.

**Release note**:

```NONE
```
2016-12-19 09:38:42 -08:00
Kubernetes Submit Queue ce67a03b81 Merge pull request #38958 from jszczepkowski/ha-validation-fix
Automatic merge from submit-queue (batch tested with PRs 38942, 38958)

Added MULTIZONE flag to e2e remove master script.

Added MULTIZONE flag to e2e remove master script. The script is used by HA tests which set-up multizone cluster.
2016-12-19 08:45:45 -08:00
Kubernetes Submit Queue ac36b0eae3 Merge pull request #38942 from brendandburns/flake
Automatic merge from submit-queue (batch tested with PRs 38942, 38958)

Refactor port allocation logic a little, deflake tests.

This should fix #38323 for real.

@xiangpengzhao @rmmh @justinsb 

Switch to manually cranking over the port filling logic, since there was a race between the allocate logic and the check to see if the port was freed up.
2016-12-19 08:45:43 -08:00
Kubernetes Submit Queue d6046aab0e Merge pull request #38903 from deads2k/cli-16-fix-generic-create
Automatic merge from submit-queue (batch tested with PRs 38080, 38903)

prevent negotation on connections that dont' require it

Not all clientconfigs require negotiation.  Before there was a patch (hardcoded to an instance variable) that allowed this.  This updates the code to allow the same as before.

@kubernetes/sig-cli @fabianofranz

@ncdc you'll want to pick
2016-12-19 07:42:45 -08:00
Kubernetes Submit Queue 40bed8e189 Merge pull request #38080 from kargakis/requeue-on-selector-updates
Automatic merge from submit-queue

controller: sync deployments once they don't overlap anymore

Fixes https://github.com/kubernetes/kubernetes/issues/34458.

@kubernetes/deployment
2016-12-19 07:31:15 -08:00
Kubernetes Submit Queue 75dfb21018 Merge pull request #38961 from wojtek-t/avoid_allocations_in_cacher
Automatic merge from submit-queue (batch tested with PRs 34353, 33837, 38878, 38961)

Reduce amount of allocations in cacher

Currently we are unnecessary copying stuff around in cacher.

This results in __tens of megabytes__ of allocations per __second__ in large clusters. This in turn results in expensive GC. Should help with metrics.

@gmarek
2016-12-19 06:43:43 -08:00
Kubernetes Submit Queue 5f82fe76a2 Merge pull request #38878 from kubernetes/revert-38780-ds-fix1
Automatic merge from submit-queue (batch tested with PRs 34353, 33837, 38878)

Revert "daemonset: bail out after we enqueue once"

I get overzealous sometimes.

Reverts kubernetes/kubernetes#38780
2016-12-19 06:43:00 -08:00
Kubernetes Submit Queue a366ca3732 Merge pull request #33837 from jsafrane/e2e-configmap
Automatic merge from submit-queue (batch tested with PRs 34353, 33837, 38878)

Add e2e test for configmap volume

There are two patches:
- refactor e2e volume tests to allow multiple volumes mounted into single pod
- add a test for ConfigMap volume mounted twice to test #28502
2016-12-19 06:42:58 -08:00
Kubernetes Submit Queue 377ad165a1 Merge pull request #34353 from copejon/gce-persistentvolume-testing
Automatic merge from submit-queue (batch tested with PRs 34353, 33837, 38878)

Gce persistentvolume testing

Add E2E PersistentVolume test for a GCE environment.  Tests that deleting a PV or PVC before the referencing pod does not fail on unmount and detach during pod deletion.
cc @jeffvance
2016-12-19 06:42:56 -08:00
Jerzy Szczepkowski 0c669fe6cc Added MULTIZONE flag to e2e remove master script.
Added MULTIZONE flag to e2e remove master script.
The script is used by HA tests which set-up multizone clusters.
2016-12-19 14:37:42 +01:00
Michail Kargakis 04c6fecbc7 controller: use defaultResync for the deployment controller 2016-12-19 14:04:15 +01:00
Michail Kargakis d19a1109e2 controller: sync deployments once they don't overlap anymore 2016-12-19 14:04:15 +01:00
Wojciech Tyczynski 457c9a2e6e Reduce amount of allocations in cacher 2016-12-19 13:51:07 +01:00
Kubernetes Submit Queue 2b0a1f2b11 Merge pull request #38945 from du2016/change-source-common
Automatic merge from submit-queue

common.sh should load before kube-down/kube-up function called

#38921 common.sh should load before kube-down/kube-up function called,now it load in  kube-down/kube-up function,so can't find verify-kube-binaries command
2016-12-19 02:35:25 -08:00
Jan Safranek 396e2bc581 Add configmap test 2016-12-19 10:33:38 +01:00
Jan Safranek 2970761cf1 Multiple volumes for e2e tests
Allow mounting of multiple volumes in volume e2e tests.
2016-12-19 10:33:38 +01:00
Kubernetes Submit Queue 16f7cd189c Merge pull request #38394 from xiang90/fix_rollback
Automatic merge from submit-queue

rollback: use 237 wal pkg

/cc @wojtek-t @mml
2016-12-19 00:30:21 -08:00
du2016 39a25edb71 common.sh should load before kube-down/kube-up function called 2016-12-19 02:50:33 -05:00
Brendan Burns 47b79de76e Refactor port allocation logic a little, deflake tests. 2016-12-18 21:18:34 -08:00
Kubernetes Submit Queue 7e1cacaa91 Merge pull request #38440 from justinsb/conversion_gen_safe_at_some_speeds
Automatic merge from submit-queue

conversion-gen: add --skip-unsafe flag

We should expose the SkipUnsafe option, for legacy compatability, so
that conversion-go can be used in other projects, and for platforms
where unsafe is not available.

Make unsafe code generation the default though, and have the help text
hint that the resulting code is sub-optimal.
2016-12-18 09:16:16 -08:00
Kubernetes Submit Queue 01b9d49673 Merge pull request #38677 from bruceauyeung/k8s-branch-fix-typos-and-comments
Automatic merge from submit-queue

path.Clean paths in GlobalEnvParams and remove unnecessary path.Join

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

1. clean all paths in `GlobalEnvParams`
1. remove unnecessary path.Join call in `pki.go`
2. fix some typos and comment errors

Signed-off-by: bruceauyeung <ouyang.qinhua@zte.com.cn>
2016-12-17 12:16:47 -08:00
Kubernetes Submit Queue 5b2823adb9 Merge pull request #38191 from sttts/sttts-move-master-options
Automatic merge from submit-queue

Move non-generic apiserver code out of the generic packages
2016-12-17 01:25:45 -08:00
Kubernetes Submit Queue a4577e70ab Merge pull request #38808 from du2016/change-heapster-version
Automatic merge from submit-queue (batch tested with PRs 38906, 38808)

change the version in the yaml file

change the version in heapster-controller.yaml with image version
2016-12-17 00:41:24 -08:00
Kubernetes Submit Queue f45fa0b76e Merge pull request #38906 from euank/coreos-owners
Automatic merge from submit-queue

cluster/gce/coreos: add OWNERS

See #33965 for context.

The code in `cluster/gce/coreos` has mostly been written/maintained by @yifan-gu and myself thusfar, so I added our names to the owner list.

@ethernetdan has also volunteered as well (thanks!).

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

cc @roberthbailey
2016-12-17 00:22:54 -08:00
bruceauyeung 02c1814113 remove unnecessary path.Join
Signed-off-by: bruceauyeung <ouyang.qinhua@zte.com.cn>
2016-12-17 11:57:00 +08:00
Kubernetes Submit Queue 9ba4a0effc Merge pull request #38908 from spxtr/wow
Automatic merge from submit-queue

Remove two zany unit tests.

These two tests aren't unit tests in the usual sense. We can consider switching them to run as verify checks, but I'm not convinced that they're even necessary.

They essentially work by searching their code for public functions with signatures that look like `FitPredicate`, then they shell out to grep to see that they're used somewhere in the source tree. This will never work in bazel.
2016-12-16 19:34:04 -08:00
Kubernetes Submit Queue 6c9a4139f3 Merge pull request #38614 from soltysh/remove_ext_jobs
Automatic merge from submit-queue (batch tested with PRs 37468, 36546, 38713, 38902, 38614)

Remove extensions/v1beta1 Job

Fixes https://github.com/kubernetes/kubernetes/issues/32763. This endpoint was deprecated in 1.5 and was planned to be removed in 1.6. 

**Release note**:
```release-note
Remove extensions/v1beta1 Jobs resource, and job/v1beta1 generator.
```
2016-12-16 18:55:37 -08:00
Kubernetes Submit Queue 9929fd3f52 Merge pull request #38902 from madhusudancs/fed-newtest-build-no-jq
Automatic merge from submit-queue (batch tested with PRs 37468, 36546, 38713, 38902, 38614)

[Federation] Remove unnecessary functions from develop.sh as part of …

This is part of a big refactor to deprecate a short-lived deploy.sh
mechanism that nobody really could use due to bugs.


```release-note
`federation/deploy/deploy.sh` was an interim solution introduced in Kubernetes v1.4 to simplify the federation control plane deployment experience. Now that we have `kubefed`, we are deprecating `deploy.sh` scripts.
```

cc @kubernetes/sig-federation-misc @irfanurrehman
2016-12-16 18:55:34 -08:00
Kubernetes Submit Queue 9a1c607673 Merge pull request #38713 from alejandroEsc/ae/localup1
Automatic merge from submit-queue (batch tested with PRs 37468, 36546, 38713, 38902, 38614)

local-up-cluster additions

**What this PR does / why we need it**:
Changes to local-cluster-up: These include: 1)  a simple additional help option. 2) additional error message to not being able to run `docker ps`. 3) fail faster when etcd is not found in path. Hopefully these make developing a bit more pleasant.

**Release note**:
```NONE
```
2016-12-16 18:55:32 -08:00
Kubernetes Submit Queue ff3d4a55b0 Merge pull request #36546 from MrHohn/e2e-firewall
Automatic merge from submit-queue (batch tested with PRs 37468, 36546, 38713, 38902, 38614)

Adds e2e firewall tests for LoadBalancer service, ingress, and e2e cluster

Fixes #25488 and fixes #31827.

This PR adds e2e firewall test for LoadBalancer type service, ingress and e2e cluster.

Test details for LoadBalancer type service as below:
- Verifies corresponding firewall rule has correct `sourceRanges`, `ports and protocols` and `target tags`.
- Verifies requests can reach all expected instances.
- Verifies requests can not reach instances that are not included.

Overview of the test procedure:
- Creates a LoadBalancer type service.
- Validates the corresponding firewall rule.
- Creates netexec pods as service backends.
- Sends requests from outside of the cluster and examine hitting all instances in range.
- Removes tags from one of the instances in order to get it out of firewall rule's range.
- Sends requests from outside of the cluster and examine not hitting this instance.
- Recovers tags for this instances and verifies its traffic is back.

@bprashanth @bowei @thockin
2016-12-16 18:55:30 -08:00
Kubernetes Submit Queue 425bae0cdf Merge pull request #37468 from bruceauyeung/k8s-branch-rename-constant-MaxPatchConflicts-to-maxRetryWhenPatchConflicts
Automatic merge from submit-queue (batch tested with PRs 37468, 36546, 38713, 38902, 38614)

rename constant MaxPatchConflicts to maxRetryWhenPatchConflicts

**What this PR does / why we need it**:
1. literally `MaxPatchConflicts` means max number of patch conflicts allowed during a patch operation. but actually in codes it is used to indicate max number of patch retry when patch conflicts happened.
2. there is no need to export this constant because it is only used in `resthandler.go`  and shouldn't be used in other packages.


Signed-off-by: bruceauyeung <ouyang.qinhua@zte.com.cn>
2016-12-16 18:55:28 -08:00
Kubernetes Submit Queue 699964c972 Merge pull request #38836 from bprashanth/kubelet_critical
Automatic merge from submit-queue

Admit critical pods in the kubelet

Haven't verified in a live cluster yet, just unittested, so applying do-not-merge label.
2016-12-16 17:21:46 -08:00
Kubernetes Submit Queue f86c0476ec Merge pull request #38884 from janetkuo/fix-broken-local-up
Automatic merge from submit-queue

Document known issue for broken local-up-cluster script in 1.5.1

Ref #38847
2016-12-16 16:41:25 -08:00
Xiang Li a4778fa6cd add README file to the forked wal pkg 2016-12-16 16:33:44 -08:00
Xiang Li 82a1aca916 fix bazel 2016-12-16 16:32:16 -08:00
Xiang Li b6939b288a rollback: use 237 wal pkg 2016-12-16 16:27:25 -08:00
Matt Liggett 69cd805532 Merge pull request #38804 from Random-Liu/disable-au
Node E2E: Disable AU in node e2e test.
2016-12-16 15:32:23 -08:00
Maciej Szulik 4188bb71d3 Update test_owners.csv 2016-12-17 00:10:38 +01:00
Maciej Szulik 43dfcd7a59 Remove extensions/v1beta1 Job - staging client 2016-12-17 00:10:35 +01:00
Maciej Szulik cdec94523e Remove extensions/v1beta1 Job - generated changes 2016-12-17 00:07:25 +01:00
Maciej Szulik 9f064c57ce Remove extensions/v1beta1 Job 2016-12-17 00:07:24 +01:00
Maciej Szulik e0ecb09fda Fix variable name conflict in clientset generation 2016-12-17 00:07:24 +01:00
Janet Kuo 142c0b86f3 Document known issue for broken local-up-cluster script in 1.5.1 2016-12-16 15:01:22 -08:00
Joe Finney b4c87a94a8 Remove two zany unit tests. 2016-12-16 14:49:05 -08:00
Zihong Zheng b43e2134a2 Adds e2e firewall tests.
For LoadBalancer type service:
- Verifies corresponding firewall rule has correct sourceRanges, ports
  & protocols, target tags.
- Verifies requests can reach all expected instances.
- Verifies requests can not reach instances that are not included.

For Ingress resrouce:
- Verifies the ingress firewall rule has correct sourceRanges, target
  tags and tcp ports.

For general e2e cluster:
- Verifies all required firewall rules has correct sourceRange, ports
  & protocols, source tags and target tags.
- Verifies well know ports on master and nodes are not
  exposed externally
2016-12-16 14:31:30 -08:00
Kubernetes Submit Queue 9010d27108 Merge pull request #38900 from bprashanth/ing_feat
Automatic merge from submit-queue

Remove a space in ingress e2e title
2016-12-16 14:15:59 -08:00
Euan Kemp 9a8c6ac41e cluster/gce/coreos: add OWNERS 2016-12-16 14:08:54 -08:00
Kubernetes Submit Queue d381ff422f Merge pull request #38895 from dashpole/e2e_node_timeout
Automatic merge from submit-queue (batch tested with PRs 38888, 38895)

InodeEviction Test failing because of docker race condition.

The inode eviciton test was failing because of a bug in docker/docker#21215.
Inode eviction test triggers garbage collection of images, which causes an error if kubernetes tries to "docker images list" at the same time.
This is not relevant to the inode eviction test, so do not cause the test to fail if this race occurs.
@Random-Liu
2016-12-16 13:35:28 -08:00
Kubernetes Submit Queue 29e7096b8e Merge pull request #38888 from madhusudancs/fed-newtest-add-develop
Automatic merge from submit-queue (batch tested with PRs 38888, 38895)

Bundle federation/develop directory in the test tarball for federation testing.

cc @kubernetes/sig-federation-misc
2016-12-16 13:35:26 -08:00
deads2k 9b507e8603 prevent negotation on connections that dont' require it 2016-12-16 16:26:55 -05:00