Commit Graph

38297 Commits

Author SHA1 Message Date
deads2k
f56cbfa8d5 add healthz to genericapiserver 2016-11-01 14:39:33 -04:00
Kubernetes Submit Queue
37122c2636 Merge pull request #35937 from zmerlynn/fix-e2e-for-taints
Automatic merge from submit-queue

e2e: Fix GetReadySchedulableNodesOrDie for taints

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

This changes framework.GetReadySchedulableNodesOrDie and
framework.GetMasterAndWorkerNodesOrDie so that nodes that can't take a
generic fake pod due to a taint/toleration mismatch aren't returned.

This is a rehash of #35210, but pulls in the scheduler code.

**Which issue this PR fixes**: c.f. #35210

**Special notes for your reviewer**: I think it's gross that we keep having to manually compute this in e2es. Maybe we need a bug for that?
2016-11-01 11:08:44 -07:00
Kubernetes Submit Queue
a02750dd8f Merge pull request #35490 from bprashanth/dockershim_probe_net
Automatic merge from submit-queue

Migrate Network Logic into runtime II

- [x] Probing
- [x] Bootstrap CIDR
- [x] Setup/Teardown
- [ ] ~~Hostports~~
- [ ] ~~BWShaping~~

https://github.com/kubernetes/kubernetes/pull/34780
2016-11-01 11:08:34 -07:00
Kubernetes Submit Queue
2244bfed81 Merge pull request #35137 from dashpole/per_container_inode_eviction
Automatic merge from submit-queue

Eviction manager evicts based on inode consumption

Fixes: #32526 Integrate Cadvisor per-container inode stats into the summary api.  Make the eviction manager act based on inode consumption to evict pods using the most inodes.

This PR is pending on a cadvisor godeps update which will be included in PR #35136
2016-11-01 10:32:09 -07:00
Zach Loafman
155a601425 e2e: Fix GetReadySchedulableNodesOrDie for taints
This changes framework.GetReadySchedulableNodesOrDie and
framework.GetMasterAndWorkerNodesOrDie so that nodes that can't take a
generic fake pod due to a taint/toleration mismatch aren't returned.

This is a rehash of #35210, but pulls in the scheduler code.
2016-11-01 09:40:02 -07:00
Kubernetes Submit Queue
1fa8369074 Merge pull request #35639 from ncdc/lister-gen
Automatic merge from submit-queue

Add tooling to generate listers

Add lister-gen tool to auto-generate listers. So far this PR only demonstrates replacing the manually-written `StoreToLimitRangeLister` with the generated `LimitRangeLister`, as it's a small and easy swap.

cc @deads2k @liggitt @sttts @nikhiljindal @lavalamp @smarterclayton @derekwaynecarr  @kubernetes/sig-api-machinery @kubernetes/rh-cluster-infra
2016-11-01 09:29:06 -07:00
Kubernetes Submit Queue
7d10cffc37 Merge pull request #32383 from sttts/sttts-sysctl-infra-only
Automatic merge from submit-queue

Only set sysctls for infra containers

We did set the sysctls for each container in a pod. This opens up a way to set un-whitelisted sysctls during upgrade from v1.3:
- set annotation in v1.3 with an un-whitelisted sysctl. Set restartPolicy=Always
- upgrade cluster to v1.4
- kill container process
- un-whitelisted sysctl is set on restart of the killed container.
2016-11-01 08:47:38 -07:00
Wojciech Tyczynski
63954ccd0e Merge pull request #35964 from wojtek-t/fix_large_density_test
Fix density test in large clusters
2016-11-01 15:13:38 +01:00
Kubernetes Submit Queue
094a00158c Merge pull request #35933 from mikedanese/lower-level
Automatic merge from submit-queue

kubeadm: lower default component logging level

v=4 is likely to causer perf issues and v=2 is the default in most other setups including scale e2e. ref https://github.com/kubernetes/kubernetes/issues/26637#issuecomment-256859739
2016-11-01 06:50:32 -07:00
Kubernetes Submit Queue
4eb1c2baa9 Merge pull request #35795 from deads2k/api-33-clean-master.go
Automatic merge from submit-queue

remove non-reuseable bits of MasterServer

Scrub `master.go` again.  I think I'm pretty happy with this shape.  I may promote `InstallAPIs` since we're likely to want it downstream.
2016-11-01 06:50:23 -07:00
Dr. Stefan Schimanski
962e7534b4 Only set sysctls for infra containers 2016-11-01 13:14:16 +01:00
Kubernetes Submit Queue
44b684ad53 Merge pull request #33663 from pmorie/selinux-fixes
Automatic merge from submit-queue

SELinux Overhaul

Overhauls handling of SELinux in Kubernetes.  TLDR: Kubelet dir no longer has to be labeled `svirt_sandbox_file_t`.

Fixes #33351 and #33510.  Implements #33951.
2016-11-01 05:04:17 -07:00
Kubernetes Submit Queue
6babfb6ccc Merge pull request #35109 from sttts/sttts-sni
Automatic merge from submit-queue

Add SNI support to the apiserver

This PR adds the `--tls-sni-key-cert` flag to the apiserver. It can be passed multiple times in the following ways:

``` shell
$ apiserver \
    --tls-sni-cert-key '*.example.com,example.com: example.key,example.crt' \
    --tls-sni-cert-key 'foo.key,foo.crt'
```

The first variant explicitly sets the accepted domain names, the second variant reads the common names and DNS names from the certificate itself.

If no domain name matches, the existing certificate (`--tls-cert-file`) is used.

``` golang
    fs.Var(config.NewNamedCertKeyArray(&s.SNICertKeys), "tls-sni-cert-key", ""+
        "A pair of x509 certificate and private key file paths, optionally prefixed with a list of "+
        "domain patterns which are fully qualified domain names, possibly with prefixed wildcard "+
        "segments. If no domain patterns are provided, the names of the certificate are "+
        "extracted. Non-wildcard matches trump over wildcard matches, explicit domain patterns "+
        "trump over extracted names. For multiple key/certificate pairs, use the "+
        "--tls-sni-key-cert multiple times. "+
        "Examples: \"example.key,example.crt\" or \"*.foo.com,foo.com:foo.key,foo.crt\".")
```

``` release-note
Add SNI support to the apiserver

Pass multiple certificates and domain name patterns with `--tls-sni-cert-key` and the right certificate will be chosen depending on the url the client is using.
```
2016-11-01 04:25:54 -07:00
Wojciech Tyczynski
b6a5992ee5 Fix density test in large clusters 2016-11-01 11:03:21 +01:00
Dr. Stefan Schimanski
7c5e354e82 Update bazel 2016-11-01 09:50:56 +01:00
Dr. Stefan Schimanski
5dc31d35a0 Add --tls-sni-cert-key to the apiserver for SNI support 2016-11-01 09:50:56 +01:00
Dr. Stefan Schimanski
d0b3981f07 Make GenericApiServer.Run interruptable and fail on first listen 2016-11-01 09:50:56 +01:00
Kubernetes Submit Queue
cc84673ebe Merge pull request #35928 from caesarxuchao/remove-clientset-archive
Automatic merge from submit-queue

Remove archived clientset

We have [client-go](https://github.com/kubernetes/client-go) now, no need to keep archived clientset in master of k8s.io/kubernetes.

Also updated the client-gen README.

Fix https://github.com/kubernetes/kubernetes/issues/35803.

cc @mikedanese 

```release-note
The main kubernetes repository stops hosting archived version of released clients. Please use [client-go](https://github.com/kubernetes/client-go).
```
2016-11-01 00:41:46 -07:00
Kubernetes Submit Queue
030c74623d Merge pull request #35764 from xilabao/break-endless-loop
Automatic merge from submit-queue

break endless loop in vagrant cluster

reference to [vagrant.md](https://github.com/kubernetes/kubernetes/blob/master/docs/devel/local-cluster/vagrant.md) troubleshooting ,  I guess we could set a time to return from endless loop to register with cloud provider.
2016-10-31 23:54:09 -07:00
Kubernetes Submit Queue
a9b183db7f Merge pull request #32557 from asalkeld/an-ingress-docs
Automatic merge from submit-queue

Correct the article in generated documents

**What this PR does / why we need it**:
Fix the article in generated docs for "create/delete [article] [kind]"

**Which issue this PR fixes**
fixes #32305

**Special notes for your reviewer**:
None

**Release note**:

``` release-note
Correct the article in generated documents
```

For example:
  "a Ingress" > "an Ingress"
2016-10-31 23:16:46 -07:00
Kubernetes Submit Queue
38a2c94f31 Merge pull request #35670 from smarterclayton/expose_unsafe_2
Automatic merge from submit-queue

Alternative unsafe copy

Have run this for 2 hours in the stresser without an error (no guarantee).

@wojtek-t can we do a 500 kubemark run with this prior to merge?
2016-10-31 22:42:55 -07:00
Kubernetes Submit Queue
3a2ae34ea9 Merge pull request #35608 from soltysh/pdb_admission
Automatic merge from submit-queue

Update PodAntiAffinity to ignore calls to subresources

@smarterclayton I hit this when I was trying to evict a pod, apparently k8s does not have this particular admission plugin on by default. ptal

@mml @davidopp fyi
2016-10-31 20:58:33 -07:00
Kubernetes Submit Queue
d461262bf8 Merge pull request #35497 from liggitt/node-address
Automatic merge from submit-queue

Allow apiserver to choose preferred kubelet address type

Follow up to #33718 to stay compatible with clusters using DNS names for master->node communications. Adds the `--kubelet-preferred-address-types` apiserver flag for clusters that prefer a different node address type.

```release-note
The apiserver can now select which type of kubelet-reported address to use for master->node communications, using the --kubelet-preferred-address-types flag.
```
2016-10-31 18:33:27 -07:00
bprashanth
2941aea88e Use podCIDR as a proxy for kubenet.Status 2016-10-31 18:05:30 -07:00
Kubernetes Submit Queue
482475e378 Merge pull request #35813 from mtaufen/bump-rkt-stage1-fly
Automatic merge from submit-queue

Bump rkt stage1-fly to 1.18.0

Bumps stage1-fly version.
2016-10-31 16:14:44 -07:00
Kubernetes Submit Queue
4bae0f3a96 Merge pull request #35927 from timstclair/summary-test
Automatic merge from submit-queue

Bump Kubelet workingset upper bound

For https://github.com/kubernetes/kubernetes/issues/34990

Follow up to https://github.com/kubernetes/kubernetes/pull/35828, because working memory is too high now too.
2016-10-31 15:34:18 -07:00
Kubernetes Submit Queue
079630a522 Merge pull request #35922 from a-robinson/crdb
Automatic merge from submit-queue

Improve CockroachDB example

This is primarily about pulling in the init container to make the config more production-ready, but I've pulled in a few other small improvements that have been made since this was first contributed.

* Use an init container to eliminate potential edge case where losing
  the first pet's data could cause it to start a second logical cluster
* Exec the cockroach binary so that it runs as PID 1 in the container
* Make some small improvements to the README

@bprashanth 

```release-note
```
2016-10-31 15:34:09 -07:00
Mike Danese
58479268e8 kubeadm: lower default component logging level 2016-10-31 15:29:10 -07:00
Fabio Yeon
d06408e441 Update CHANGELOG.md for v1.3.10. 2016-10-31 15:15:41 -07:00
Chao Xu
4cf7adf937 remove release_1_4
remove archived federation clientsets
update README
2016-10-31 15:05:26 -07:00
Michael Taufen
7a6d26bab7 Bump rkt stage1-fly to 1.18.0 2016-10-31 14:54:39 -07:00
Kubernetes Submit Queue
8b3db0bc98 Merge pull request #35915 from caesarxuchao/stop-update-client-go
Automatic merge from submit-queue

Disable update-staging-client-go until code freeze

I don't want to cause developer friction until code freeze. I'll babysit the script for the moment to keep the client-go up-to-date.

Should had done this in #34489.
2016-10-31 14:35:04 -07:00
Kubernetes Submit Queue
cbabb03acc Merge pull request #34841 from derekwaynecarr/quota-shared-informer
Automatic merge from submit-queue

quota controller uses informers if available for pod calculation

This PR does the following:
1. plumb informer factory into quota registry and evaluators
2. pod quota evaluator uses informers for determining aggregrate usage instead of making direct calls
3. admission code path does not use informers because
   1. we do not want to add new watches in apiserver
   2. admission code path does not require aggregate usage calculation

As a result, quota controller is much faster in re-calculating quota usage when it observes a pod deletion.

Follow-on PRs will make similar changes for other informer backed resources (pvcs next).

/cc @deads2k @mfojtik @smarterclayton @kubernetes/rh-cluster-infra
2016-10-31 14:34:57 -07:00
Tim St. Clair
8330b081bc
Bump Kubelet workingset upper bound 2016-10-31 13:51:07 -07:00
bprashanth
11638e9ee1 Unittests 2016-10-31 13:05:20 -07:00
bprashanth
aee5b8099c Skip legacy features (hostport, bwshaping) 2016-10-31 13:05:19 -07:00
bprashanth
7c2aeecdd2 SetUp/TearDown II: remove code from kuberuntime 2016-10-31 13:05:19 -07:00
bprashanth
706207904e SetUp/TearDown I: dockershim 2016-10-31 13:05:19 -07:00
bprashanth
48db726342 Split network.Host into LegacyHost and NamespaceGetter 2016-10-31 13:05:19 -07:00
bprashanth
9c585baf1f Teach cri about podCIDR 2016-10-31 13:05:18 -07:00
bprashanth
c11aa2eb77 Teach dockershim to probe for network plugins 2016-10-31 13:05:13 -07:00
Jordan Liggitt
1a7f7c5399
Allow apiserver to choose preferred kubelet address type 2016-10-31 16:02:38 -04:00
Alex Robinson
6b98de39a5 Improve CockroachDB example
* Use an init container to eliminate potential edge case where losing
  the first pet's could cause it to start a second logical cluster
* Exec the cockroach binary so that it runs as PID 1 in the container
* Make some small improvements to the README
2016-10-31 15:48:41 -04:00
David Ashpole
4ca7f9f972 deal with nil inodes and inodesfree 2016-10-31 12:10:23 -07:00
David Ashpole
b8fc546d60 eviction manager ecivts pod using the most inodes. 2016-10-31 11:32:49 -07:00
Chao Xu
04f291e5b4 disable update-staging-client-go until code freeze 2016-10-31 10:57:14 -07:00
Tim Hockin
e6b2517feb Merge pull request #32931 from philips/delete-kube-dns-docs-in-build
build: kube-dns: remove out of date docs
2016-10-31 18:10:35 +01:00
Kubernetes Submit Queue
dad0445bcb Merge pull request #35894 from gmarek/race
Automatic merge from submit-queue

Initialize CIDR allocator before registering handle functions

Currently we start shared informers after everything is already created, but this change make it future-proof.

cc @davidopp @kevin-wangzefeng @foxish
2016-10-31 09:28:38 -07:00
Kubernetes Submit Queue
24899597b9 Merge pull request #32687 from jouve/resource_type
Automatic merge from submit-queue

update list of vailable resources

Hi,

kubectl get --help produce a list of resource types and aliases :

```
Valid resource types include:
   * clusters (valid only for federation apiservers)
   * componentstatuses (aka 'cs')
   ...
```

``` release-note
Update the list of resources in kubectl get --help
```

The list is currently outdated (for exemple missing networkpolicy).

http://kubernetes.io/docs/user-guide/kubectl-overview/#resource-types has the same data and is also outdated.

The patch updates these 2 lists.
2016-10-31 09:28:29 -07:00
Kubernetes Submit Queue
3e3ff447d2 Merge pull request #30285 from jsafrane/external-provisioner-proposal
Automatic merge from submit-queue

Proposal for external dynamic provisioners

@kubernetes/sig-storage
2016-10-31 08:45:08 -07:00