Commit Graph

2250 Commits

Author SHA1 Message Date
Kubernetes Submit Queue
33f873dbbe
Merge pull request #55262 from liggitt/schedulercache
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Fix 'Schedulercache is corrupted' error

Fixes #50916

If an Assume()ed pod is Add()ed with a different nodeName, the podStates view of the pod is not corrected to reflect the actual nodeName. On the next Update(), the scheduler observes the mismatch and process exits.

```release-note
Fixed 'Schedulercache is corrupted' error in kube-scheduler
```
2017-11-07 23:23:33 -08:00
Jordan Liggitt
a366e6ced0
Fix 'Schedulercache is corrupted' error 2017-11-07 18:49:49 -05:00
Kubernetes Submit Queue
e1de2ad507
Merge pull request #52562 from ironcladlou/kube-scheduler-config
Automatic merge from submit-queue (batch tested with PRs 53592, 52562, 55175, 55213). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Refactor kube-scheduler config API, command, and server setup

Refactor the kube-scheduler configuration API, command setup, and server setup according to the guidelines established in #32215 and using the kube-proxy refactor (#34727) as a model of a well factored component adhering to said guidelines.

* Config API: clarify meaning and use of algorithm source by replacing modality derived from bools and string emptiness checks with an explicit AlgorithmSource type hierarchy.
* Config API: consolidate client connection config with common structs.
* Config API: split and simplify healthz/metrics server configuration.
* Config API: clarify leader election configuration.
* Config API: improve defaulting.
* CLI: deprecate all flags except `--config`.
* CLI: port all flags to new config API.
* CLI: refactor to match kube-proxy Cobra command style.
* Server: refactor away configurator.go to clarify application wiring.
* Server: refactor to more clearly separate wiring/setup from running.

Fixes https://github.com/kubernetes/kubernetes/issues/52428.

@kubernetes/api-reviewers 
@kubernetes/sig-cluster-lifecycle-pr-reviews 
@kubernetes/sig-scheduling-pr-reviews 

/cc @ncdc @timothysc @bsalamat

```release-note
The kube-scheduler command now supports a `--config` flag which is the location of a file containing a serialized scheduler configuration. Most other kube-scheduler flags are now deprecated.
```
2017-11-07 11:21:19 -08:00
Kubernetes Submit Queue
26986e8407
Merge pull request #55058 from deads2k/admission-11-split-doubles
Automatic merge from submit-queue (batch tested with PRs 53273, 55058, 55237, 50140). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

split some admission plugins into mutation and validation halves

Splits the podnodeselector, serviceaccount, and priority admission plugins into validating and mutating admission plugins.

@kubernetes/sig-api-machinery-pr-reviews
2017-11-07 09:39:39 -08:00
Kubernetes Submit Queue
d33077526a
Merge pull request #53273 from mikedanese/authtristate
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

add support for short-circuit deny in union authorizer

This change has no behavioral changes.

Fixes https://github.com/kubernetes/kubernetes/issues/51862

```release-note
Add support for the webhook authorizer to make a Deny decision that short-circuits the union authorizer and immediately returns Deny. 
```
2017-11-07 09:25:37 -08:00
Dan Mace
efb2bb71cd Refactor scheduler config API
Refactor the kube-scheduler configuration API, command setup, and server
setup according to the guidelines established in #32215 and using the
kube-proxy refactor (#34727) as a model of a well factored component
adhering to said guidelines.

* Config API: clarify meaning and use of algorithm source by replacing
modality derived from bools and string emptiness checks with an explicit
AlgorithmSource type hierarchy.
* Config API: consolidate client connection config with common structs.
* Config API: split and simplify healthz/metrics server configuration.
* Config API: clarify leader election configuration.
* Config API: improve defaulting.
* CLI: deprecate all flags except `--config`.
* CLI: port all flags to new config API.
* CLI: refactor to match kube-proxy Cobra command style.
* Server: refactor away configurator.go to clarify application wiring.
* Server: refactor to more clearly separate wiring/setup from running.

Fixes #52428.
2017-11-07 09:41:39 -05:00
Dan Mace
25ca287707 Update generated files 2017-11-07 09:41:35 -05:00
Kubernetes Submit Queue
32b761edf3
Merge pull request #54871 from anfernee/clean
Automatic merge from submit-queue (batch tested with PRs 55114, 52976, 54871, 55122, 55140). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

haveSame is suboptimal, fix it as well as the name

**What this PR does / why we need it**: Make the util function simpler, also faster by reducing a for loop.

**Special notes for your reviewer**:

**Release note**:
```release-note
None
```
2017-11-06 23:19:18 -08:00
David Eads
2c671614df split some admission plugins into mutation and validation halves 2017-11-06 09:07:00 -05:00
Kubernetes Submit Queue
aed9ad4b08
Merge pull request #54778 from stewart-yu/addexplain
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Add comment for preempt sunction.

**What this PR does / why we need it**:
Add explain for preempt sunction.

**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
```
2017-11-04 15:42:20 -07:00
Mike Danese
12125455d8 move authorizers over to new interface 2017-11-03 13:46:28 -07:00
Kubernetes Submit Queue
6fe3a4d82a
Merge pull request #55000 from deads2k/admission-10-split
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

make easy validation admission plugins into validators

This switches "easy" admission plugins that need to be validators to be validators.  I also did one simple plugin to do both mutation and validation for practice.

@kubernetes/sig-api-machinery-pr-reviews @caesarxuchao
2017-11-03 06:33:17 -07:00
David Eads
75c448dbc7 make easy validation admission plugins into validators 2017-11-03 07:54:39 -04:00
Yongkun Anfernee Gui
5d268eb1d3 haveSame is suboptimal, fix it as well as the name 2017-11-02 22:50:32 -07:00
Kubernetes Submit Queue
2084f7f4f3
Merge pull request #54488 from lichuqiang/plugin_base
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Add admission handler for device resources allocation

**What this PR does / why we need it**:
Add admission handler for device resources allocation to fail fast during pod creation

**Which issue this PR fixes** 
fixes #51592

**Special notes for your reviewer**:
@jiayingz Sorry, there is something wrong with my branch in #51895. And I think the existing comments in the PR might be too long for others to view. So I closed it and opened the new one, as we have basically reach an agreement on the implement :)
I have covered the functionality and unit test part here, and would set about the e2e part ASAP

/cc @jiayingz @vishh @RenaudWasTaken 

**Release note**:

```release-note
NONE
```
2017-11-02 17:24:06 -07:00
Kubernetes Submit Queue
40212c17cd
Merge pull request #54484 from sttts/sttts-split-psp
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

admission: wire through validating-only admission

Based on https://github.com/kubernetes/kubernetes/pull/54232.

This is important and required for beta because it affects the shape of the webhook admission plugins that are going to be produced and is needed to make sure that our existing chain continues to properly verify and protect the API objects based on their final state after webhook admission mutators run.

We discussed this in the October 11 API machinery call with @erictune and @caesarxuchao and we agreed to do this as a requirement for beta. See this part of the recording: https://www.youtube.com/watch?v=mrgDPHyr4VY#t=325 .
2017-11-02 04:02:34 -07:00
Kubernetes Submit Queue
e989ca4e63
Merge pull request #54812 from aveshagarwal/master-pod-toleration-restrictions-issues
Automatic merge from submit-queue (batch tested with PRs 54800, 53898, 54812, 54921, 53558). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Allow override of cluster level (default, whitelist) tolerations by namespace level empty (default, whitelist) tolerations.

Currently In PodTolerationRestriction admission plugin, if namespace level default and whitelist of tolerations are nil or empty, they do not override cluster level default and whitelist tolerations. 

This PR fixes the plugin to not override cluster level tolerations only when namespace level toleration are nil. IOW, if namespace level toleration are empty, they override cluster level tolerations. To be more clear, if following annotations are set to empty, they override cluster level tolerations.
 
``` 
scheduler.alpha.kubernetes.io/defaultTolerations : ""
scheduler.alpha.kubernetes.io/tolerationsWhitelist: ""
```

This behavior is inline with PodNodeSelector admission plugin too.

@sjenning @derekwaynecarr 

**Release Note**:

```release-note
In PodTolerationRestriction admisson plugin, if namespace level tolerations are empty, now they override cluster level tolerations. 
```
2017-11-02 03:14:21 -07:00
Dr. Stefan Schimanski
aedcf681b3 admission: rename Validate{ -> Initialization}, Validat{ingAdmit -> e} 2017-11-02 09:29:55 +01:00
Dr. Stefan Schimanski
2452afffe0 admission: wire create+update validation func into kube registries 2017-11-02 09:29:16 +01:00
Dr. Stefan Schimanski
74b4223ab8 admission: complete plumbing of validation admission 2017-11-02 08:45:41 +01:00
Dr. Stefan Schimanski
970d2553cc admission: { -> Mutating}Admit(admission.Attributes) 2017-11-02 08:45:41 +01:00
lichuqiang
ebd445eb8c add admission handler for device resources allocation 2017-11-02 09:17:48 +08:00
Yongkun Anfernee Gui
8465625bd7 Optimize the suboptimal image locality algorithm 2017-11-01 10:18:59 -07:00
Avesh Agarwal
13289d2218 Allow override of cluster level (default, whitelist) tolerations by namespace level empty (default, whitelist) tolerations. 2017-11-01 10:47:52 -04:00
Stewart-YU
bc2845ee08 Add explain for preempt sunction. 2017-10-31 08:27:52 +08:00
Kubernetes Submit Queue
7a944a69d6
Merge pull request #54485 from sttts/sttts-unify-admission-constructors
Automatic merge from submit-queue (batch tested with PRs 54761, 54748, 53991, 54485, 46951). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

admission: unify plugin constructors

It's common in Go to return the actual object in constructors, not **one interface**
it implements. This allows us to implement multiple interfaces, but only have
one constructor. As having private types in constructors, we export all plugin structs, of course with private fields.

Note: super interfaces do not work if there are overlapping methods.
2017-10-30 15:38:33 -07:00
Dr. Stefan Schimanski
ce6ecbbc54 Remove all api.Scheme references by using explicit package aliases 2017-10-30 19:54:02 +01:00
Kubernetes Submit Queue
9c35982e1a
Merge pull request #54691 from wackxu/priadm
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

update wrong group for priorityclasses

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

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #54690

**Special notes for your reviewer**:

assgin @bsalamat 

**Release note**:

```release-note
NONE
```
2017-10-30 10:31:03 -07:00
Dr. Stefan Schimanski
131905cdb8 admission: unify plugin constructors 2017-10-30 16:56:38 +01:00
Kubernetes Submit Queue
7c96feb298
Merge pull request #54047 from kuramal/my
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Optimize Repeated registration of AlgorithmProvider when ApplyFeatureGates

**What this PR does / why we need it**:
modified ApplyFeatureGates() just add/del features, cancel the register of all AlgorithmProvider.

 there is Repeated registration of all AlgorithmProvider when ApplyFeatureGates() runs;
AlgorithmProvider have already registered when  package defaults loaded;
I think ApplyFeatureGates() is just add/del features, it needn't  register all AlgorithmProvider again
**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
```
2017-10-29 20:42:48 -07:00
wackxu
2430c24c61 update wrong group for priorityclasses 2017-10-28 10:35:41 +08:00
Kubernetes Submit Queue
27ef37a32d
Merge pull request #54320 from derekwaynecarr/quota-update
Automatic merge from submit-queue (batch tested with PRs 54331, 54655, 54320, 54639, 54288). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Ability to do object count quota for all namespaced resources

**What this PR does / why we need it**:
- Defines syntax for generic object count quota `count/<resource>.<group>`
- Migrates existing objects to support new syntax with old syntax
- Adds support to quota all standard namespace resources 
- Updates the controller to do discovery and replenishment on those resources
- Updates unit tests
- Tweaks admission configuration around quota
- Add e2e test for replicasets (demonstrate dynamic generic counting)

```
$  kubectl create quota test --hard=count/deployments.extensions=2,count/replicasets.extensions=4,count/pods=3,count/secrets=4
resourcequota "test" created
$ kubectl run nginx --image=nginx --replicas=2
$ kubectl describe quota
Name:                         test
Namespace:                    default
Resource                      Used  Hard
--------                      ----  ----
count/deployments.extensions  1     2
count/pods                    2     3
count/replicasets.extensions  1     4
count/secrets                 1     4
```

**Special notes for your reviewer**:
- simple object count quotas no longer require writing code
- deferring support for custom resources pending investigation about how to share caches with garbage collector.  in addition, i would like to see how this integrates with downstream quota usage in openshift.

**Release note**:
```release-note
Object count quotas supported on all standard resources using `count/<resource>.<group>` syntax
```
2017-10-27 15:42:24 -07:00
Kubernetes Submit Queue
7d628db7cb
Merge pull request #54657 from anfernee/sched
Automatic merge from submit-queue (batch tested with PRs 54635, 54250, 54657, 54696, 54700). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Remove unused pods in genericScheduler

**What this PR does / why we need it**: source code cleanup. less confusion.

**Special notes for your reviewer**:

**Release note**:

```release-note
None
```
2017-10-27 14:38:26 -07:00
Derek Carr
5b4ca14307 Update admission control framework for quota 2017-10-27 11:08:14 -04:00
Kubernetes Submit Queue
3a5eab236f Merge pull request #54538 from WIZARD-CXY/updatehostPort
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

delete the hostport from usedmap

**What this PR does / why we need it**:
delete the hostport record when pod is not on the host

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

Facilitate the further pr https://github.com/kubernetes/kubernetes/pull/52421. Because the code which detects the conflict between wantports and existingports is not quite clean now.
Besides remove the unused port from map will save the memory.

**Special notes for your reviewer**:

I and the original coder @k82cn agreed to make this change

**Release note**:

```release-note
NONE
```
2017-10-27 02:18:41 -07:00
Yongkun Anfernee Gui
6ef9cf4936 Remove unused pods in genericScheduler 2017-10-26 15:23:26 -07:00
David Eads
8c1fe1f61a move webhook admission to generic apiserver 2017-10-26 07:45:49 -04:00
Kubernetes Submit Queue
32847aab1d Merge pull request #51783 from jiulongzaitian/myfeature3
Automatic merge from submit-queue (batch tested with PRs 54081, 54271, 51783, 54600, 54594). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

move getMaxVols function to predicates.go and add some NewVolumeCount…

…Predicate funcs

Signed-off-by: zhangjie <zhangjie0619@yeah.net>



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

**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
move getMaxVols function to predicates.go and add some NewVolumeCountPredicate funcs
```
2017-10-26 02:43:48 -07:00
zhangjie
968df828a0 move getMaxVols function to predicates.go and change the params of NewMaxPDVolumeCountPredicate funcs
Signed-off-by: zhangjie <zhangjie0619@yeah.net>
2017-10-26 12:11:09 +08:00
kuramal
a48cc26443 Optimize Repeated registration of AlgorithmProvider when ApplyFeatureGates
Add InsertPredicateKeyToAlgorithmProviderMap() and RemovePredicateKeyFromAlgorithmProviderMap() to insert/remove fit predicate key of all algorithmProviders which in algorithmProviderMap
Add Func RemovePredicateKeyFromAlgoProvider() AND InsertPredicateKeyToAlgoProvider() which can insert/remove fit predicate key to specific algorithmProvider
2017-10-26 10:10:48 +08:00
Kubernetes Submit Queue
17638ee018 Merge pull request #54414 from deads2k/admission-08-options
Automatic merge from submit-queue (batch tested with PRs 53760, 48996, 51267, 54414). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

update admission webhook to handle multiple auth domains

Fixes https://github.com/kubernetes/kubernetes/issues/54404

Adds some wiring to have the admission plugin accept a config file for per-apiserver configuration.

@kubernetes/sig-auth-api-reviews @deads2k @ericchiang @liggitt in particular
@kubernetes/sig-api-machinery-pr-reviews @lavalamp @caesarxuchao @sttts @cheftako

```release-note
generic webhook admission now takes a config file which describes how to authenticate to webhook servers
```
2017-10-25 17:37:11 -07:00
chenxingyu
195de933f1 delete the hostport from usedmap 2017-10-25 15:49:18 +08:00
Bobby (Babak) Salamat
ccc62412a3 Autogenerated files 2017-10-24 16:04:58 -07:00
Bobby (Babak) Salamat
5497e893ae Change type of scheduling queue from cache.FIFO to a new interface 2017-10-24 16:04:58 -07:00
Haoran Wang
4b185b4db9 update scheduler to use schedulerName selector 2017-10-24 13:31:38 +08:00
Kubernetes Submit Queue
6a444673b9 Merge pull request #53914 from bsalamat/pdb
Automatic merge from submit-queue (batch tested with PRs 53903, 53914, 54374). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Add PodDisruptionBudget to scheduler cache.

**What this PR does / why we need it**:
This is the first step to add support for PodDisruptionBudget during preemption. This PR adds PDB to scheduler cache.

**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**: None

**Release note**:

```release-note
Add PodDisruptionBudget to scheduler cache.
```

ref/ #53913
2017-10-23 12:27:06 -07:00
Kubernetes Submit Queue
fd3878a59d Merge pull request #54333 from liggitt/webhook-service-resolver
Automatic merge from submit-queue (batch tested with PRs 54363, 54333). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Ensure port on resolved service host

The resolved host should include a port so it can be used by dialers directly. It's also not necessary to reparse the URL when constructing directly.

```release-note
NONE
```
2017-10-23 11:30:25 -07:00
David Eads
fd4ab3e061 update admission webhook to handle multiple auth domains 2017-10-23 12:33:54 -04:00
Kubernetes Submit Queue
3812230f4d Merge pull request #48963 from Raffo/master
Automatic merge from submit-queue (batch tested with PRs 52792, 48963). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.

Fix range for min value in imagepolicy admission 

**What this PR does / why we need it**:
This makes the range for imagepolicy admission work properly. Currently if we specify `1s` for the value of `AllowTTL` or `DenyTTL` in the configuration yaml, we get a message like the following: 

```
Error: failed to initialize plugins: Couldn't init admission plugin "ImagePolicyWebhook": valid value is between 1s and 30m0s, got 1s
```

This is due to a wrong comparison which is fixed in this PR. The rest of the PR just adds tests for this behaviour.
2017-10-21 15:10:12 -07:00
Bobby (Babak) Salamat
efc151f46b Autogenerated files 2017-10-20 23:22:03 -07:00