Commit Graph

424 Commits

Author SHA1 Message Date
Jeff Grafton
aee5f457db update BUILD files 2017-10-15 18:18:13 -07:00
Kubernetes Submit Queue
731f42172b Merge pull request #51266 from resouer/not-ready
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>.

Refactor node taint conditions

**What this PR does / why we need it**:
We should use `not-ready` etc as node condition taint key.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: 
fixes #51246 

**Special notes for your reviewer**:

**Release note**:

```release-note
Use `not-ready` to replace `notReady` in node condition taint keys.
```
2017-10-04 06:56:44 -07:00
Kubernetes Submit Queue
bfb7f3c2a7 Merge pull request #53135 from jsafrane/fix-predicate-counting
Automatic merge from submit-queue (batch tested with PRs 53135, 52512, 48339). 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>.

Fixed counting of unbound PVCs towards limit of attached volumes.

Count unbound PVCs to the limit of attached volumes to a node. 

When MaxPDVolumeCountPredicate is in doubt (e.g. PVC or PV is missing), it assumes the volume is attached. It should assume the same when it encounters an unbound PVC. In any case, it should not return an error, it would stop scheduling all pods with a PVC.

Fixes: #53134

```release-note
NONE
```
2017-10-03 13:08:13 -07:00
Jan Safranek
2caae38d32 Fixed counting of unbound PVCs towards limit of attached volumes.
There are two ways how a scheduled pod can get its PVCs unbound:
- admin forcefuly unbinds it
- user deletes original PVC that was bound when the pod was scheduled and
  creates a new one with the same name that does not get bound from some
  reason.

In both cases we don't know where the original PVC pointed at and if we
should account it to the limit of attached AWS EBS / GCE PDs etc.

The common pattern here is to count it in when in doubt.
2017-10-02 15:49:34 +02:00
Suraj Deshmukh
bf25a2753b remove unused function addStorageLimit
In package `plugin/pkg/scheduler/algorithm/predicates` remove unused
unexported function `addStorageLimit`.
2017-09-29 12:17:52 +05:30
Kubernetes Submit Queue
6e9012c94c Merge pull request #52137 from aveshagarwal/master-scheduler-resources-consolidation
Automatic merge from submit-queue (batch tested with PRs 51902, 52718, 52687, 52137, 52697). 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>..

Consolidate extended resources and hugepages in Scheduler

Fixes https://github.com/kubernetes/kubernetes/issues/51732

@bsalamat @derekwaynecarr @sjenning @kubernetes/sig-scheduling-pr-reviews
2017-09-23 19:49:54 -07:00
Kubernetes Submit Queue
507e21e7b9 Merge pull request #51889 from guangxuli/pod_affinity_error
Automatic merge from submit-queue (batch tested with PRs 52168, 48939, 51889, 52051, 50396). 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 specific errors for pod affinity predicates

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

Add specific error for pod affinity predicates

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

fix https://github.com/kubernetes/kubernetes/issues/51655

**Special notes for your reviewer**:
none
**Release note**:
none
2017-09-23 13:40:50 -07:00
Avesh Agarwal
ae05a6da34 Consolidate extended resources and hugepages in Scheduler. 2017-09-18 09:32:16 -04:00
Harry Zhang
2afab02349 Use NC to fix deprecated taint key name 2017-09-15 23:54:15 +08:00
Harry Zhang
221db2bd6b Refactor node taint conditions 2017-09-15 14:34:11 +08:00
Gavin
a724a0fcdc Add specific errors for pod affinity predicates 2017-09-13 23:47:58 +08:00
Kubernetes Submit Queue
83b4c0ac84 Merge pull request #48226 from wongma7/pd-predicate-log
Automatic merge from submit-queue (batch tested with PRs 48226, 52046, 52231, 52344, 52352)

Log get PVC/PV errors in MaxPD predicate only at high verbosity

The error is effectively ignored since even if a PVC/PV doesn't exist it gets counted, and it's rarely actionable either so let's reduce the verbosity.

Basically a user somewhere on the cluster will have to have done something "wrong" for this error to occur, e.g. if *,while the pod is running, pod's PVC is deleted or pods' PVC's PV is deleted. And from that point forward the logs will be spammed every time the predicate is evaluated on a node where that "wrong" pod exists

**Release note**:

```release-note
NONE
```
2017-09-12 19:39:52 -07:00
Kubernetes Submit Queue
0ae98b6ffe Merge pull request #52146 from resouer/eclass-fix
Automatic merge from submit-queue

Note equivalence class for dev and other fix

**What this PR does / why we need it**:
1. Add a note for predicate developers to respect equivalence class design
2. Add comments and re-ordered the related data structure, ref https://github.com/kubernetes/community/pull/1031
3. Fix some nits (typo, code length etc)


**Special notes for your reviewer**:

**Release note**:

```release-note
Scheduler predicate developer should respect equivalence class cache
```
2017-09-12 04:36:10 -07:00
gmarek
b152d08ba4 fix condition-taint labels 2017-09-11 10:19:51 +02:00
Harry Zhang
71babd1496 Note equivalence class for dev and other fix 2017-09-09 22:35:16 +08:00
Kubernetes Submit Queue
f695a3120a Merge pull request #50949 from bsalamat/preemption_eviction
Automatic merge from submit-queue

Add pod preemption to the scheduler

**What this PR does / why we need it**:
This is the last of a series of PRs to add priority-based preemption to the scheduler. This PR connects the preemption logic to the scheduler workflow.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #48646

**Special notes for your reviewer**:
This PR includes other PRs which are under review (#50805, #50405, #50190). All the new code is located in 43627afdf9.

**Release note**:

```release-note
Add priority-based preemption to the scheduler.
```

ref/ #47604

/assign @davidopp 

@kubernetes/sig-scheduling-pr-reviews
2017-09-08 14:19:42 -07:00
Bobby (Babak) Salamat
4a08dff168 Add pod eviction logic for scheduler preemption
Add Preempt to scheduler interface
Add preemption to the scheduling workflow
Minor changes to the scheduler integration test library
2017-09-07 15:31:55 -07:00
Derek Carr
41a4e2ccad Scheduler support for hugepages 2017-09-05 09:46:08 -04:00
saadali
3b834cf665 Modify VolumeZonePredicate to handle multi-zone PV
Modifies the VolumeZonePredicate to handle a PV that belongs to more
then one zone or region. This is indicated by the zone or region label
value containing a comma separated list.
2017-09-04 20:13:32 -07:00
Jing Xu
4d6da1fd9a Change SizeLimit to a pointer
This PR fixes issue #50121
2017-08-30 11:50:35 -07:00
Bobby (Babak) Salamat
20931d7a68 Add specific types for PredicateMetadata and PredicateMetadataProducer 2017-08-30 10:32:09 -07:00
Bobby (Babak) Salamat
a4edc6c871 Add preemption victim selector logic to scheduler 2017-08-30 10:32:09 -07:00
sakeven
0420c8f048 [Scheduler] regroup packages
Signed-off-by: sakeven <jc5930@sina.cn>
2017-08-30 11:14:44 +08:00
Kubernetes Submit Queue
4ba2b625c5 Merge pull request #50805 from bsalamat/preemption_metacompute
Automatic merge from submit-queue

Add support to modify precomputed predicate metadata upon adding/removal of a pod

**What this PR does / why we need it**: This PR adds capability to change precomputed predicate metadata and let's us add/remove pods to the precomputed metadata efficiently without the need ot recomputing everything upon addition/removal of pods. This PR is needed as a part of adding preemption logic to the scheduler.

**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**:
To make the review process a bit easier, there are three commits. The cleanup commit is only moving code and renaming some functions, without logic changes.

**Release note**:

```release-note
NONE
```
ref/ #47604
ref/ #48646

/assign @wojtek-t 

@kubernetes/sig-scheduling-pr-reviews @davidopp
2017-08-28 05:11:19 -07:00
Kubernetes Submit Queue
daf591c193 Merge pull request #51117 from k82cn/k8s_50360_2
Automatic merge from submit-queue

Moved node condition filter into a predicates.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #50360

**Release note**:

```release-note
A new predicates, named 'CheckNodeCondition', was added to replace node condition filter. 'NetworkUnavailable', 'OutOfDisk' and 'NotReady' maybe reported as a reason when failed to schedule pods.
```
2017-08-28 01:22:27 -07:00
Bobby (Babak) Salamat
87d406569d bazel update 2017-08-28 00:12:46 -07:00
Bobby (Babak) Salamat
264ca7d158 Add support to recompute partial predicate metadata upon adding/removing pods 2017-08-28 00:12:46 -07:00
Klaus Ma
18dc690c7c Moved node condition filter into a predicates. 2017-08-26 09:08:07 +08:00
NickrenREN
c95ecbc7e8 Local storage does not manage overlay any more 2017-08-25 09:56:36 +08:00
Kubernetes Submit Queue
178a5ff314 Merge pull request #50665 from xiangpengzhao/hardcode-to-const
Automatic merge from submit-queue (batch tested with PRs 50257, 50247, 50665, 50554, 51077)

Replace hard-code "cpu" and "memory" to consts

**What this PR does / why we need it**:
There are many places using hard coded "cpu" and "memory" as resource name. This PR replace them to consts.

**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**:
/kind cleanup

**Release note**:

```release-note
NONE
```
2017-08-23 02:35:09 -07:00
Kubernetes Submit Queue
cb8ade18c6 Merge pull request #50950 from k82cn/revert_50360
Automatic merge from submit-queue

Revert #50362.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: part of #50884

**Release note**:

```release-note
None
```
2017-08-21 16:50:53 -07:00
Kubernetes Submit Queue
0f8eaa45dd Merge pull request #49976 from aveshagarwal/master-pod-affinities-topology-key
Automatic merge from submit-queue (batch tested with PRs 50531, 50853, 49976, 50939, 50607)

Do not allow empty topology key for pod affinities.

**What this PR does / why we need it**:
This PR do not allow empty topology key for all 4 pod affinities.

**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
Affinity in annotations alpha feature is no longer supported in 1.8. Anyone upgrading from 1.7 with AffinityInAnnotation feature enabled must ensure pods (specifically with pod anti-affinity PreferredDuringSchedulingIgnoredDuringExecution) with empty TopologyKey fields must be removed before upgrading to 1.8.
```
@kubernetes/sig-scheduling-bugs  @bsalamat @davidopp
2017-08-21 15:46:20 -07:00
Klaus Ma
df3a699069 Revert #50362. 2017-08-19 10:24:50 +08:00
Connor Doyle
630af5422b OIR predicate includes namespaced resources. 2017-08-16 15:29:24 -07:00
xiangpengzhao
1c4dbcf5ca Replace hard-code "cpu" and "memory" to consts 2017-08-16 16:37:50 +08:00
Kubernetes Submit Queue
2820b45caa Merge pull request #50362 from k82cn/k8s_50360
Automatic merge from submit-queue

Moved node condition filter into a predicates.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #50360 

**Release note**:
```release-note
A new predicates, named 'CheckNodeCondition', was added to replace node condition filter. 'NetworkUnavailable', 'OutOfDisk' and 'NotReady' maybe reported as a reason when failed to schedule pods.
```
2017-08-12 15:10:31 -07:00
Jeff Grafton
a7f49c906d Use buildozer to delete licenses() rules except under third_party/ 2017-08-11 09:32:39 -07:00
Jeff Grafton
33276f06be Use buildozer to remove deprecated automanaged tags 2017-08-11 09:31:50 -07:00
Klaus Ma
e9738c0ce6 Moved node condition filter into a predicates. 2017-08-11 16:43:33 +08:00
Kubernetes Submit Queue
90a45b2df3 Merge pull request #49547 from k82cn/k8s_42001_0
Automatic merge from submit-queue (batch tested with PRs 50119, 48366, 47181, 41611, 49547)

Task 0: Added node taints labels and feature flags

**What this PR does / why we need it**:
Added node taint const for node condition.

**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: part of #42001

**Release note**:
```release-note
None
```
2017-08-04 14:29:42 -07:00
Harry Zhang
a0787358b5 Cover get equivalence cache in core
Fix testing method
2017-08-03 23:03:52 +08:00
Klaus Ma
c8ecd92269 Moved node condition check into Predicats. 2017-08-03 15:39:11 +08:00
Avesh Agarwal
0dad8dd459 Do not allow empty topology key for pod affinities. 2017-08-02 09:41:29 -04:00
Klaus Ma
ec4aa192cc Added taints node by condition feature flag. 2017-07-31 19:30:34 +08:00
Klaus Ma
c85e4dc1de Added node taints labels. 2017-07-25 15:21:51 +08:00
Kubernetes Submit Queue
e623fed778 Merge pull request #48636 from jingxu97/July/allocatable
Automatic merge from submit-queue (batch tested with PRs 48636, 49088, 49251, 49417, 49494)

Fix issues for local storage allocatable feature

This PR fixes the following issues:
1. Use ResourceStorageScratch instead of ResourceStorage API to represent
local storage capacity
2. In eviction manager, use container manager instead of node provider
(kubelet) to retrieve the node capacity and reserved resources. Node
provider (kubelet) has a feature gate so that storagescratch information
may not be exposed if feature gate is not set. On the other hand,
container manager has all the capacity and allocatable resource
information.

This PR fixes issue #47809
2017-07-24 19:30:33 -07:00
xiangpengzhao
d9d3396566 Remove affinity annotations leftover 2017-07-18 19:42:52 +08:00
Kubernetes Submit Queue
9995212ed3 Merge pull request #48869 from sakeven/rm_error
Automatic merge from submit-queue

[Scheduler] Remove error since err is always nil

Signed-off-by: sakeven <jc5930@sina.cn>



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

No need to log error since err is always nil.

**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**:

```
NONE
```
2017-07-17 23:08:39 -07:00
Jacob Simpson
29c1b81d4c Scripted migration from clientset_generated to client-go. 2017-07-17 15:05:37 -07:00
Kubernetes Submit Queue
a9afb931d4 Merge pull request #48805 from sakeven/use_const
Automatic merge from submit-queue (batch tested with PRs 48262, 48805)

[Scheduler] Use const value maxPriority instead of immediate value 10

Signed-off-by: sakeven <jc5930@sina.cn>



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

Use const value maxPriority instead of immediate value 10.

**Special notes for your reviewer**:

**Release note**:

```
NONE
```
2017-07-17 04:40:53 -07:00