Skip to content

Commit 0d6ab7f

Browse files
kannon92kannon92
committed
generate api
1 parent a8ed737 commit 0d6ab7f

File tree

12 files changed

+386
-132
lines changed

12 files changed

+386
-132
lines changed

‎config/schemapatch/kueue-operator.crd.yaml

+3-6
Original file line numberDiff line numberDiff line change
@@ -2,7 +2,7 @@ apiVersion: apiextensions.k8s.io/v1
22
kind: CustomResourceDefinition
33
metadata:
44
annotations:
5-
controller-gen.kubebuilder.io/version: v0.17.2
5+
controller-gen.kubebuilder.io/version: v0.4.1
66
name: kueues.operator.openshift.io
77
spec:
88
group: operator.openshift.io
@@ -11,7 +11,7 @@ spec:
1111
listKind: KueueList
1212
plural: kueues
1313
singular: kueue
14-
scope: Namespaced
14+
scope: Cluster
1515
versions:
1616
- name: v1alpha1
1717
schema:
@@ -175,8 +175,6 @@ spec:
175175
x-kubernetes-validations:
176176
- rule: self.matches(r'^([a-z0-9]([-a-z0-9]*[a-z0-9])?(\.[a-z0-9]([-a-z0-9]*[a-z0-9])?)*/)?([a-z0-9]([-a-z0-9]*[a-z0-9])?)$') && self.size() <= 317
177177
x-kubernetes-list-type: atomic
178-
<<<<<<< HEAD
179-
=======
180178
kueueGangSchedulingPolicy:
181179
description: |-
182180
kueueGangSchedulingPolicy controls how Kueue admits workloads.
@@ -200,7 +198,7 @@ spec:
200198
- Parallel
201199
- Sequential
202200
x-kubernetes-validations:
203-
- rule: self.policy==ByWorkload
201+
- rule: self.policy=='ByWorkload'
204202
message: byWorkload is only valid if policy equals ByWorkload
205203
policy:
206204
description: |-
@@ -249,7 +247,6 @@ spec:
249247
enum:
250248
- QueueNameRequired
251249
- QueueNameOptional
252-
>>>>>>> 22bc930 (add FairSharing, QueueLabelPolicy and GangScheduling)
253250
logLevel:
254251
description: |-
255252
logLevel is an intent based logging for an overall component. It does not give fine grained control, but it is a

‎config/schemapatch/operator.openshift.io_kueues.yaml

+1-4
Original file line numberDiff line numberDiff line change
@@ -175,8 +175,6 @@ spec:
175175
x-kubernetes-validations:
176176
- rule: self.matches(r'^([a-z0-9]([-a-z0-9]*[a-z0-9])?(\.[a-z0-9]([-a-z0-9]*[a-z0-9])?)*/)?([a-z0-9]([-a-z0-9]*[a-z0-9])?)$') && self.size() <= 317
177177
x-kubernetes-list-type: atomic
178-
<<<<<<< HEAD
179-
=======
180178
kueueGangSchedulingPolicy:
181179
description: |-
182180
kueueGangSchedulingPolicy controls how Kueue admits workloads.
@@ -200,7 +198,7 @@ spec:
200198
- Parallel
201199
- Sequential
202200
x-kubernetes-validations:
203-
- rule: self.policy==ByWorkload
201+
- rule: self.policy=='ByWorkload'
204202
message: byWorkload is only valid if policy equals ByWorkload
205203
policy:
206204
description: |-
@@ -249,7 +247,6 @@ spec:
249247
enum:
250248
- QueueNameRequired
251249
- QueueNameOptional
252-
>>>>>>> 22bc930 (add FairSharing, QueueLabelPolicy and GangScheduling)
253250
logLevel:
254251
description: |-
255252
logLevel is an intent based logging for an overall component. It does not give fine grained control, but it is a

‎deploy/crd/kueue-operator.crd.yaml

+72
Original file line numberDiff line numberDiff line change
@@ -171,6 +171,78 @@ spec:
171171
required:
172172
- frameworks
173173
type: object
174+
kueueGangSchedulingPolicy:
175+
description: |-
176+
kueueGangSchedulingPolicy controls how Kueue admits workloads.
177+
Gang Scheduling is the act of all or nothing scheduling.
178+
Kueue provides this ability.
179+
This field is optional.
180+
properties:
181+
byWorkload:
182+
description: |-
183+
byWorkload controls how admission is done.
184+
When admission is set to Sequential, only pods from the currently processing workload will be admitted.
185+
Once all pods from the current workload are admitted, and ready, Kueue will process the next workload.
186+
Sequential processing may slow down admission when the cluster has sufficient capacity for multiple workloads,
187+
but provides a higher guarantee of workloads scheduling all pods together successfully.
188+
When set to Parallel, pods from any workload will be admitted at any time.
189+
This may lead to a deadlock where workloads are in contention for cluster capacity and
190+
pods from another workload having successfully scheduled prevent pods from the current workload scheduling.
191+
enum:
192+
- Parallel
193+
- Sequential
194+
type: string
195+
x-kubernetes-validations:
196+
- message: byWorkload is only valid if policy equals ByWorkload
197+
rule: self.policy=='ByWorkload'
198+
policy:
199+
description: |-
200+
policy allows for changing the kinds of gang scheduling Kueue does.
201+
This is an optional field.
202+
The allowed values are ByWorkload and Disabled.
203+
The default value will be Disabled.
204+
ByWorkload allows for configuration how admission is performed
205+
for Kueue.
206+
enum:
207+
- ByWorkload
208+
- Disabled
209+
type: string
210+
type: object
211+
preemption:
212+
description: |-
213+
preemption is the process of evicting one or more admitted Workloads to accommodate another Workload.
214+
Kueue has classical preemption and preemption via fair sharing.
215+
properties:
216+
preemptionStrategy:
217+
description: |-
218+
preemptionStrategy are the types of preemption kueue allows.
219+
Kueue has two types of preemption: classical and fair sharing.
220+
enum:
221+
- Classical
222+
- FairSharing
223+
type: string
224+
type: object
225+
queueLabelPolicy:
226+
description: |-
227+
queueLabelPolicy controls how kueue manages workloads
228+
The default behavior of Kueue will manage workloads that have a queue-name label.
229+
This field is optional.
230+
properties:
231+
queueLabelPolicy:
232+
description: |-
233+
queueLabelPolicy controls whether or not Kueue reconciles
234+
jobs that don't set the label kueue.x-k8s.io/queue-name.
235+
The allowed values are QueueNameRequired and QueueNameOptional.
236+
If set to QueueNameRequired, then those jobs will be suspended and never started unless
237+
they are assigned a queue and eventually admitted. This also applies to
238+
jobs created before starting the kueue controller.
239+
Defaults to QueueNameRequired; therefore, those jobs are not managed and if they are created
240+
unsuspended, they will start immediately.
241+
enum:
242+
- QueueNameRequired
243+
- QueueNameOptional
244+
type: string
245+
type: object
174246
required:
175247
- integrations
176248
type: object

‎deploy/examples/ibm.yaml

+20
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,20 @@
1+
apiVersion: operator.openshift.io/v1alpha1
2+
kind: Kueue
3+
metadata:
4+
labels:
5+
app.kubernetes.io/name: kueue-operator
6+
app.kubernetes.io/managed-by: kustomize
7+
name: cluster
8+
namespace: openshift-kueue-operator
9+
spec:
10+
managementState: Managed
11+
config:
12+
integrations:
13+
frameworks:
14+
- AppWrapper
15+
kueueGangSchedulingPolicy:
16+
policy: Disabled
17+
queueLabelPolicy:
18+
queueLabelPolicy: QueueNameOptional
19+
preemption:
20+
preemptionStrategy: FairSharing

‎deploy/examples/job.yaml ‎deploy/examples/openshift.yaml

File renamed without changes.

‎deploy/examples/rhoai_kueue.yaml

+12-12
Original file line numberDiff line numberDiff line change
@@ -7,18 +7,18 @@ metadata:
77
name: cluster
88
namespace: openshift-kueue-operator
99
spec:
10+
managementState: Managed
1011
config:
1112
integrations:
1213
frameworks:
13-
- "batch/job"
14-
- "kubeflow.org/mpijob"
15-
- "ray.io/rayjob"
16-
- "ray.io/raycluster"
17-
- "jobset.x-k8s.io/jobset"
18-
- "kubeflow.org/mxjob"
19-
- "kubeflow.org/paddlejob"
20-
- "kubeflow.org/pytorchjob"
21-
- "kubeflow.org/tfjob"
22-
- "kubeflow.org/xgboostjob"
23-
externalFrameworks:
24-
- "AppWrapper.v1beta2.workload.codeflare.dev"
14+
- RayJob
15+
- RayCluster
16+
- PyTorchJob
17+
kueueGangSchedulingPolicy:
18+
policy: ByWorkload
19+
byWorkload: Parallel
20+
queueLabelPolicy:
21+
queueLabelPolicy: QueueNameRequired
22+
preemption:
23+
preemptionStrategy: Classical
24+

‎manifests/kueue-operator.crd.yaml

+72
Original file line numberDiff line numberDiff line change
@@ -171,6 +171,78 @@ spec:
171171
required:
172172
- frameworks
173173
type: object
174+
kueueGangSchedulingPolicy:
175+
description: |-
176+
kueueGangSchedulingPolicy controls how Kueue admits workloads.
177+
Gang Scheduling is the act of all or nothing scheduling.
178+
Kueue provides this ability.
179+
This field is optional.
180+
properties:
181+
byWorkload:
182+
description: |-
183+
byWorkload controls how admission is done.
184+
When admission is set to Sequential, only pods from the currently processing workload will be admitted.
185+
Once all pods from the current workload are admitted, and ready, Kueue will process the next workload.
186+
Sequential processing may slow down admission when the cluster has sufficient capacity for multiple workloads,
187+
but provides a higher guarantee of workloads scheduling all pods together successfully.
188+
When set to Parallel, pods from any workload will be admitted at any time.
189+
This may lead to a deadlock where workloads are in contention for cluster capacity and
190+
pods from another workload having successfully scheduled prevent pods from the current workload scheduling.
191+
enum:
192+
- Parallel
193+
- Sequential
194+
type: string
195+
x-kubernetes-validations:
196+
- message: byWorkload is only valid if policy equals ByWorkload
197+
rule: self.policy=='ByWorkload'
198+
policy:
199+
description: |-
200+
policy allows for changing the kinds of gang scheduling Kueue does.
201+
This is an optional field.
202+
The allowed values are ByWorkload and Disabled.
203+
The default value will be Disabled.
204+
ByWorkload allows for configuration how admission is performed
205+
for Kueue.
206+
enum:
207+
- ByWorkload
208+
- Disabled
209+
type: string
210+
type: object
211+
preemption:
212+
description: |-
213+
preemption is the process of evicting one or more admitted Workloads to accommodate another Workload.
214+
Kueue has classical preemption and preemption via fair sharing.
215+
properties:
216+
preemptionStrategy:
217+
description: |-
218+
preemptionStrategy are the types of preemption kueue allows.
219+
Kueue has two types of preemption: classical and fair sharing.
220+
enum:
221+
- Classical
222+
- FairSharing
223+
type: string
224+
type: object
225+
queueLabelPolicy:
226+
description: |-
227+
queueLabelPolicy controls how kueue manages workloads
228+
The default behavior of Kueue will manage workloads that have a queue-name label.
229+
This field is optional.
230+
properties:
231+
queueLabelPolicy:
232+
description: |-
233+
queueLabelPolicy controls whether or not Kueue reconciles
234+
jobs that don't set the label kueue.x-k8s.io/queue-name.
235+
The allowed values are QueueNameRequired and QueueNameOptional.
236+
If set to QueueNameRequired, then those jobs will be suspended and never started unless
237+
they are assigned a queue and eventually admitted. This also applies to
238+
jobs created before starting the kueue controller.
239+
Defaults to QueueNameRequired; therefore, those jobs are not managed and if they are created
240+
unsuspended, they will start immediately.
241+
enum:
242+
- QueueNameRequired
243+
- QueueNameOptional
244+
type: string
245+
type: object
174246
required:
175247
- integrations
176248
type: object

‎manifests/operator.openshift.io_kueues.yaml

+1-4
Original file line numberDiff line numberDiff line change
@@ -171,8 +171,6 @@ spec:
171171
required:
172172
- frameworks
173173
type: object
174-
<<<<<<< HEAD
175-
=======
176174
kueueGangSchedulingPolicy:
177175
description: |-
178176
kueueGangSchedulingPolicy controls how Kueue admits workloads.
@@ -196,7 +194,7 @@ spec:
196194
type: string
197195
x-kubernetes-validations:
198196
- message: byWorkload is only valid if policy equals ByWorkload
199-
rule: self.policy==ByWorkload
197+
rule: self.policy=='ByWorkload'
200198
policy:
201199
description: |-
202200
policy allows for changing the kinds of gang scheduling Kueue does.
@@ -245,7 +243,6 @@ spec:
245243
- QueueNameOptional
246244
type: string
247245
type: object
248-
>>>>>>> 22bc930 (add FairSharing, QueueLabelPolicy and GangScheduling)
249246
required:
250247
- integrations
251248
type: object

0 commit comments

Comments
 (0)