when using spark with volcano ,queue allocated exceeds capability about volcano HOT 13 CLOSED

Maybe I know why. The overused check is only supported in proportion plugin, capacity plugin has no overused check.
Be relative to resource compare only consider requested dimension #3537
proportion plugin . The example I tested with nginx works, but the spark native task still doesn't

I did not use the spark official website configuration
remove
"--conf spark.kubernetes.scheduler.volcano.podGroupTemplateFile=/home/hadoop/yt/volcano/podgroup-test.yaml
--conf spark.kubernetes.driver.pod.featureSteps=org.apache.spark.deploy.k8s.features.VolcanoFeatureStep
--conf spark.kubernetes.executor.pod.featureSteps=org.apache.spark.deploy.k8s.features.VolcanoFeatureStep"
add
" --conf spark.kubernetes.scheduler.name=volcano
--conf spark.kubernetes.driver.annotation.scheduling.volcano.sh/queue-name=yt-queue
--conf spark.kubernetes.executor.annotation.scheduling.volcano.sh/queue-name=yt-queue "
Meet expectations

from volcano.

Please paste your config.

from volcano.

请粘贴您的配置。
test_queue.yaml:
apiVersion: scheduling.volcano.sh/v1beta1
kind: Queue
metadata:
name: test
spec:
reclaimable: false
weight: 1
capability:
cpu: "5"
memory: "10240Mi"

podgroup.yaml
apiVersion: scheduling.volcano.sh/v1beta1
kind: PodGroup
metadata:
generation: 5
name: test
namespace: pjl
spec:
minMember: 1
minResources:
cpu: "3"
memory: "6144Mi"
queue: test

spark-shell:
/data/module/spark-3.3.2-bin-hadoop3/bin/spark-submit
--master k8s://https://xxxxx:6443
--class org.apache.spark.examples.streaming.NetworkWordCount
--name spark-wc-test
--deploy-mode cluster
--conf spark.executor.memory=2g
--conf spark.executor.cores=2
--conf spark.driver.memory=1g
--conf spark.kubernetes.container.image=spark:v3.3.2-vol-rc02
--conf spark.kubernetes.authenticate.driver.serviceAccountName=spark
--conf spark.kubernetes.namespace=pjl
--conf spark.kubernetes.executor.podNamePrefix=sparkwc
--conf spark.kubernetes.kubeconfig=~/.kube/config
--conf spark.kubernetes.scheduler.volcano.podGroupTemplateFile=/home/hadoop/yt/volcano/podgroup-test.yaml
--conf spark.kubernetes.driver.pod.featureSteps=org.apache.spark.deploy.k8s.features.VolcanoFeatureStep
--conf spark.kubernetes.executor.pod.featureSteps=org.apache.spark.deploy.k8s.features.VolcanoFeatureStep
local:///opt/spark/examples/jars/spark-examples_2.12-3.3.2.jar csprod01 12138

What seems to be the problem?

from volcano.

请粘贴您的配置。
volcano-scheduler-configmap is default:
apiVersion: v1
data:
volcano-scheduler.conf: |
actions: "enqueue, allocate, backfill"
tiers:
- plugins:
- name: priority
- name: gang
enablePreemptable: false
- name: conformance
- plugins:
- name: overcommit
- name: drf
enablePreemptable: false
- name: predicates
- name: proportion
- name: nodeorder
- name: binpack
kind: ConfigMap
metadata:
annotations:
meta.helm.sh/release-name: volcano
meta.helm.sh/release-namespace: pjl
creationTimestamp: "2024-06-27T07:18:58Z"
labels:
app.kubernetes.io/managed-by: Helm
name: volcano-scheduler-configmap
namespace: pjl
resourceVersion: "35749706"
uid: fbd974b1-d45f-4d94-94f6-bded38e0b175

from volcano.

I think that scheduler conf will be more useful - actions and tiers with plugins (You can find them in config map in Volcano namespace).

from volcano.

I think that scheduler conf will be more useful - actions and tiers with plugins (You can find them in config map in Volcano namespace).

That's it
volcano-scheduler-configmap is default:
apiVersion: v1
data:
volcano-scheduler.conf: |
actions: "enqueue, allocate, backfill"
tiers:

• plugins:
• name: priority
• name: gang
  enablePreemptable: false
• name: conformance
• plugins:
• name: overcommit
• name: drf
  enablePreemptable: false
• name: predicates
• name: proportion
• name: nodeorder
• name: binpack
  kind: ConfigMap
  metadata:
  annotations:
  meta.helm.sh/release-name: volcano
  meta.helm.sh/release-namespace: pjl
  creationTimestamp: "2024-06-27T07:18:58Z"
  labels:
  app.kubernetes.io/managed-by: Helm
  name: volcano-scheduler-configmap
  namespace: pjl
  resourceVersion: "35749706"
  uid: fbd974b1-d45f-4d94-94f6-bded38e0b175

from volcano.

I use the exact same configuration : docs/user-guide/how_to_use_capacity_plugin.md
but it still didn't work as described in the documentation。

from volcano.

Maybe I know why. The overused check is only supported in proportion plugin, capacity plugin has no overused check.

Be relative to resource compare only consider requested dimension #3537

from volcano.

Maybe I know why. The overused check is only supported in proportion plugin, capacity plugin has no overused check.

Be relative to resource compare only consider requested dimension #3537

But I tried both plugins separately, and neither worked.
I checked my configurations according to the user-guide and the website and couldn't find the problem.
Did I have to configure all the queues to be filled with k8s resources？ But that's not what I expected

from volcano.

Maybe I know why. The overused check is only supported in proportion plugin, capacity plugin has no overused check.

Be relative to resource compare only consider requested dimension #3537
proportion plugin . The example I tested with nginx works, but the spark native task still doesn't

from volcano.

Maybe I know why. The overused check is only supported in proportion plugin, capacity plugin has no overused check.
Be relative to resource compare only consider requested dimension #3537
proportion plugin . The example I tested with nginx works, but the spark native task still doesn't

I did not use the spark official website configuration remove "--conf spark.kubernetes.scheduler.volcano.podGroupTemplateFile=/home/hadoop/yt/volcano/podgroup-test.yaml --conf spark.kubernetes.driver.pod.featureSteps=org.apache.spark.deploy.k8s.features.VolcanoFeatureStep --conf spark.kubernetes.executor.pod.featureSteps=org.apache.spark.deploy.k8s.features.VolcanoFeatureStep" add " --conf spark.kubernetes.scheduler.name=volcano --conf spark.kubernetes.driver.annotation.scheduling.volcano.sh/queue-name=yt-queue --conf spark.kubernetes.executor.annotation.scheduling.volcano.sh/queue-name=yt-queue " Meet expectations
use the spark official website configuration is ok
/data/module/spark-3.3.2-bin-hadoop3/bin/spark-submit
--master k8s://https://xxxxx:6443
--class org.apache.spark.examples.streaming.NetworkWordCount
--name spark-wc-test1
--deploy-mode cluster
--conf spark.executor.memory=2g
--conf spark.executor.cores=2
--conf spark.driver.memory=1g
--conf spark.kubernetes.container.image=spark:v3.3.2-dynamic-vol-rc02
--conf spark.kubernetes.authenticate.driver.serviceAccountName=spark
--conf spark.kubernetes.namespace=pjl
--conf spark.kubernetes.executor.podNamePrefix=sparkwc
--conf spark.kubernetes.kubeconfig=~/.kube/config
--conf spark.kubernetes.scheduler.name=volcano
--conf spark.kubernetes.scheduler.volcano.podGroupTemplateFile=/home/hadoop/yt/volcano/podgroup-test.yaml
--conf spark.kubernetes.driver.pod.featureSteps=org.apache.spark.deploy.k8s.features.VolcanoFeatureStep
--conf spark.kubernetes.executor.pod.featureSteps=org.apache.spark.deploy.k8s.features.VolcanoFeatureStep
local:///opt/spark/examples/jars/spark-examples_2.12-3.3.2.jar csprod01 12138
I missed it earlier:
--conf spark.kubernetes.scheduler.name=volcano \

from volcano.

/close

from volcano.

@Monokaix: Closing this issue.

from volcano.