Yikun commented on a change in pull request #35776: URL: https://github.com/apache/spark/pull/35776#discussion_r822248403
########## File path: resource-managers/kubernetes/core/src/test/resources/executor-podgroup-template.yml ########## @@ -0,0 +1,25 @@ +# +# Licensed to the Apache Software Foundation (ASF) under one or more +# contributor license agreements. See the NOTICE file distributed with +# this work for additional information regarding copyright ownership. +# The ASF licenses this file to You under the Apache License, Version 2.0 +# (the "License"); you may not use this file except in compliance with +# the License. You may obtain a copy of the License at +# +# http://www.apache.org/licenses/LICENSE-2.0 +# +# Unless required by applicable law or agreed to in writing, software +# distributed under the License is distributed on an "AS IS" BASIS, +# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +# See the License for the specific language governing permissions and +# limitations under the License. +# +apiVersion: scheduling.volcano.sh/v1beta1 +kind: PodGroup +spec: + minMember: 1000 + minResources: + cpu: "4" + memory: "16Gi" + priorityClassName: executor-priority + queue: executor-queue Review comment: nit: new line ########## File path: resource-managers/kubernetes/core/src/main/scala/org/apache/spark/deploy/k8s/features/VolcanoFeatureStep.scala ########## @@ -43,19 +44,27 @@ private[spark] class VolcanoFeatureStep extends KubernetesDriverCustomFeatureCon } override def getAdditionalPreKubernetesResources(): Seq[HasMetadata] = { - val podGroup = new PodGroupBuilder() - .editOrNewMetadata() - .withName(podGroupName) - .withNamespace(namespace) - .endMetadata() - .editOrNewSpec() - .endSpec() + val client = new DefaultVolcanoClient - queue.foreach(podGroup.editOrNewSpec().withQueue(_).endSpec()) + val template = if (kubernetesConf.isInstanceOf[KubernetesDriverConf]) { + kubernetesConf.get(KUBERNETES_DRIVER_PODGROUP_TEMPLATE_FILE) + } else { + kubernetesConf.get(KUBERNETES_EXECUTOR_PODGROUP_TEMPLATE_FILE) + } + val pg = template.map(client.podGroups.load(_).get).getOrElse(new PodGroup()) + var metadata = pg.getMetadata + if (metadata == null) metadata = new ObjectMeta + metadata.setName(podGroupName) + metadata.setNamespace(namespace) + pg.setMetadata(metadata) - priorityClassName.foreach(podGroup.editOrNewSpec().withPriorityClassName(_).endSpec()) + var spec = pg.getSpec + if (spec == null) spec = new PodGroupSpec + queue.foreach(spec.setQueue(_)) + priorityClassName.foreach(spec.setPriorityClassName(_)) Review comment: I have only one left conern in priorityClassName about configuration overwrite priority. Compare to pod template this might be a reasonable overwirte sort for me: - Top 1 configuration value (like queue) - Top 2 config template value (like driver/executor template) - Top 3 default value/feature (like priorityClassName which reuse pod priority as default value) So do you think we should change it to: ``` // Overwrite if queue configuration specified queue.foreach(spec.setQueue(_)) // Reuse Pod priority if priority is not set in template if (spec.getPriorityClassName == null) priorityClassName.foreach(spec.setPriorityClassName(_)) ``` Then we can use config template to overwrite priority the default feature step value if user needed. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
