submit spark application java.lang.ClassNotFoundException: org.apache.spark.deploy.k8s.features.VolcanoFeatureStep

Affected Stackable version

spark operator version: 25.11.0

Affected Apache Spark-on-Kubernetes version

spark version: 3.5.7-stackable25.11.0

Current and expected behavior

提交spark任务的yaml如下
`apiVersion: spark.stackable.tech/v1alpha1
kind: SparkApplication
metadata:
name: spark-streaming
namespace: spark
spec:
sparkImage:
productVersion: 3.5.7
mode: cluster
mainClass: org.apache.spark.examples.SparkPi
mainApplicationFile: "local:///stackable/spark/examples/jars/spark-examples.jar"
args:
- "1000"
sparkConf:
spark.kubernetes.submission.waitAppCompletion: "false"
# 删除固定的 pod name，让 operator 管理
spark.kubernetes.driver.pod.name: "spark-streaming-driver"
spark.kubernetes.executor.podNamePrefix: "spark-streaming"

# --- Celeborn 配置 ---
spark.shuffle.manager: "org.apache.spark.shuffle.celeborn.SparkShuffleManager"
spark.celeborn.master.endpoints: "celeborn-master-0.celeborn-master-svc.spark.svc.cluster.local"
spark.celeborn.client.push.enabled: "true"
spark.celeborn.client.fetch.enabled: "true"

spark.kubernetes.scheduler.name: "volcano"
spark.kubernetes.driver.pod.featureSteps: "org.apache.spark.deploy.k8s.features.VolcanoFeatureStep"
spark.kubernetes.executor.pod.featureSteps: "org.apache.spark.deploy.k8s.features.VolcanoFeatureStep"
# --- 动态分配与资源 ---
spark.dynamicAllocation.enabled: "false"
spark.executor.instances: "2"
spark.executor.cores: "2"
spark.executor.memory: "1g"
spark.driver.cores: "1"
spark.driver.memory: "512m"

#在 Spark Conf 中也指定 SA，确保 Spark 内部客户端知道用哪个 SA
spark.kubernetes.driver.serviceAccountName: "spark-app-sa"
spark.kubernetes.executor.serviceAccountName: "spark-app-sa"

driver:
#使用 podOverrides 指定 K8s 原生的 serviceAccountName
podOverrides:
spec:
serviceAccountName: "spark-app-sa"
config:
resources:
cpu:
min: "1"
max: "2"
memory:
limit: "1Gi"

executor:
replicas: 1
podOverrides:
spec:
serviceAccountName: "spark-app-sa"
config:
resources:
cpu:
min: "1700m"
max: "3"
memory:
limit: "2Gi"
`
报错日志如下：

Possible solution

No response

Additional context

No response

Environment

No response

Would you like to work on fixing this bug?

None

Start with the SparkApplication YAML and the Java ClassNotFoundException for org.apache.spark.deploy.k8s.features.VolcanoFeatureStep. Check whether that feature step is available in the declared Spark 3.5.7 image and whether the Volcano-related settings are supported, then reproduce submission with the provided configuration. Done means the application submits without this class-loading failure.