Apache spark 错误H2O群集的大小应为3，但为2

我正在尝试使用中的步骤在Kubernetes上运行H2O SW

我启动了一个测试软件应用程序

$ bin/spark-submit \
--master k8s://$KUBERNETES_ENDPOINT \
--deploy-mode cluster \
--class ai.h2o.sparkling.InitTest \
--conf spark.scheduler.minRegisteredResourcesRatio=1 \
--conf spark.kubernetes.container.image=h2oai/sparkling-water-scala:3.32.0.2-1-2.4 \
--conf spark.executor.instances=3 \
local:///opt/sparkling-water/tests/initTest.jar

似乎UI流正在正确运行，因为我可以在执行此操作后访问它

$ kubectl port-forward ai-h2o-sparkling-inittest-1606331533023-driver 54322:54322

当查看创建的SparklingWater吊舱的日志时，我看到以下内容

$ kubectl logs ai-h2o-sparkling-inittest-1606331533023-driver

20/11/25 19:14:14 INFO SignalUtils: Registered signal handler for INT
20/11/25 19:14:22 INFO Server: jetty-9.4.z-SNAPSHOT; built: 2018-06-05T18:24:03.829Z; git: d5fc0523cfa96bfebfbda19606cad384d772f04c; jvm 1.8.0_275-b01
20/11/25 19:14:23 INFO ContextHandler: Started a.h.o.e.j.s.ServletContextHandler@5af7a7{/,null,AVAILABLE}
20/11/25 19:14:23 INFO AbstractConnector: Started ServerConnector@63f4e498{HTTP/1.1,[http/1.1]}{0.0.0.0:54321}
20/11/25 19:14:23 INFO Server: Started @90939ms
20/11/25 19:14:23 INFO RestApiUtils: H2O node http://10.244.1.4:54321/3/Cloud successfully responded for the GET.
20/11/25 19:14:23 INFO H2OContext: Sparkling Water 3.32.0.2-1-2.4 started, status of context: 
Sparkling Water Context:
 * Sparkling Water Version: 3.32.0.2-1-2.4
 * H2O name: root
 * cluster size: 2
 * list of used nodes:
  (executorId, host, port)
  ------------------------
  (0,10.244.1.4,54321)
  (1,10.244.0.10,54321)
  ------------------------

  Open H2O Flow in browser: http://ai-h2o-sparkling-inittest-1606331533023-driver-svc.default.svc:54321 (CMD + click in Mac OSX)

     
Exception in thread "main" java.lang.RuntimeException: H2O cluster should be of size 3 but is 2
    at ai.h2o.sparkling.InitTest$.main(InitTest.scala:34)
    at ai.h2o.sparkling.InitTest.main(InitTest.scala)
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:498)
    at org.apache.spark.deploy.JavaMainApplication.start(SparkApplication.scala:52)
    at org.apache.spark.deploy.SparkSubmit.org$apache$spark$deploy$SparkSubmit$$runMain(SparkSubmit.scala:845)
    at org.apache.spark.deploy.SparkSubmit.doRunMain$1(SparkSubmit.scala:161)
    at org.apache.spark.deploy.SparkSubmit.submit(SparkSubmit.scala:184)
    at org.apache.spark.deploy.SparkSubmit.doSubmit(SparkSubmit.scala:86)
    at org.apache.spark.deploy.SparkSubmit$$anon$2.doSubmit(SparkSubmit.scala:920)
    at org.apache.spark.deploy.SparkSubmit$.main(SparkSubmit.scala:929)
    at org.apache.spark.deploy.SparkSubmit.main(SparkSubmit.scala)

当查看SW创建的POD时，我看到一个处于挂起状态（从不进入运行状态）

---

有没有办法解决这个问题？

这似乎是因为k8s集群没有足够的CPU（这是一个小型集群）

启动SW时减少执行器的数量（从3个减少到2个）修复了该问题

bin/spark-submit \      
--master k8s://$KUBERNETES_ENDPOINT \
--deploy-mode cluster \
--class ai.h2o.sparkling.InitTest \
--conf spark.scheduler.minRegisteredResourcesRatio=1 \
--conf spark.kubernetes.container.image=h2oai/sparkling-water-scala:3.32.0.2-1-2.4 \
--conf spark.executor.instances=2 \
local:///opt/sparkling-water/tests/initTest.jar

bin/spark-submit \      
--master k8s://$KUBERNETES_ENDPOINT \
--deploy-mode cluster \
--class ai.h2o.sparkling.InitTest \
--conf spark.scheduler.minRegisteredResourcesRatio=1 \
--conf spark.kubernetes.container.image=h2oai/sparkling-water-scala:3.32.0.2-1-2.4 \
--conf spark.executor.instances=2 \
local:///opt/sparkling-water/tests/initTest.jar