Github user dvogelbacher commented on a diff in the pull request:
https://github.com/apache/spark/pull/21366#discussion_r194180839
--- Diff:
resource-managers/kubernetes/core/src/main/scala/org/apache/spark/scheduler/cluster/k8s/ExecutorPodsLifecycleManager.scala
---
@@ -0,0 +1,146 @@
+/*
+ * 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.
+ */
+package org.apache.spark.scheduler.cluster.k8s
+
+import com.google.common.cache.Cache
+import io.fabric8.kubernetes.api.model.Pod
+import io.fabric8.kubernetes.client.KubernetesClient
+import scala.collection.JavaConverters._
+import scala.collection.mutable
+
+import org.apache.spark.SparkConf
+import org.apache.spark.deploy.k8s.Config._
+import org.apache.spark.scheduler.ExecutorExited
+import org.apache.spark.util.Utils
+
+private[spark] class ExecutorPodsLifecycleManager(
+ conf: SparkConf,
+ executorBuilder: KubernetesExecutorBuilder,
+ kubernetesClient: KubernetesClient,
+ snapshotsStore: ExecutorPodsSnapshotsStore,
+ // Use a best-effort to track which executors have been removed
already. It's not generally
+ // job-breaking if we remove executors more than once but it's ideal
if we make an attempt
+ // to avoid doing so. Expire cache entries so that this data structure
doesn't grow beyond
+ // bounds.
+ removedExecutorsCache: Cache[java.lang.Long, java.lang.Long]) {
+
+ import ExecutorPodsLifecycleManager._
+
+ private val eventProcessingInterval =
conf.get(KUBERNETES_EXECUTOR_EVENT_PROCESSING_INTERVAL)
+
+ def start(schedulerBackend: KubernetesClusterSchedulerBackend): Unit = {
+ snapshotsStore.addSubscriber(eventProcessingInterval) {
+ onNextSnapshot(schedulerBackend, _)
+ }
+ }
+
+ private def onNextSnapshot(
+ schedulerBackend: KubernetesClusterSchedulerBackend,
+ snapshot: ExecutorPodsSnapshot): Unit = {
+ val execIdsRemovedInThisRound = mutable.HashSet.empty[Long]
+ snapshot.executorPods.foreach { case (execId, state) =>
+ state match {
+ case PodDeleted(pod) =>
+ removeExecutorFromSpark(schedulerBackend, pod, execId)
+ execIdsRemovedInThisRound += execId
+ case errorOrSucceeded @ (PodFailed(_) | PodSucceeded(_)) =>
+ removeExecutorFromK8s(errorOrSucceeded.pod)
+ removeExecutorFromSpark(schedulerBackend, errorOrSucceeded.pod,
execId)
+ execIdsRemovedInThisRound += execId
+ case _ =>
+ }
+ }
+
+ // Reconcile the case where Spark claims to know about an executor but
the corresponding pod
+ // is missing from the cluster. This would occur if we miss a deletion
event and the pod
+ // transitions immediately from running io absent.
+ (schedulerBackend.getExecutorIds().map(_.toLong).toSet
+ -- snapshot.executorPods.keySet
+ -- execIdsRemovedInThisRound).foreach { missingExecutorId =>
+ if (removedExecutorsCache.getIfPresent(missingExecutorId) == null) {
+ val exitReason = ExecutorExited(
+ UNKNOWN_EXIT_CODE,
+ exitCausedByApp = false,
+ s"The executor with ID $missingExecutorId was not found in the
cluster but we didn't" +
+ s" get a reason why. Marking the executor as failed. The
executor may have been" +
+ s" deleted but the driver missed the deletion event.")
+ schedulerBackend.doRemoveExecutor(missingExecutorId.toString,
exitReason)
+ }
+ }
+ }
+
+ private def removeExecutorFromK8s(updatedPod: Pod): Unit = {
+ // If deletion failed on a previous try, we can try again if resync
informs us the pod
+ // is still around.
+ // Delete as best attempt - duplicate deletes will throw an exception
but the end state
+ // of getting rid of the pod is what matters.
+ Utils.tryLogNonFatalError {
+ kubernetesClient
+ .pods()
+ .withName(updatedPod.getMetadata.getName)
+ .delete()
+ }
+ }
+
+ private def removeExecutorFromSpark(
+ schedulerBackend: KubernetesClusterSchedulerBackend,
+ pod: Pod,
+ execId: Long): Unit = {
+ if (removedExecutorsCache.getIfPresent(execId) == null) {
+ removedExecutorsCache.put(execId, execId)
+ val exitReason = findExitReason(pod, execId)
+ schedulerBackend.doRemoveExecutor(execId.toString, exitReason)
+ }
+ }
+
+ private def findExitReason(pod: Pod, execId: Long): ExecutorExited = {
+ val exitCode = findExitCode(pod)
+ val (exitCausedByApp, exitMessage) = if (isDeleted(pod)) {
+ (false, s"The executor with id $execId was deleted by a user or the
framework.")
+ } else {
+ val msg = exitReasonMessage(pod, execId, exitCode)
+ (true, msg)
+ }
+ ExecutorExited(exitCode, exitCausedByApp, exitMessage)
+ }
+
+ private def exitReasonMessage(pod: Pod, execId: Long, exitCode: Int) = {
+ s"""
+ |The executor with id $execId exited with exit code $exitCode.
+ |The API gave the following brief reason: ${pod.getStatus.getReason}
+ |The API gave the following message: ${pod.getStatus.getMessage}
+ |The API gave the following container statuses:
+ |
+
|${pod.getStatus.getContainerStatuses.asScala.map(_.toString).mkString("\n===\n")}
+ """.stripMargin
+ }
+
+ private def isDeleted(pod: Pod): Boolean =
pod.getMetadata.getDeletionTimestamp != null
--- End diff --
this method is duplicated in `ExecutorPodsSnapshot.scala`, can we get rid
of it here?
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
