diff options
author | zhonghaihua <793507405@qq.com> | 2016-04-01 16:23:14 -0500 |
---|---|---|
committer | Tom Graves <tgraves@yahoo-inc.com> | 2016-04-01 16:23:14 -0500 |
commit | bd7b91cefb0d192d808778e6182dcdd2c143e132 (patch) | |
tree | ed8f76bab3aa5042e7f3fa88b4ef2dcd5eb0ddcd | |
parent | 3e991dbc310a4a33eec7f3909adce50bf8268d04 (diff) | |
download | spark-bd7b91cefb0d192d808778e6182dcdd2c143e132.tar.gz spark-bd7b91cefb0d192d808778e6182dcdd2c143e132.tar.bz2 spark-bd7b91cefb0d192d808778e6182dcdd2c143e132.zip |
[SPARK-12864][YARN] initialize executorIdCounter after ApplicationMaster killed for max n…
Currently, when max number of executor failures reached the `maxNumExecutorFailures`, `ApplicationMaster` will be killed and re-register another one.This time, `YarnAllocator` will be created a new instance.
But, the value of property `executorIdCounter` in `YarnAllocator` will reset to `0`. Then the Id of new executor will starting from `1`. This will confuse with the executor has already created before, which will cause FetchFailedException.
This situation is just in yarn client mode, so this is an issue in yarn client mode. For more details, [link to jira issues SPARK-12864](https://issues.apache.org/jira/browse/SPARK-12864)
This PR introduce a mechanism to initialize `executorIdCounter` after `ApplicationMaster` killed.
Author: zhonghaihua <793507405@qq.com>
Closes #10794 from zhonghaihua/initExecutorIdCounterAfterAMKilled.
4 files changed, 29 insertions, 2 deletions
diff --git a/core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedClusterMessage.scala b/core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedClusterMessage.scala index 8d5c11dc36..46a829114e 100644 --- a/core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedClusterMessage.scala +++ b/core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedClusterMessage.scala @@ -30,6 +30,8 @@ private[spark] object CoarseGrainedClusterMessages { case object RetrieveSparkProps extends CoarseGrainedClusterMessage + case object RetrieveLastAllocatedExecutorId extends CoarseGrainedClusterMessage + // Driver to executors case class LaunchTask(data: SerializableBuffer) extends CoarseGrainedClusterMessage diff --git a/core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedSchedulerBackend.scala b/core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedSchedulerBackend.scala index eb4f5331d6..70470cc6d2 100644 --- a/core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedSchedulerBackend.scala +++ b/core/src/main/scala/org/apache/spark/scheduler/cluster/CoarseGrainedSchedulerBackend.scala @@ -79,6 +79,9 @@ class CoarseGrainedSchedulerBackend(scheduler: TaskSchedulerImpl, val rpcEnv: Rp // Executors that have been lost, but for which we don't yet know the real exit reason. protected val executorsPendingLossReason = new HashSet[String] + // The num of current max ExecutorId used to re-register appMaster + protected var currentExecutorIdCounter = 0 + class DriverEndpoint(override val rpcEnv: RpcEnv, sparkProperties: Seq[(String, String)]) extends ThreadSafeRpcEndpoint with Logging { @@ -156,6 +159,9 @@ class CoarseGrainedSchedulerBackend(scheduler: TaskSchedulerImpl, val rpcEnv: Rp // in this block are read when requesting executors CoarseGrainedSchedulerBackend.this.synchronized { executorDataMap.put(executorId, data) + if (currentExecutorIdCounter < executorId.toInt) { + currentExecutorIdCounter = executorId.toInt + } if (numPendingExecutors > 0) { numPendingExecutors -= 1 logDebug(s"Decremented number of pending executors ($numPendingExecutors left)") diff --git a/yarn/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocator.scala b/yarn/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocator.scala index 7d71a642f6..b0bfe855e9 100644 --- a/yarn/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocator.scala +++ b/yarn/src/main/scala/org/apache/spark/deploy/yarn/YarnAllocator.scala @@ -40,6 +40,7 @@ import org.apache.spark.internal.config._ import org.apache.spark.rpc.{RpcCallContext, RpcEndpointRef} import org.apache.spark.scheduler.{ExecutorExited, ExecutorLossReason} import org.apache.spark.scheduler.cluster.CoarseGrainedClusterMessages.RemoveExecutor +import org.apache.spark.scheduler.cluster.CoarseGrainedClusterMessages.RetrieveLastAllocatedExecutorId import org.apache.spark.util.ThreadUtils /** @@ -83,8 +84,23 @@ private[yarn] class YarnAllocator( new ConcurrentHashMap[ContainerId, java.lang.Boolean]) @volatile private var numExecutorsRunning = 0 - // Used to generate a unique ID per executor - private var executorIdCounter = 0 + + /** + * Used to generate a unique ID per executor + * + * Init `executorIdCounter`. when AM restart, `executorIdCounter` will reset to 0. Then + * the id of new executor will start from 1, this will conflict with the executor has + * already created before. So, we should initialize the `executorIdCounter` by getting + * the max executorId from driver. + * + * And this situation of executorId conflict is just in yarn client mode, so this is an issue + * in yarn client mode. For more details, can check in jira. + * + * @see SPARK-12864 + */ + private var executorIdCounter: Int = + driverRef.askWithRetry[Int](RetrieveLastAllocatedExecutorId) + @volatile private var numExecutorsFailed = 0 @volatile private var targetNumExecutors = diff --git a/yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnSchedulerBackend.scala b/yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnSchedulerBackend.scala index a8781636f2..5aeaf44732 100644 --- a/yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnSchedulerBackend.scala +++ b/yarn/src/main/scala/org/apache/spark/scheduler/cluster/YarnSchedulerBackend.scala @@ -292,6 +292,9 @@ private[spark] abstract class YarnSchedulerBackend( logWarning("Attempted to kill executors before the AM has registered!") context.reply(false) } + + case RetrieveLastAllocatedExecutorId => + context.reply(currentExecutorIdCounter) } override def onDisconnected(remoteAddress: RpcAddress): Unit = { |