maryannxue commented on a change in pull request #24706: [SPARK-23128][SQL] A new approach to do adaptive execution in Spark SQL URL: https://github.com/apache/spark/pull/24706#discussion_r292994737
########## File path: sql/core/src/main/scala/org/apache/spark/sql/execution/adaptive/AdaptiveSparkPlanExec.scala ########## @@ -0,0 +1,446 @@ +/* + * 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.sql.execution.adaptive + +import java.util +import java.util.concurrent.LinkedBlockingQueue + +import scala.collection.JavaConverters._ +import scala.collection.concurrent.TrieMap +import scala.collection.mutable +import scala.concurrent.ExecutionContext + +import org.apache.spark.SparkException +import org.apache.spark.rdd.RDD +import org.apache.spark.sql.SparkSession +import org.apache.spark.sql.catalyst.InternalRow +import org.apache.spark.sql.catalyst.expressions.Attribute +import org.apache.spark.sql.catalyst.plans.logical.{LogicalPlan, ReturnAnswer} +import org.apache.spark.sql.catalyst.rules.{Rule, RuleExecutor} +import org.apache.spark.sql.execution._ +import org.apache.spark.sql.execution.adaptive.AdaptiveSparkPlanExec._ +import org.apache.spark.sql.execution.exchange._ +import org.apache.spark.sql.execution.ui.SparkListenerSQLAdaptiveExecutionUpdate +import org.apache.spark.sql.internal.SQLConf +import org.apache.spark.util.ThreadUtils + +/** + * A root node to execute the query plan adaptively. It splits the query plan into independent + * stages and executes them in order according to their dependencies. The query stage + * materializes its output at the end. When one stage completes, the data statistics of the + * materialized output will be used to optimize the remainder of the query. + * + * To create query stages, we traverse the query tree bottom up. When we hit an exchange node, + * and if all the child query stages of this exchange node are materialized, we create a new + * query stage for this exchange node. The new stage is then materialized asynchronously once it + * is created. + * + * When one query stage finishes materialization, the rest query is re-optimized and planned based + * on the latest statistics provided by all materialized stages. Then we traverse the query plan + * again and create more stages if possible. After all stages have been materialized, we execute + * the rest of the plan. + */ +case class AdaptiveSparkPlanExec( + initialPlan: SparkPlan, + @transient session: SparkSession, + @transient subqueryMap: Map[Long, ExecSubqueryExpression], + @transient stageCache: TrieMap[SparkPlan, QueryStageExec]) + extends LeafExecNode { + + @transient private val executionId = Option( + session.sparkContext.getLocalProperty(SQLExecution.EXECUTION_ID_KEY)).map(_.toLong) + + @transient private val lock = new Object() + + // The logical plan optimizer for re-optimizing the current logical plan. + @transient private val optimizer = new RuleExecutor[LogicalPlan] { + // TODO add more optimization rules + override protected def batches: Seq[Batch] = Seq() + } + + // A list of physical plan rules to be applied before creation of query stages. The physical + // plan should reach a final status of query stages (i.e., no more addition or removal of + // Exchange nodes) after running these rules. + @transient private val queryStagePreparationRules: Seq[Rule[SparkPlan]] = Seq( + PlanAdaptiveSubqueries(subqueryMap), + EnsureRequirements(conf) + ) + + // A list of physical optimizer rules to be applied to a new stage before its execution. These + // optimizations should be stage-independent. + @transient private val queryStageOptimizerRules: Seq[Rule[SparkPlan]] = Seq( + CollapseCodegenStages(conf) + ) + + private var currentStageId = 0 + + @volatile private var currentPhysicalPlan = initialPlan + + @volatile private var isFinalPlan = false + + @volatile private var fallback = false + + /** + * Return type for `createQueryStages` + * @param newPlan the new plan with created query stages. + * @param allChildStagesMaterialized whether all child stages have been materialized. + * @param newStages the newly created query stages, including new reused query stages. + */ + private case class CreateStageResult( + newPlan: SparkPlan, + allChildStagesMaterialized: Boolean, + newStages: Seq[(Exchange, QueryStageExec)]) + + def executedPlan: SparkPlan = currentPhysicalPlan + + override def conf: SQLConf = session.sessionState.conf + + override def output: Seq[Attribute] = initialPlan.output + + override def doCanonicalize(): SparkPlan = initialPlan.canonicalized + + override def doExecute(): RDD[InternalRow] = if (isFinalPlan) { + currentPhysicalPlan.execute() + } else { + lock.synchronized { + var currentLogicalPlan = currentPhysicalPlan.logicalLink.get + val (r, p) = createQueryStagesOrFallback(currentPhysicalPlan, currentLogicalPlan) + var result = r + var logicalPlan = p + val events = new LinkedBlockingQueue[StageMaterializationEvent]() + val errors = new mutable.ArrayBuffer[SparkException]() + while (!result.allChildStagesMaterialized) { + currentPhysicalPlan = result.newPlan + currentLogicalPlan = logicalPlan + currentPhysicalPlan.setTagValue(SparkPlan.LOGICAL_PLAN_TAG, currentLogicalPlan) + onUpdatePlan() + + // Start materialization of all new stages. + result.newStages.map(_._2).foreach { stage => + stage.materialize().onComplete { res => + if (res.isSuccess) { + events.offer(StageSuccess(stage, res.get)) + } else { + events.offer(StageFailure(stage, res.failed.get)) + } + }(AdaptiveSparkPlanExec.executionContext) + } + + // Wait on the next completed stage, which indicates new stats are available and probably + // new stages can be created. There might be other stages that finish at around the same + // time, so we process those stages too in order to reduce re-planning. + val nextMsg = events.take() + val rem = new util.ArrayList[StageMaterializationEvent]() + events.drainTo(rem) + (Seq(nextMsg) ++ rem.asScala).foreach { + case StageSuccess(stage, res) => + stage.resultOption = Some(res) + case StageFailure(stage, ex) => + errors.append( + new SparkException(s"Fail to materialize query stage: ${stage.treeString}", ex)) + } + + // In case of errors, we cancel all running stages and throw exception. + if (errors.nonEmpty) { + try { + currentPhysicalPlan.foreach { + case s: QueryStageExec => s.cancel() + case _ => + } + } finally { + val ex = new SparkException( + "Adaptive execution failed due to stage materialization failures.", errors.head) + errors.tail.foreach(ex.addSuppressed) + throw ex + } + } + + // Do re-planning and try creating new stages on the new physical plan. + val (newPhysicalPlan, newLogicalPlan) = reOptimize(currentLogicalPlan) + currentPhysicalPlan = newPhysicalPlan + currentLogicalPlan = newLogicalPlan + val (r, p) = createQueryStagesOrFallback(currentPhysicalPlan, currentLogicalPlan) + result = r + logicalPlan = p + } + + // Run the final plan when there's no more unfinished stages. + currentPhysicalPlan = applyPhysicalRules(result.newPlan, queryStageOptimizerRules) + currentPhysicalPlan.setTagValue(SparkPlan.LOGICAL_PLAN_TAG, logicalPlan) + logDebug(s"Final plan: $currentPhysicalPlan") + onUpdatePlan() + isFinalPlan = true + currentPhysicalPlan.execute() + } + } + + override def generateTreeString( + depth: Int, + lastChildren: Seq[Boolean], + append: String => Unit, + verbose: Boolean, + prefix: String = "", + addSuffix: Boolean = false, + maxFields: Int): Unit = { + currentPhysicalPlan.generateTreeString( + depth, lastChildren, append, verbose, "", addSuffix = false, maxFields) + } + + /** + * Try creating new query stages and updating the logical plan accordingly. Return the + * `CreateStageResult` along with the updated logical plan if successful; otherwise, turn on + * the "fallback" mode, which means no new stages will be created and we just wait for all the + * existing stages to complete and execute the rest of the plan. + */ + private def createQueryStagesOrFallback( + physicalPlan: SparkPlan, + logicalPlan: LogicalPlan): (CreateStageResult, LogicalPlan) = { + var result = createQueryStages(physicalPlan) + var newLogicalPlan = logicalPlan + try { + newLogicalPlan = updateLogicalPlan(logicalPlan, result.newPlan, result.newStages) + } catch { + case e: AmbiguousLogicalMappingException => + logWarning("Fall back to non-adaptive mode for the rest of the plan due to ambiguous " + + s"logical plan mapping for node ${e.plan}.") + fallback = true + result = createQueryStages(physicalPlan) + assert(result.newStages.isEmpty, "Fallback mode should not create new stages.") + } + (result, newLogicalPlan) + } + + /** + * This method is called recursively to traverse the plan tree bottom-up and create a new query + * stage or try reusing an existing stage if the current node is an [[Exchange]] node and all of + * its child stages have been materialized. + * + * With each call, it returns: + * 1) The new plan replaced with [[QueryStageExec]] nodes where new stages are created. + * 2) Whether the child query stages (if any) of the current node have all been materialized. + * 3) A list of the new query stages that have been created. + */ + private def createQueryStages(plan: SparkPlan): CreateStageResult = plan match { + case e: Exchange if !fallback => + // First have a quick check in the `stageCache` without having to traverse down the node. + stageCache.get(e.canonicalized) match { + case Some(existingStage) if conf.exchangeReuseEnabled => + val reusedStage = reuseQueryStage(existingStage, e.output) + // When reusing a stage, we treat it a new stage regardless of whether the existing stage + // has been materialized or not. Thus we won't skip re-optimization for a reused stage. + CreateStageResult(newPlan = reusedStage, + allChildStagesMaterialized = false, newStages = Seq((e, reusedStage))) + + case _ => + val result = createQueryStages(e.child) + val newPlan = e.withNewChildren(Seq(result.newPlan)).asInstanceOf[Exchange] + // Create a query stage only when all the child query stages are ready. + if (result.allChildStagesMaterialized) { + var newStage = newQueryStage(newPlan) + if (conf.exchangeReuseEnabled) { + // Check the `stageCache` again for reuse. If a match is found, ditch the new stage + // and reuse the existing stage found in the `stageCache`, otherwise update the + // `stageCache` with the new stage. + val queryStage = stageCache.getOrElseUpdate(e.canonicalized, newStage) + if (queryStage.ne(newStage)) { + newStage = reuseQueryStage(queryStage, e.output) + } + } + + // We've created a new stage, which is obviously not ready yet. + CreateStageResult(newPlan = newStage, + allChildStagesMaterialized = false, newStages = Seq((e, newStage))) + } else { + CreateStageResult(newPlan = newPlan, + allChildStagesMaterialized = false, newStages = result.newStages) + } + } + + case q: QueryStageExec => + CreateStageResult(newPlan = q, + allChildStagesMaterialized = q.resultOption.isDefined, newStages = Seq.empty) + + case _ => + if (plan.children.isEmpty) { + CreateStageResult(newPlan = plan, allChildStagesMaterialized = true, newStages = Seq.empty) + } else { + val results = plan.children.map(createQueryStages) + CreateStageResult( + newPlan = plan.withNewChildren(results.map(_.newPlan)), + allChildStagesMaterialized = results.forall(_.allChildStagesMaterialized), + newStages = results.flatMap(_.newStages)) + } + } + + private def newQueryStage(e: Exchange): QueryStageExec = { + val optimizedPlan = applyPhysicalRules(e.child, queryStageOptimizerRules) + val queryStage = e match { + case s: ShuffleExchangeExec => + ShuffleQueryStageExec(currentStageId, s.copy(child = optimizedPlan)) + case b: BroadcastExchangeExec => + BroadcastQueryStageExec(currentStageId, b.copy(child = optimizedPlan)) + } + currentStageId += 1 + queryStage + } + + private def reuseQueryStage(s: QueryStageExec, output: Seq[Attribute]): QueryStageExec = { + val queryStage = ReusedQueryStageExec(currentStageId, s, output) + currentStageId += 1 + queryStage + } + + /** + * Returns the updated logical plan after new query stages have been created and the physical + * plan has been updated with the newly created stages. + * 1. If the new query stage can be mapped to an integral logical sub-tree, replace the + * corresponding logical sub-tree with a leaf node [[LogicalQueryStage]] referencing the new + * query stage. For example: + * Join SMJ SMJ + * / \ / \ / \ + * r1 r2 => Xchg1 Xchg2 => Stage1 Stage2 + * | | + * r1 r2 + * The updated plan node will be: + * Join + * / \ + * LogicalQueryStage1(Stage1) LogicalQueryStage2(Stage2) + * + * 2. Otherwise (which means the new query stage can only be mapped to part of a logical + * sub-tree), replace the corresponding logical sub-tree with a leaf node + * [[LogicalQueryStage]] referencing to the top physical node into which this logical node is + * transformed during physical planning. For example: + * Agg HashAgg HashAgg + * | | | + * child => Xchg => Stage1 + * | + * HashAgg + * | + * child + * The updated plan node will be: + * LogicalQueryStage(HashAgg - Stage1) + */ + private def updateLogicalPlan( + logicalPlan: LogicalPlan, + physicalPlan: SparkPlan, + newStages: Seq[(Exchange, QueryStageExec)]): LogicalPlan = { + var currentLogicalPlan = logicalPlan + newStages.foreach { case (exhange: Exchange, stage: QueryStageExec) => + // Get the corresponding logical node for `exchange`. If `exchange` has been transformed from + // a `Repartition`, it should have `logicalLink` available by itself; otherwise traverse down + // to find the first node that is not generated by `EnsureRequirements`. + val logicalNodeOpt = exhange.logicalLink.orElse(exhange.collectFirst { + case p if p.logicalLink.isDefined => p.logicalLink.get + }) + assert(logicalNodeOpt.isDefined) + val logicalNode = logicalNodeOpt.get + val physicalNode = physicalPlan.collectFirst { + case p if p.eq(stage) || p.logicalLink.exists(logicalNode.eq) => p + } + assert(physicalNode.isDefined) + // Replace the corresponding logical node with LogicalQueryStage + val newLogicalNode = LogicalQueryStage(logicalNode, physicalNode.get) + var cnt = 0 + val newLogicalPlan = currentLogicalPlan.transformDown { + case p if p.eq(logicalNode) => + cnt += 1 + newLogicalNode + } + if (cnt > 1) { + throw AmbiguousLogicalMappingException(logicalNode) + } + assert(cnt == 1, + s"logicalNode: $logicalNode; " + + s"logicalPlan: $currentLogicalPlan " + + s"physicalPlan: $physicalPlan" + + s"stage: $stage") + currentLogicalPlan = newLogicalPlan + } + currentLogicalPlan + } + + /** + * Re-optimize and run physical planning on the current logical plan based on the latest stats. + */ + private def reOptimize(logicalPlan: LogicalPlan): (SparkPlan, LogicalPlan) = { + logicalPlan.invalidateStatsCache() + val optimized = optimizer.execute(logicalPlan) + SparkSession.setActiveSession(session) + val sparkPlan = session.sessionState.planner.plan(ReturnAnswer(optimized)).next() + val newPlan = applyPhysicalRules(sparkPlan, queryStagePreparationRules) + (newPlan, optimized) + } + + /** + * Notify the listeners of the physical plan change. + */ + private def onUpdatePlan(): Unit = { + executionId.foreach { id => + val exec = SQLExecution.getQueryExecution(id) + if (exec != null) { + session.sparkContext.listenerBus.post(SparkListenerSQLAdaptiveExecutionUpdate( + id, + exec.toString, Review comment: Not sure I understand it. The physical plan has a link to the logical plan, is that sufficient? ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
