SPARK-3025 [SQL]: Allow JDBC clients to set a fair scheduler pool

docs/sql-programming-guide.md

@@ -605,6 +605,11 @@ Configuration of Hive is done by placing your `hive-site.xml` file in `conf/`.
 
 You may also use the beeline script comes with Hive.
 
+To set a [Fair Scheduler](job-scheduling.html#fair-scheduler-pools) pool for a JDBC client session,
+users can set the `spark.sql.thriftserver.scheduler.pool` variable:
+
+    SET spark.sql.thriftserver.scheduler.pool=accounting;
+
 ### Migration Guide for Shark Users
 
 #### Reducer number

sql/core/src/main/scala/org/apache/spark/sql/SQLConf.scala

@@ -32,6 +32,9 @@ private[spark] object SQLConf {
   val CODEGEN_ENABLED = "spark.sql.codegen"
   val DIALECT = "spark.sql.dialect"
 
+  // This is only used for the thriftserver
+  val THRIFTSERVER_POOL = "spark.sql.thriftserver.scheduler.pool"
+
   object Deprecated {
     val MAPRED_REDUCE_TASKS = "mapred.reduce.tasks"
   }

sql/hive-thriftserver/src/main/scala/org/apache/spark/sql/hive/thriftserver/server/SparkSQLOperationManager.scala

@@ -17,24 +17,24 @@
 
 package org.apache.spark.sql.hive.thriftserver.server
 
-import scala.collection.JavaConversions._
-import scala.collection.mutable.ArrayBuffer
-import scala.math.{random, round}
-
 import java.sql.Timestamp
 import java.util.{Map => JMap}
 
+import scala.collection.JavaConversions._
+import scala.collection.mutable.{ArrayBuffer, Map}
+import scala.math.{random, round}
+
 import org.apache.hadoop.hive.common.`type`.HiveDecimal
 import org.apache.hadoop.hive.metastore.api.FieldSchema
 import org.apache.hive.service.cli._
 import org.apache.hive.service.cli.operation.{ExecuteStatementOperation, Operation, OperationManager}
 import org.apache.hive.service.cli.session.HiveSession
-
 import org.apache.spark.Logging
+import org.apache.spark.sql.{Row => SparkRow, SQLConf, SchemaRDD}
+import org.apache.spark.sql.catalyst.plans.logical.SetCommand
 import org.apache.spark.sql.catalyst.types._
-import org.apache.spark.sql.hive.thriftserver.ReflectionUtils
 import org.apache.spark.sql.hive.{HiveContext, HiveMetastoreTypes}
-import org.apache.spark.sql.{SchemaRDD, Row => SparkRow}
+import org.apache.spark.sql.hive.thriftserver.ReflectionUtils
 
 /**
  * Executes queries using Spark SQL, and maintains a list of handles to active queries.
@@ -43,6 +43,9 @@ class SparkSQLOperationManager(hiveContext: HiveContext) extends OperationManage
   val handleToOperation = ReflectionUtils
     .getSuperField[JMap[OperationHandle, Operation]](this, "handleToOperation")
 
+  // TODO: Currenlty this will grow infinitely, even as sessions expire
+  val sessionToActivePool = Map[HiveSession, String]()
+
   override def newExecuteStatementOperation(
       parentSession: HiveSession,
       statement: String,
@@ -130,8 +133,18 @@ class SparkSQLOperationManager(hiveContext: HiveContext) extends OperationManage
         try {
           result = hiveContext.sql(statement)
           logDebug(result.queryExecution.toString())
+          result.queryExecution.logical match {
+            case SetCommand(Some(key), Some(value)) if (key == SQLConf.THRIFTSERVER_POOL) =>
+              sessionToActivePool(parentSession) = value
+              logInfo(s"Setting spark.scheduler.pool=$value for future statements in this session.")
+            case _ =>
+          }
+
           val groupId = round(random * 1000000).toString
           hiveContext.sparkContext.setJobGroup(groupId, statement)
+          sessionToActivePool.get(parentSession).foreach { pool =>
+            hiveContext.sparkContext.setLocalProperty("spark.scheduler.pool", pool)
+          }
           iter = {
             val resultRdd = result.queryExecution.toRdd
             val useIncrementalCollect =