Skip to content

[SPARK-26992][STS] Fix STS scheduler pool correct delivery #23895

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
cxzl25 wants to merge 2 commits into from
Closed

[SPARK-26992][STS] Fix STS scheduler pool correct delivery #23895

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
4e735d4
Fix STS scheduler pool correct delivery
cxzl25 Feb 26, 2019
8ae9369
Merge branch 'master' into thrift_server_scheduler_pool_pollute
cxzl25 Apr 2, 2019
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 16 additions & 6 deletions ...rc/main/scala/org/apache/spark/sql/hive/thriftserver/SparkExecuteStatementOperation.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,7 +109,7 @@ private[hive] class SparkExecuteStatementOperation(
}
}

def getNextRowSet(order: FetchOrientation, maxRowsL: Long): RowSet = {
def getNextRowSet(order: FetchOrientation, maxRowsL: Long): RowSet = withSchedulerPool {
validateDefaultFetchOrientation(order)
assertState(OperationState.FINISHED)
setHasResultSet(true)
Expand Down Expand Up @@ -210,7 +210,7 @@ private[hive] class SparkExecuteStatementOperation(
}
}

private def execute(): Unit = {
private def execute(): Unit = withSchedulerPool {
statementId = UUID.randomUUID().toString
logInfo(s"Running query '$statement' with $statementId")
setState(OperationState.RUNNING)
Expand All @@ -225,10 +225,6 @@ private[hive] class SparkExecuteStatementOperation(
statementId,
parentSession.getUsername)
sqlContext.sparkContext.setJobGroup(statementId, statement)
val pool = sessionToActivePool.get(parentSession.getSessionHandle)
// It may have unpredictably behavior since we use thread pools to execute quries and
// the 'spark.scheduler.pool' may not be 'default' when we did not set its value.(SPARK-26914)
sqlContext.sparkContext.setLocalProperty(SparkContext.SPARK_SCHEDULER_POOL, pool)
try {
result = sqlContext.sql(statement)
logDebug(result.queryExecution.toString())
Expand Down Expand Up @@ -291,6 +287,20 @@ private[hive] class SparkExecuteStatementOperation(
sqlContext.sparkContext.cancelJobGroup(statementId)
}
}

private def withSchedulerPool[T](body: => T): T = {
val pool = sessionToActivePool.get(parentSession.getSessionHandle)
if (pool != null) {
Copy link
Member

@srowen srowen Apr 2, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Although this looks like it undoes the previous fix, you're saying it will always be null when this method starts right? It seems easier to unconditionally set it, unless there's some overhead I'm missing

Copy link
Contributor Author

@cxzl25 cxzl25 Apr 2, 2019

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Because only the user submitted sql has set spark.sql.thriftserver.scheduler.pool=xxx, sessionToActivePool will put xxx value, will not be the default scheduler, so the default scheduler pool name is null

Unconditional setting it is worried about ambiguity, because only after the put value, there will be a need to take this value, although there is no overhead

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

Lines 236 to 239 in 43bf4ae

case SetCommand(Some((SQLConf.THRIFTSERVER_POOL.key, Some(value)))) =>
sessionToActivePool.put(parentSession.getSessionHandle, value)
logInfo(s"Setting ${SparkContext.SPARK_SCHEDULER_POOL}=$value for future statements " +
"in this session.")

sqlContext.sparkContext.setLocalProperty(SparkContext.SPARK_SCHEDULER_POOL, pool)
}
try {
body
} finally {
if (pool != null) {
sqlContext.sparkContext.setLocalProperty(SparkContext.SPARK_SCHEDULER_POOL, null)
}
}
}
}

object SparkExecuteStatementOperation {
Expand Down