[GitHub] spark pull request #19440: [SPARK-21871][SQL] Fix infinite loop when bytecod...
Github user asfgit closed the pull request at: https://github.com/apache/spark/pull/19440 --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #19440: [SPARK-21871][SQL] Fix infinite loop when bytecod...
Github user maropu commented on a diff in the pull request:
https://github.com/apache/spark/pull/19440#discussion_r143094785
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/WholeStageCodegenExec.scala
---
@@ -392,12 +392,16 @@ case class WholeStageCodegenExec(child: SparkPlan)
extends UnaryExecNode with Co
// Check if compiled code has a too large function
if (maxCodeSize > sqlContext.conf.hugeMethodLimit) {
- logWarning(s"Found too long generated codes and JIT optimization
might not work: " +
-s"the bytecode size was $maxCodeSize, this value went over the
limit " +
+ logInfo(s"Found too long generated codes and JIT optimization might
not work: " +
+s"the bytecode size ($maxCodeSize) is above the limit " +
s"${sqlContext.conf.hugeMethodLimit}, and the whole-stage codegen
was disabled " +
s"for this plan. To avoid this, you can raise the limit " +
-s"${SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.key}:\n$treeString")
- return child.execute()
+s"`${SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.key}`:\n$treeString")
+ child match {
+// For batch file source scan, we should continue executing it
+case f: FileSourceScanExec if f.supportsBatch => // do nothing
--- End diff --
yea, I totally agree that we need to refactor this in future. Anyway, it's
ok for now.
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #19440: [SPARK-21871][SQL] Fix infinite loop when bytecod...
Github user viirya commented on a diff in the pull request:
https://github.com/apache/spark/pull/19440#discussion_r143092189
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/WholeStageCodegenExec.scala
---
@@ -392,12 +392,16 @@ case class WholeStageCodegenExec(child: SparkPlan)
extends UnaryExecNode with Co
// Check if compiled code has a too large function
if (maxCodeSize > sqlContext.conf.hugeMethodLimit) {
- logWarning(s"Found too long generated codes and JIT optimization
might not work: " +
-s"the bytecode size was $maxCodeSize, this value went over the
limit " +
+ logInfo(s"Found too long generated codes and JIT optimization might
not work: " +
+s"the bytecode size ($maxCodeSize) is above the limit " +
s"${sqlContext.conf.hugeMethodLimit}, and the whole-stage codegen
was disabled " +
s"for this plan. To avoid this, you can raise the limit " +
-s"${SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.key}:\n$treeString")
- return child.execute()
+s"`${SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.key}`:\n$treeString")
+ child match {
+// For batch file source scan, we should continue executing it
--- End diff --
It's better to explain why we should continue it. Otherwise later readers
may not understand it immediately.
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #19440: [SPARK-21871][SQL] Fix infinite loop when bytecod...
Github user gatorsmile commented on a diff in the pull request:
https://github.com/apache/spark/pull/19440#discussion_r143091966
--- Diff:
sql/core/src/test/scala/org/apache/spark/sql/execution/WholeStageCodegenSuite.scala
---
@@ -185,4 +185,22 @@ class WholeStageCodegenSuite extends SparkPlanTest
with SharedSQLContext {
val (_, maxCodeSize2) = CodeGenerator.compile(codeWithLongFunctions)
assert(maxCodeSize2 >
SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.defaultValue.get)
}
+
+ test("returning batch for wide table") {
+import testImplicits._
+withTempPath { dir =>
+ val path = dir.getCanonicalPath
+ val df = spark.range(10).select(Seq.tabulate(201) {i => ('id +
i).as(s"c$i")} : _*)
+ df.write.mode(SaveMode.Overwrite).parquet(path)
+
+ withSQLConf(SQLConf.WHOLESTAGE_MAX_NUM_FIELDS.key -> "202",
+SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.key -> "8000") {
+// donot return batch, because whole stage codegen is disabled for
wide table (>202 columns)
--- End diff --
this is copied and pasted. will fix it.
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #19440: [SPARK-21871][SQL] Fix infinite loop when bytecod...
Github user viirya commented on a diff in the pull request:
https://github.com/apache/spark/pull/19440#discussion_r143091744
--- Diff:
sql/core/src/test/scala/org/apache/spark/sql/execution/WholeStageCodegenSuite.scala
---
@@ -185,4 +185,22 @@ class WholeStageCodegenSuite extends SparkPlanTest
with SharedSQLContext {
val (_, maxCodeSize2) = CodeGenerator.compile(codeWithLongFunctions)
assert(maxCodeSize2 >
SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.defaultValue.get)
}
+
+ test("returning batch for wide table") {
+import testImplicits._
+withTempPath { dir =>
+ val path = dir.getCanonicalPath
+ val df = spark.range(10).select(Seq.tabulate(201) {i => ('id +
i).as(s"c$i")} : _*)
+ df.write.mode(SaveMode.Overwrite).parquet(path)
+
+ withSQLConf(SQLConf.WHOLESTAGE_MAX_NUM_FIELDS.key -> "202",
+SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.key -> "8000") {
+// donot return batch, because whole stage codegen is disabled for
wide table (>202 columns)
--- End diff --
Is this comment wrong or I misunderstand it? Looks like it returns batch as
it asserts `supportsBatch`.
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #19440: [SPARK-21871][SQL] Fix infinite loop when bytecod...
Github user gatorsmile commented on a diff in the pull request:
https://github.com/apache/spark/pull/19440#discussion_r143089714
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/WholeStageCodegenExec.scala
---
@@ -392,12 +392,16 @@ case class WholeStageCodegenExec(child: SparkPlan)
extends UnaryExecNode with Co
// Check if compiled code has a too large function
if (maxCodeSize > sqlContext.conf.hugeMethodLimit) {
- logWarning(s"Found too long generated codes and JIT optimization
might not work: " +
-s"the bytecode size was $maxCodeSize, this value went over the
limit " +
+ logInfo(s"Found too long generated codes and JIT optimization might
not work: " +
+s"the bytecode size ($maxCodeSize) is above the limit " +
s"${sqlContext.conf.hugeMethodLimit}, and the whole-stage codegen
was disabled " +
s"for this plan. To avoid this, you can raise the limit " +
-s"${SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.key}:\n$treeString")
- return child.execute()
+s"`${SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.key}`:\n$treeString")
+ child match {
+// For batch file source scan, we should continue executing it
+case f: FileSourceScanExec if f.supportsBatch => // do nothing
--- End diff --
If we do it in `FileSourceScanExec `, we are unable to know which causes
the fallback. Now, we have at least two reasons that trigger the fallback.
Ideally, we should not call `WholeStageCodegenExec ` in `doExecute`.
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #19440: [SPARK-21871][SQL] Fix infinite loop when bytecod...
Github user maropu commented on a diff in the pull request:
https://github.com/apache/spark/pull/19440#discussion_r143086096
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/WholeStageCodegenExec.scala
---
@@ -392,12 +392,16 @@ case class WholeStageCodegenExec(child: SparkPlan)
extends UnaryExecNode with Co
// Check if compiled code has a too large function
if (maxCodeSize > sqlContext.conf.hugeMethodLimit) {
- logWarning(s"Found too long generated codes and JIT optimization
might not work: " +
-s"the bytecode size was $maxCodeSize, this value went over the
limit " +
+ logInfo(s"Found too long generated codes and JIT optimization might
not work: " +
+s"the bytecode size ($maxCodeSize) is above the limit " +
s"${sqlContext.conf.hugeMethodLimit}, and the whole-stage codegen
was disabled " +
s"for this plan. To avoid this, you can raise the limit " +
-s"${SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.key}:\n$treeString")
- return child.execute()
+s"`${SQLConf.WHOLESTAGE_HUGE_METHOD_LIMIT.key}`:\n$treeString")
+ child match {
+// For batch file source scan, we should continue executing it
+case f: FileSourceScanExec if f.supportsBatch => // do nothing
--- End diff --
I feel a little weird `WholeStageCodegenExec` has specific error handling
for each spark plan. Could we handle this error inside `FileSourceScanExec`?
For example, how about checking if `parent.isInstanceOf[WholeStageCodegenExec]`
in `FileSourceScanExec`?
---
-
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
[GitHub] spark pull request #19440: [SPARK-21871][SQL] Fix infinite loop when bytecod...
GitHub user gatorsmile opened a pull request: https://github.com/apache/spark/pull/19440 [SPARK-21871][SQL] Fix infinite loop when bytecode size is larger than spark.sql.codegen.hugeMethodLimit ## What changes were proposed in this pull request? When exceeding `spark.sql.codegen.hugeMethodLimit`, the runtime fallbacks to the Volcano iterator solution. This could cause an infinite loop when `FileSourceScanExec` can use the columnar batch to read the data. This PR is to fix the issue. ## How was this patch tested? Added a test You can merge this pull request into a Git repository by running: $ git pull https://github.com/gatorsmile/spark testt Alternatively you can review and apply these changes as the patch at: https://github.com/apache/spark/pull/19440.patch To close this pull request, make a commit to your master/trunk branch with (at least) the following in the commit message: This closes #19440 commit 6e2f53238856942e40a3301108f37a3a5cc17bca Author: gatorsmile Date: 2017-10-05T18:20:48Z fix. --- - To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org
