[SPARK-22033] BufferHolder, other size checks should account for the specific VM array size limitations

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 2.2.0
• Fix Version/s: 2.3.0
• Component/s: SQL
• Assignee: Sean R. Owen
• Reporter: Vadim Semenov
• Created: 15/Sep/17 22:06
• Updated: 09/Apr/18 20:49
• Resolved: 23/Sep/17 14:41

Description

User may get the following OOM Error while running a job with heavy aggregations

```
java.lang.OutOfMemoryError: Requested array size exceeds VM limit
at org.apache.spark.sql.catalyst.expressions.codegen.BufferHolder.grow(BufferHolder.java:73)
at org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter.write(UnsafeRowWriter.java:235)
at org.apache.spark.sql.catalyst.expressions.codegen.UnsafeRowWriter.write(UnsafeRowWriter.java:228)
at org.apache.spark.sql.catalyst.expressions.GeneratedClass$SpecificUnsafeProjection.apply(Unknown Source)
at org.apache.spark.sql.execution.aggregate.AggregationIterator$$anonfun$generateResultProjection$2.apply(AggregationIterator.scala:254)
at org.apache.spark.sql.execution.aggregate.AggregationIterator$$anonfun$generateResultProjection$2.apply(AggregationIterator.scala:247)
at org.apache.spark.sql.execution.aggregate.ObjectAggregationIterator.next(ObjectAggregationIterator.scala:88)
at org.apache.spark.sql.execution.aggregate.ObjectAggregationIterator.next(ObjectAggregationIterator.scala:33)
at scala.collection.Iterator$$anon$11.next(Iterator.scala:409)
at org.apache.spark.shuffle.sort.UnsafeShuffleWriter.write(UnsafeShuffleWriter.java:167)
at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:96)
at org.apache.spark.scheduler.ShuffleMapTask.runTask(ShuffleMapTask.scala:53)
at org.apache.spark.scheduler.Task.run(Task.scala:108)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:335)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
```

The [`BufferHolder.grow` tries to create a byte array of `Integer.MAX_VALUE` here](https://github.com/apache/spark/blob/v2.2.0/sql/catalyst/src/main/java/org/apache/spark/sql/catalyst/expressions/codegen/BufferHolder.java#L72) but the maximum size of an array depends on specifics of a VM.

The safest value seems to be `Integer.MAX_VALUE - 8`
http://hg.openjdk.java.net/jdk8/jdk8/jdk/file/tip/src/share/classes/java/util/ArrayList.java#l229

In my JVM:
```
java -version
openjdk version "1.8.0_141"
OpenJDK Runtime Environment (build 1.8.0_141-b16)
OpenJDK 64-Bit Server VM (build 25.141-b16, mixed mode)
```

the max is `new Array[Byte](Integer.MAX_VALUE - 2)`

Issue Links

• relates to: SPARK-22222 Fix the ARRAY_MAX in BufferHolder and add a test
• links to: [Github] Pull Request #19266 (srowen)