[SPARK-23524] Big local shuffle blocks should not be checked for corruption.

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.2.1
• Fix Version/s: 2.3.1, 2.4.0
• Component/s: Spark Core
• Assignee: Jin Xing
• Reporter: Jin Xing
• Created: 27/Feb/18 09:55
• Updated: 08/Mar/18 04:09
• Resolved: 08/Mar/18 04:09

Description

In current code, all local blocks will be checked for corruption no matter it's big or not.  The reasons are as below:

1. Size in FetchResult for local block is set to be 0 (https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/storage/ShuffleBlockFetcherIterator.scala#L327)
2. SPARK-4105 meant to only check the small blocks(size<maxBytesInFlight/3), but for reason 1, below check will be invalid. https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/storage/ShuffleBlockFetcherIterator.scala#L420

Issue Links

• links to: [Github] Pull Request #20685 (jinxing64)