[SPARK-3630] Identify cause of Kryo+Snappy PARSING_ERROR

• Type: Task
• Status: Resolved
• Priority: Major
• Resolution: Done
• Affects Version/s: 1.1.0, 1.2.0
• Component/s: Spark Core
• Assignee: Josh Rosen
• Reporter: Andrew Ash
• Created: 22/Sep/14 01:28
• Updated: 22/Aug/16 07:15
• Resolved: 03/Aug/16 11:24

Description

A recent GraphX commit caused non-deterministic exceptions in unit tests so it was reverted (see SPARK-3400).

Separately, aash observed the same exception stacktrace in an application-specific Kryo registrator:

com.esotericsoftware.kryo.KryoException: java.io.IOException: failed to uncompress the chunk: PARSING_ERROR(2)
com.esotericsoftware.kryo.io.Input.fill(Input.java:142) com.esotericsoftware.kryo.io.Input.require(Input.java:169) 
com.esotericsoftware.kryo.io.Input.readInt(Input.java:325) com.esotericsoftware.kryo.io.Input.readFloat(Input.java:624) 
com.esotericsoftware.kryo.serializers.DefaultSerializers$FloatSerializer.read(DefaultSerializers.java:127) 
com.esotericsoftware.kryo.serializers.DefaultSerializers$FloatSerializer.read(DefaultSerializers.java:117) 
com.esotericsoftware.kryo.Kryo.readClassAndObject(Kryo.java:732) 
com.esotericsoftware.kryo.serializers.CollectionSerializer.read(CollectionSerializer.java:109) 
com.esotericsoftware.kryo.serializers.CollectionSerializer.read(CollectionSerializer.java:18) 
com.esotericsoftware.kryo.Kryo.readClassAndObject(Kryo.java:732)
...

This ticket is to identify the cause of the exception in the GraphX commit so the faulty commit can be fixed and merged back into master.

Issue Links

• is duplicated by: SPARK-6977 PARSING_ERROR(2) in Spark Streaming
• is duplicated by: SPARK-3743 noisy logging when context is stopped
• is related to: SPARK-3426 Sort-based shuffle compression behavior is inconsistent
• is related to: SPARK-3948 Sort-based shuffle can lead to assorted stream-corruption exceptions
• is related to: SPARK-3958 Possible stream-corruption issues in TorrentBroadcast
• is related to: SPARK-4019 Shuffling with more than 2000 reducers may drop all data when partitons are mostly empty or cause deserialization errors if at least one partition is empty
• is related to: SPARK-4105 FAILED_TO_UNCOMPRESS(5) errors when fetching shuffle data with sort-based shuffle
• is related to: SPARK-4133 PARSING_ERROR(2) when upgrading issues from 1.0.2 to 1.1.0
• is related to: SPARK-4056 Upgrade snappy-java to 1.1.1.5
• relates to: SPARK-3400 GraphX unit tests fail nondeterministically