ConfX created MAPREDUCE-7447:
--------------------------------
Summary: Unnecessary NPE encountered when starting
CryptoOutputStream with encrypted-intermediate-data
Key: MAPREDUCE-7447
URL: https://issues.apache.org/jira/browse/MAPREDUCE-7447
Project: Hadoop Map/Reduce
Issue Type: Bug
Reporter: ConfX
Attachments: reproduce.sh
h2. What happened?
Got NullPointerException when initializing a {{{}CryptoOutputStream{}}}.
h2. Where's the bug?
In line 106 of {{{}CryptoOutputStream{}}},the code lacks a check to verify
whether the key parameter is null or not.
{noformat}
public CryptoOutputStream(OutputStream out, CryptoCodec codec,
int bufferSize, byte[] key, byte[] iv, long streamOffset,
boolean closeOutputStream)
throws IOException {
...
this.key = key.clone();{noformat}
As a result, when the configuration provides a null key, the key.clone()
operation will throw a NullPointerException.
It is essential to add a null check for the key parameter before using it.
h2. How to reproduce?
(1) set {{mapreduce.job.encrypted-intermediate-data}} to {{true}}
(2) run
{{org.apache.hadoop.mapreduce.task.reduce.TestMergeManager#testLargeMemoryLimits}}
h2. Stacktrace
h2. Stacktrace
{noformat}
java.lang.NullPointerException
at
org.apache.hadoop.crypto.CryptoOutputStream.<init>(CryptoOutputStream.java:106)
at
org.apache.hadoop.fs.crypto.CryptoFSDataOutputStream.<init>(CryptoFSDataOutputStream.java:38)
at
org.apache.hadoop.mapreduce.CryptoUtils.wrapIfNecessary(CryptoUtils.java:141)
at
org.apache.hadoop.mapreduce.security.IntermediateEncryptedStream.wrapIfNecessary(IntermediateEncryptedStream.java:46)
at
org.apache.hadoop.mapreduce.task.reduce.OnDiskMapOutput.<init>(OnDiskMapOutput.java:87)
at
org.apache.hadoop.mapreduce.task.reduce.MergeManagerImpl.reserve(MergeManagerImpl.java:274)
at
org.apache.hadoop.mapreduce.task.reduce.TestMergeManager.verifyReservedMapOutputType(TestMergeManager.java:309)
at
org.apache.hadoop.mapreduce.task.reduce.TestMergeManager.testLargeMemoryLimits(TestMergeManager.java:303){noformat}
For an easy reproduction, run the reproduce.sh in the attachment.
We are happy to provide a patch if this issue is confirmed.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: mapreduce-dev-unsubscr...@hadoop.apache.org
For additional commands, e-mail: mapreduce-dev-h...@hadoop.apache.org
