[
https://issues.apache.org/jira/browse/FLINK-6044?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15924222#comment-15924222
]
ASF GitHub Bot commented on FLINK-6044:
---------------------------------------
GitHub user StefanRRichter opened a pull request:
https://github.com/apache/flink/pull/3533
[FLINK-6044] Replace calls to InputStream#read(...) with the indended…
This PR fixes FLINK-6044. On top of that, I searched through the code for
calls to `InputStream#read(...)` that ignore their return value and replaced
them with the intended `InputStream#readFully(...)` calls.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/StefanRRichter/flink FLINK-6044-readFully
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/flink/pull/3533.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 #3533
----
commit 342144a967c5d93ba87f1d2850b88dad7c796f81
Author: Stefan Richter <[email protected]>
Date: 2017-03-14T13:45:00Z
[FLINK-6044] Replace calls to InputStream#read(...) with the indended
InputStream#readFully(...)
----
> TypeSerializerSerializationProxy.read() doesn't verify the read buffer length
> -----------------------------------------------------------------------------
>
> Key: FLINK-6044
> URL: https://issues.apache.org/jira/browse/FLINK-6044
> Project: Flink
> Issue Type: Bug
> Components: Type Serialization System
> Affects Versions: 1.2.0
> Environment: Ubuntu server 12.04.5 64 bit
> java version "1.8.0_111"
> Java(TM) SE Runtime Environment (build 1.8.0_111-b14)
> Java HotSpot(TM) 64-Bit Server VM (build 25.111-b14, mixed mode)
> Reporter: Avihai Berkovitz
> Assignee: Stefan Richter
> Priority: Critical
>
> The read() method of TypeSerializerSerializationProxy creates a buffers and
> tries to fill it by calling the read() method of the given DataInputView, but
> never checks the return value. The actual size read from the stream might be
> smaller than the buffer size, and the rest of the buffer is filled with
> zeroes, causing the deserialization to fail.
> It happened to me using a RocksDB state backend backed by S3. The setup was
> done according to
> https://ci.apache.org/projects/flink/flink-docs-release-1.2/setup/aws.html#s3-simple-storage-service
> and everything worked correctly until I upgraded to Flink 1.2.0.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
