[ 
https://issues.apache.org/jira/browse/FLINK-2545?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14716152#comment-14716152
 ] 

Chengxiang Li commented on FLINK-2545:
--------------------------------------

Hi, [~greghogan], do you mind to share how to reproduce this issue? The 
exception caused by a negative array index, which represents the count of 
bucket member. It should be very simple to add a check of the count value to 
fix this issue. But the count should never be negative according to the 
hashtable implementation, so if it's possible, i want to reproduce this issue 
to check if there is other hidden issues behind this.

> NegativeArraySizeException while creating hash table bloom filters
> ------------------------------------------------------------------
>
>                 Key: FLINK-2545
>                 URL: https://issues.apache.org/jira/browse/FLINK-2545
>             Project: Flink
>          Issue Type: Bug
>          Components: Distributed Runtime
>    Affects Versions: master
>            Reporter: Greg Hogan
>            Assignee: Chengxiang Li
>
> The following exception occurred a second time when I immediately re-ran my 
> application, though after recompiling and restarting Flink the subsequent 
> execution ran without error.
> java.lang.Exception: The data preparation for task '...' , caused an error: 
> null
>       at 
> org.apache.flink.runtime.operators.RegularPactTask.run(RegularPactTask.java:465)
>       at 
> org.apache.flink.runtime.operators.RegularPactTask.invoke(RegularPactTask.java:354)
>       at org.apache.flink.runtime.taskmanager.Task.run(Task.java:581)
>       at java.lang.Thread.run(Thread.java:745)
> Caused by: java.lang.NegativeArraySizeException
>       at 
> org.apache.flink.runtime.operators.hash.MutableHashTable.buildBloomFilterForBucket(MutableHashTable.java:1160)
>       at 
> org.apache.flink.runtime.operators.hash.MutableHashTable.buildBloomFilterForBucketsInPartition(MutableHashTable.java:1143)
>       at 
> org.apache.flink.runtime.operators.hash.MutableHashTable.spillPartition(MutableHashTable.java:1117)
>       at 
> org.apache.flink.runtime.operators.hash.MutableHashTable.insertBucketEntry(MutableHashTable.java:946)
>       at 
> org.apache.flink.runtime.operators.hash.MutableHashTable.insertIntoTable(MutableHashTable.java:868)
>       at 
> org.apache.flink.runtime.operators.hash.MutableHashTable.buildInitialTable(MutableHashTable.java:692)
>       at 
> org.apache.flink.runtime.operators.hash.MutableHashTable.open(MutableHashTable.java:455)
>       at 
> org.apache.flink.runtime.operators.hash.ReusingBuildSecondHashMatchIterator.open(ReusingBuildSecondHashMatchIterator.java:93)
>       at 
> org.apache.flink.runtime.operators.JoinDriver.prepare(JoinDriver.java:195)
>       at 
> org.apache.flink.runtime.operators.RegularPactTask.run(RegularPactTask.java:459)
>       ... 3 more



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to