Shawn Chang created HUDI-5936:
---------------------------------

             Summary: Potential serialization issue when FileStatus is not 
serializable
                 Key: HUDI-5936
                 URL: https://issues.apache.org/jira/browse/HUDI-5936
             Project: Apache Hudi
          Issue Type: Bug
            Reporter: Shawn Chang
            Assignee: Shawn Chang


Hadoop3's FileStatus is serializable and won't have this issue. However, when 
users run Hudi on older Hadoop or customized FileSystem implementation whose 
FileStatus is not serializable then it's possible to run into serialization 
issue.

 

Exception:
{code:java}
com.esotericsoftware.kryo.KryoException: 
java.util.ConcurrentModificationException Caused by: 
java.util.ConcurrentModificationException at 
java.util.Vector$Itr.checkForComodification(Vector.java:1212 ) at 
java.util.Vector$Itr.next(Vector.java:1165 ) at 
com.esotericsoftware.kryo.serializers.CollectionSerializer.write(CollectionSerializer.java:99
 ) at 
com.esotericsoftware.kryo.serializers.CollectionSerializer.write(CollectionSerializer.java:40
 ) at com.esotericsoftware.kryo.Kryo.writeObject(Kryo.java:575 ) at 
com.esotericsoftware.kryo.serializers.ObjectField.write(ObjectField.java:79 ) 
{code}
 

The LOC that causes this issue: 
https://github.com/apache/hudi/blob/master/hudi-common/src/main/java/org/apache/hudi/metadata/FileSystemBackedTableMetadata.java#L109



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to