[
https://issues.apache.org/jira/browse/HIVE-7711?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14123099#comment-14123099
]
Dr. Christian Betz commented on HIVE-7711:
------------------------------------------
Hi,
sorry for answering so late, but I had some other issues to work on more
urgently. I really appreciate you fast answer though :)
# I did check sourcecode of hive and kryo (in addition to documentation) and
found a solution/workaround: Just providing my own *"does nothing"-Serializer*
and "attaching" it to my GenericUDF-Class using the *@DefaultSerializer*
annotation (see https://github.com/EsotericSoftware/kryo#default-serializers)
solved the problem.
\\
I do not have my a code snippet right now to include/attach, but it basically
is an implementation of *com.esotericsoftware.kryo.Serializer* where the write
method does nothing and the read method returns a new instance of the class
given in the third parameter.
# I reviewed the patch you provided but did not build my own hive version from
it. It is basically not a solution to the problem, but emits a warning for a
missing serializer. As neither GenericUDF nor my implementation of GenericUDF
implement serializable I'm pretty sure I wouldn't see a warning message though.
However, the idea of generating a warning message is good. Maybe you could
think about implementing the above mentioned "does-nothing"-serializer for
GenericUDF (and UDF), as that mimics the behaviour of the orignal code before
you moved to kryo.
# I'm totally missing advanced documentation for all of this. I did not find
anything about "GenericUDFs should be serializable" on the web or in the
GenericUDF-source code. I think there should be something on that somewhere,
but I leave it up to you to decide on a new issue / subtask for that. However,
thanks in advance if you do it ;)
For me, you could well close that issue as my code is working right now,
however I would strongly advise to improve your codebase, your tests, and your
documentation on this issue.
> Error Serializing GenericUDF
> ----------------------------
>
> Key: HIVE-7711
> URL: https://issues.apache.org/jira/browse/HIVE-7711
> Project: Hive
> Issue Type: Bug
> Affects Versions: 0.13.0
> Reporter: Dr. Christian Betz
> Attachments: HIVE-7711.1.patch.txt
>
>
> I get an exception running a job with a GenericUDF in HIVE 0.13.0 (which was
> ok in HIVE 0.12.0).
> The org.apache.hadoop.hive.ql.plan.ExprNodeGenericFuncDesc is serialized
> using Kryo, trying to serialize stuff in my GenericUDF which is not
> serializable (doesn't implement Serializable).
> Switching to Kryo made the comment in ExprNodeGenericFuncDesc obsolte:
> "/**
> * In case genericUDF is Serializable, we will serialize the object.
> *
> * In case genericUDF does not implement Serializable, Java will remember
> the
> * class of genericUDF and creates a new instance when deserialized. This is
> * exactly what we want.
> */"
> Find the stacktrace below, however, the description above should be clear.
> Exception in thread "main"
> org.apache.hive.com.esotericsoftware.kryo.KryoException:
> java.lang.UnsupportedOperationException
> Serialization trace:
> value (java.util.concurrent.atomic.AtomicReference)
> state (clojure.lang.Atom)
> state (udfs.ArraySum)
> genericUDF (org.apache.hadoop.hive.ql.plan.ExprNodeGenericFuncDesc)
> colExprMap (org.apache.hadoop.hive.ql.exec.SelectOperator)
> childOperators (org.apache.hadoop.hive.ql.exec.TableScanOperator)
> aliasToWork (org.apache.hadoop.hive.ql.plan.MapWork)
> mapWork (org.apache.hadoop.hive.ql.plan.MapredWork)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.ObjectField.read(ObjectField.java:125)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.FieldSerializer.read(FieldSerializer.java:507)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readObject(Kryo.java:694)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.ObjectField.read(ObjectField.java:106)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.FieldSerializer.read(FieldSerializer.java:507)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readObject(Kryo.java:694)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.ObjectField.read(ObjectField.java:106)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.FieldSerializer.read(FieldSerializer.java:507)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readObject(Kryo.java:694)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.ObjectField.read(ObjectField.java:106)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.FieldSerializer.read(FieldSerializer.java:507)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readClassAndObject(Kryo.java:776)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.MapSerializer.read(MapSerializer.java:139)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.MapSerializer.read(MapSerializer.java:17)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readObject(Kryo.java:694)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.ObjectField.read(ObjectField.java:106)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.FieldSerializer.read(FieldSerializer.java:507)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readClassAndObject(Kryo.java:776)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.CollectionSerializer.read(CollectionSerializer.java:112)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.CollectionSerializer.read(CollectionSerializer.java:18)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readObject(Kryo.java:694)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.ObjectField.read(ObjectField.java:106)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.FieldSerializer.read(FieldSerializer.java:507)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readClassAndObject(Kryo.java:776)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.MapSerializer.read(MapSerializer.java:139)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.MapSerializer.read(MapSerializer.java:17)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readObject(Kryo.java:694)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.ObjectField.read(ObjectField.java:106)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.FieldSerializer.read(FieldSerializer.java:507)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readObject(Kryo.java:694)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.ObjectField.read(ObjectField.java:106)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.FieldSerializer.read(FieldSerializer.java:507)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readObject(Kryo.java:672)
> at
> org.apache.hadoop.hive.ql.exec.Utilities.deserializeObjectByKryo(Utilities.java:918)
> at
> org.apache.hadoop.hive.ql.exec.Utilities.deserializePlan(Utilities.java:826)
> at
> org.apache.hadoop.hive.ql.exec.Utilities.deserializePlan(Utilities.java:840)
> at
> org.apache.hadoop.hive.ql.exec.mr.ExecDriver.main(ExecDriver.java:738)
> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
> at
> sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
> at java.lang.reflect.Method.invoke(Method.java:606)
> at org.apache.hadoop.util.RunJar.main(RunJar.java:212)
> Caused by: java.lang.UnsupportedOperationException
> at clojure.lang.APersistentMap.put(APersistentMap.java:287)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.MapSerializer.read(MapSerializer.java:140)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.MapSerializer.read(MapSerializer.java:17)
> at
> org.apache.hive.com.esotericsoftware.kryo.Kryo.readObject(Kryo.java:694)
> at
> org.apache.hive.com.esotericsoftware.kryo.serializers.ObjectField.read(ObjectField.java:106)
> ... 41 more
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
