[jira] [Commented] (KAFKA-13386) Foreign Key Join filtering out valid records after a code change / schema evolved

["Guozhang Wang (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/KAFKA-13386?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17431420#comment-17431420 ] 

Guozhang Wang commented on KAFKA-13386:
---------------------------------------

Hello [~sduran] thanks for reporting your issue. My understanding is that when the serializer's behavior changes upon an upgrade, records may then be sending to a different partition after the upgrade, is that what you're suggesting? I'm not sure though how that would cause valid records to be filtered, could you elaborate a bit?

> Foreign Key Join filtering out valid records after a code change / schema evolved
> ---------------------------------------------------------------------------------
>
>                 Key: KAFKA-13386
>                 URL: https://issues.apache.org/jira/browse/KAFKA-13386
>             Project: Kafka
>          Issue Type: Bug
>          Components: streams
>    Affects Versions: 2.6.2
>            Reporter: Sergio Duran Vegas
>            Priority: Major
>
> The join optimization assumes the serializer is deterministic and invariant across upgrades. So in case of changes this opimitzation will drop invalid/intermediate records. In other situations we have relied on the same property, for example when computing whether an update is a duplicate result or not.
>  
> The problem is that some serializers are sadly not deterministic.
>  
> [https://github.com/apache/kafka/blob/trunk/streams/src/main/java/org/apache/kafka/streams/kstream/internals/foreignkeyjoin/SubscriptionResolverJoinProcessorSupplier.java]
>  
> {code:java}
> //If this value doesn't match the current value from the original table, it is stale and should be discarded.
>  if (java.util.Arrays.equals(messageHash, currentHash)) {{code}
>  
> A solution for this problem would be that the comparison use foreign-key reference itself instead of the whole message hash.
>  
> The bug fix proposal is to be allow the user to choose between one method of comparison or another (whole hash or Fk reference). This would fix the problem of dropping valid records on certain cases and allow the user to also choose the current optimized way of checking valid records and intermediate results dropping.
>  
>  
>  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)