[ 
https://issues.apache.org/jira/browse/FLINK-8716?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16370037#comment-16370037
 ] 

Aljoscha Krettek commented on FLINK-8716:
-----------------------------------------

I'm starting to think that this might be on purpose. We can only always read 
with the newest schema because I think we would get into trouble with RocksDB 
where we don't always read/write all data in between doing savepoints. (You 
alluded to this in your mailing list e-mail.) If we do two schema updates with 
savepoints in between we will not have the first schema anymore but we might 
have data in RocksDB that was written using the first schema.

[~StephanEwen] What do you think? You reworked those parts recently.

> AvroSerializer does not use schema of snapshot
> ----------------------------------------------
>
>                 Key: FLINK-8716
>                 URL: https://issues.apache.org/jira/browse/FLINK-8716
>             Project: Flink
>          Issue Type: Bug
>    Affects Versions: 1.4.0
>            Reporter: Arvid Heise
>            Priority: Major
>
> The new AvroSerializer stores the schema in the snapshot and uses it to 
> validate compability.
> However, it does not use the schema of the snapshot while reading the data. 
> This version will fail for any change of the data layout (so it supports more 
> or less only renaming currently).
>  
> [https://github.com/apache/flink/blob/f3a2197a23524048200ae2b4712d6ed833208124/flink-formats/flink-avro/src/main/java/org/apache/flink/formats/avro/typeutils/AvroSerializer.java#L265]
>  needs to use the schema from
>  
> [https://github.com/apache/flink/blob/f3a2197a23524048200ae2b4712d6ed833208124/flink-formats/flink-avro/src/main/java/org/apache/flink/formats/avro/typeutils/AvroSerializer.java#L188]
>  as the first parameter. Accordingly, a readSchema field need to be set
>  in #ensureCompatibility and relayed in #duplicate. Note that the readSchema 
> is passed as the write schema parameter to the DatumReader, as it was the 
> schema that was used to write the data.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to