[ 
https://issues.apache.org/jira/browse/STORM-1504?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15128669#comment-15128669
 ] 

ASF GitHub Bot commented on STORM-1504:
---------------------------------------

Github user ptgoetz commented on a diff in the pull request:

    https://github.com/apache/storm/pull/1052#discussion_r51607023
  
    --- Diff: external/storm-hdfs/README.md ---
    @@ -315,6 +314,18 @@ An `org.apache.avro.Schema` object cannot be directly 
provided since it does not
     The AvroGenericRecordBolt expects to receive tuples containing an Avro 
GenericRecord that conforms to the provided
     schema.
     
    +To use this bolt you **must** register the appropriate Kryo serializers 
with your topology configuration.  A convenience
    +method is provided for this:
    +
    +```AvroGenericRecordBolt.addAvroKryoSerializations(conf);```
    +
    +By default Storm will use the ```GenericAvroSerializer``` to handle 
serialization.  This will work, but there are much 
    +faster options available if you can pre-define the schemas you will be 
using or utilize an external schema registry. An
    +implementation using the Confluent Schema Registry is provided, but others 
can be implemented and provided to Storm.
    +Please see the javadoc for classes in org.apache.storm.hdfs.avro for 
information about using the built-in options or
    +creating your own.
    +
    --- End diff --
    
    Something's off with the markdown formatting here, such that the content 
below is all rendered as code/preformatted.


> AvroGenericRecordBolt will not work in a topology with multiple workers
> -----------------------------------------------------------------------
>
>                 Key: STORM-1504
>                 URL: https://issues.apache.org/jira/browse/STORM-1504
>             Project: Apache Storm
>          Issue Type: Bug
>            Reporter: Aaron Dossett
>            Assignee: Aaron Dossett
>            Priority: Critical
>
> It receives Avro GenericRecord objects, which leads to serialization errors 
> in a multi-worker topology.  Specifically, GenericData.Record will throw a 
> serialization error.
> A serializer for this class should be added and registered.



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

Reply via email to