beam-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Aviem Zur (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (BEAM-1810) Spark runner combineGlobally uses Kryo serialization
Date Sat, 25 Mar 2017 19:56:41 GMT

     [ https://issues.apache.org/jira/browse/BEAM-1810?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Aviem Zur updated BEAM-1810:
----------------------------
    Description: 
{{TransformTranslator#combineGlobally}} inadvertently uses Kryo serialization to serialize
data. This should never happen.

This is due to Spark's implementation of {{RDD.isEmpty}} which takes the first element to
check if there are elements in the {{RDD}}, it is then serialized with Spark's configured
serializer (In our case, Kryo).

  was:
{{TransformTranslator#combineGlobally}} uses Kryo serialization to serialize data. This should
never happen.

This is due to Spark's implementation of {{RDD.isEmpty}} which takes the first element to
check if there are elements in the {{RDD}}, it is then serialized with Spark's configured
serializer (In our case, Kryo).


> Spark runner combineGlobally uses Kryo serialization
> ----------------------------------------------------
>
>                 Key: BEAM-1810
>                 URL: https://issues.apache.org/jira/browse/BEAM-1810
>             Project: Beam
>          Issue Type: Bug
>          Components: runner-spark
>            Reporter: Aviem Zur
>            Assignee: Aviem Zur
>
> {{TransformTranslator#combineGlobally}} inadvertently uses Kryo serialization to serialize
data. This should never happen.
> This is due to Spark's implementation of {{RDD.isEmpty}} which takes the first element
to check if there are elements in the {{RDD}}, it is then serialized with Spark's configured
serializer (In our case, Kryo).



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message