beam-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Amit Sela (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (BEAM-918) Let users set STORAGE_LEVEL via SparkPipelineOptions.
Date Tue, 15 Nov 2016 11:59:59 GMT

    [ https://issues.apache.org/jira/browse/BEAM-918?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15666959#comment-15666959
] 

Amit Sela commented on BEAM-918:
--------------------------------

[~jbonofre] you're good to go ;)

> Let users set STORAGE_LEVEL via SparkPipelineOptions.
> -----------------------------------------------------
>
>                 Key: BEAM-918
>                 URL: https://issues.apache.org/jira/browse/BEAM-918
>             Project: Beam
>          Issue Type: Improvement
>          Components: runner-spark
>            Reporter: Amit Sela
>            Assignee: Jean-Baptiste Onofré
>
> Spark provides different "STORAGE_LEVEL"s for caching RDDs (disk, memory, ser/de, etc.).
> The runner decides on caching when necessary, for example when a RDD is accessed repeatedly.
>  
> For batch, we can let users set their preferred STORAGE_LEVEL via SparkPipelineOptions.
> Note: for streaming we force a "MEMORY_ONLY", since (among other things) the runner heavily
relies on stateful operations, and it's natural for streaming to happen in memory.



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

Mime
View raw message