spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sean Owen (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SPARK-6156) Not cache in memory again if put memory_and_disk level block after put it in disk after unroll unsuccess in memory.
Date Wed, 04 Mar 2015 12:24:05 GMT

    [ https://issues.apache.org/jira/browse/SPARK-6156?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14346827#comment-14346827
] 

Sean Owen commented on SPARK-6156:
----------------------------------

It looked like you filed two versions of the same issue with the same start of the description.
I assumed it was a mistake; however on re-reading it still sounds like one issue you're reporting.

> Not cache in memory again if put memory_and_disk level block after put it in disk after
unroll unsuccess in memory.
> -------------------------------------------------------------------------------------------------------------------
>
>                 Key: SPARK-6156
>                 URL: https://issues.apache.org/jira/browse/SPARK-6156
>             Project: Spark
>          Issue Type: Improvement
>          Components: Block Manager
>    Affects Versions: 1.2.1
>            Reporter: SuYan
>            Priority: Minor
>
> Current code:
> Now we want to cache a Memory_and_Disk level block
> 1. Try to put in memory and unroll unsuccessful. then reserved unroll memory because
we got a iterator from an unroll Array 
> 2. Then put into disk.
> 3. Get value from get(blockId), and iterator from that value.
> And there have some doubts:
> for process 3:
> get value use blockManager.get(blockId)
> in get(blockId) method, it will first getLocal() and then getRemote() in logical, I think
is more suitable for use getLocal(blockId) to instead get(blockId), I dont know why we call
get(blockId), have other considerations?
> and in doGetLocal method, if it get value success from disk, it will try to put in memory
again... I think, that a redundant... we just try to put in memory and failed to unroll in
process 1.
> may this pull request is useless, because may have some reasons to do that?



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

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org


Mime
View raw message