You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@carbondata.apache.org by "Ajantha Bhat (Jira)" <ji...@apache.org> on 2021/03/17 05:47:00 UTC

[jira] [Commented] (CARBONDATA-3670) Support compress offheap columnpage directly, avoding a copy of data from offhead to heap when compressed.

    [ https://issues.apache.org/jira/browse/CARBONDATA-3670?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17303118#comment-17303118 ] 

Ajantha Bhat commented on CARBONDATA-3670:
------------------------------------------

Already handled in 2.0 from https://github.com/apache/carbondata/pull/3638

> Support compress offheap columnpage directly, avoding a copy of data from offhead to heap when compressed.
> ----------------------------------------------------------------------------------------------------------
>
>                 Key: CARBONDATA-3670
>                 URL: https://issues.apache.org/jira/browse/CARBONDATA-3670
>             Project: CarbonData
>          Issue Type: Wish
>          Components: core
>    Affects Versions: 2.0.0
>            Reporter: Xingjun Hao
>            Priority: Minor
>             Fix For: 2.1.1
>
>          Time Spent: 4h 10m
>  Remaining Estimate: 0h
>
> When writing data, the columnpages are stored on the offheap,  the pages will be compressed to save storage cost. Now, in the compression processing, the data will be copied from the offheap to the heap before compressed, which leads to heavier GC overhead compared with compress offhead directly.
> To sum up, we support compress offheap columnpage directly, avoding a copy of data from offhead to heap when compressed.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)