hadoop-common-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "churro morales (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (HADOOP-13578) Add Codec for ZStandard Compression
Date Mon, 26 Sep 2016 22:05:22 GMT

    [ https://issues.apache.org/jira/browse/HADOOP-13578?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15524323#comment-15524323
] 

churro morales edited comment on HADOOP-13578 at 9/26/16 10:04 PM:
-------------------------------------------------------------------

[~jlowe] Thank you for taking a look.
bq. The BUILDING.txt addition must have been copied from snappy, since a lot of it still refers
to snappy.
Yes did a copy and paste with the Building.txt, will take a look.  

bq.  There's a TODO in the code that should get sorted out:
Yes I think I have it figured out, had to read up a bit :). Since we compress the entire buffer
at once, that means it's transformed into a single ZStandard Frame. 
So the overhead is a Magic # (4 bytes) + Header (14 bytes max) + Checksum (optional 4 bytes).
So I believe that can be a constant of 22 Bytes because of the way we compress in hadoop.
 I'll update that as well. 

bq. Shouldn't this come from IO_COMPRESSION_CODEC_ZSTD_LEVEL_DEFAULT that's already in the
config keys?
Correct!, will change that back to 3.  Also will use the constant.  

I'll incorporate your fixes.  Comments much appreciated. 






was (Author: churromorales):
[~jlowe] Thank you for taking a look.
bq. The BUILDING.txt addition must have been copied from snappy, since a lot of it still refers
to snappy.
Yes did a copy and paste with the Building.txt, will take a look.  

bq.  There's a TODO in the code that should get sorted out:
Yes I think I have it figured out, had to read up a bit :). Since we empty compress the entire
buffer at once, that means it's transformed into a single ZStandard Frame. 
So the overhead is a Magic # (4 bytes) + Header (14 bytes max) + Checksum (optional 4 bytes).
So I believe that can be a constant of 22 Bytes because of the way we compress in hadoop.
 I'll update that as well. 

bq. Shouldn't this come from IO_COMPRESSION_CODEC_ZSTD_LEVEL_DEFAULT that's already in the
config keys?
Correct!, will change that back to 3.  Also will use the constant.  

I'll incorporate your fixes.  Comments much appreciated. 





> Add Codec for ZStandard Compression
> -----------------------------------
>
>                 Key: HADOOP-13578
>                 URL: https://issues.apache.org/jira/browse/HADOOP-13578
>             Project: Hadoop Common
>          Issue Type: New Feature
>            Reporter: churro morales
>            Assignee: churro morales
>         Attachments: HADOOP-13578.patch
>
>
> ZStandard: https://github.com/facebook/zstd has been used in production for 6 months
by facebook now.  v1.0 was recently released.  Create a codec for this library.  



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

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


Mime
View raw message