hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Elek, Marton (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDDS-732) Add read method which takes offset and length in SignedChunkInputStream
Date Thu, 15 Nov 2018 11:37:00 GMT

    [ https://issues.apache.org/jira/browse/HDDS-732?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16687894#comment-16687894
] 

Elek, Marton commented on HDDS-732:
-----------------------------------

Thanks [~candychencan] to work on this issue.

I executed the unit tests and they work well but I have two comments:

1. First of all the read(byte b[], int off, int len) call could be more faster than read()
in some cases (depends from the underlying InputStream). If I understood well the original
comment from [~bharatviswa], the goal here is to use the read(byte[],int,int) on the originalStream
in our read(byte[],int,int), which is added in this patch very well.

I think it could be done with adjusting the line 82 to the read max(remainingData, currentLen)
number of bytes with read(byte[],int,int)

2. The other minor: Even if we will have this excellent read(byte[],int,int) some stupid client
may use the simple read(). With the modification in the unit test we test only the new read
method. I would test both of the reads.

> Add read method which takes offset and length in SignedChunkInputStream
> -----------------------------------------------------------------------
>
>                 Key: HDDS-732
>                 URL: https://issues.apache.org/jira/browse/HDDS-732
>             Project: Hadoop Distributed Data Store
>          Issue Type: Sub-task
>            Reporter: Bharat Viswanadham
>            Assignee: chencan
>            Priority: Major
>         Attachments: HDDS-732.001.patch, HDDS-732.002.patch
>
>
> This Jira is created from the comments in HDDS-693
>  
> {quote}We have only read(), we don't have read(byte[] b, int off, int len), we might
see some slow operation during put with SignedInputStream.  
> {quote}
> 100% agree. I didn't check any performance numbers, yet, but we need to do it sooner
or later. I would implement this method in a separate jira as it adds more complexity and
as of now I would like to support the mkdir operations of the s3a unit tests (where the size
is 0).



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

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


Mime
View raw message