[
https://issues.apache.org/jira/browse/HDFS-3246?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16786806#comment-16786806
]
Steve Loughran commented on HDFS-3246:
--------------------------------------
bq. IMO, we should support ByteBuffer read & pread in all FileSystem ?
better: HBase handles not having this.
We looked in HADOOP-14603 at adding the API To a different FS (here, the S3A
connector). Unless the underlying libraries that the store uses support this,
then its a lot of work in the connectors -and for what? Because it isn't going
to delver the speedups you expect.
Better: if a stream offers ByteBufferReadable, it is declaring that it offers
an efficient way to copy data. If a stream does not, then it is declaring that
it doesn't.
We can add a StreamCapabilities probe so you can check to see if a stream
supports it without having to make an API Call which goes on to fail -this will
let you check earlier. Would that help?
> pRead equivalent for direct read path
> -------------------------------------
>
> Key: HDFS-3246
> URL: https://issues.apache.org/jira/browse/HDFS-3246
> Project: Hadoop HDFS
> Issue Type: Improvement
> Components: hdfs-client, performance
> Affects Versions: 3.0.0-alpha1
> Reporter: Henry Robinson
> Assignee: Sahil Takiar
> Priority: Major
> Attachments: HDFS-3246.001.patch, HDFS-3246.002.patch,
> HDFS-3246.003.patch, HDFS-3246.004.patch, HDFS-3246.005.patch,
> HDFS-3246.006.patch
>
>
> There is no pread equivalent in ByteBufferReadable. We should consider adding
> one. It would be relatively easy to implement for the distributed case
> (certainly compared to HDFS-2834), since DFSInputStream does most of the
> heavy lifting.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
