[jira] [Commented] (HADOOP-14765) AdlFsInputStream to implement CanUnbuffer

["Atul Sikaria (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HADOOP-14765?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16170829#comment-16170829 ] 

Atul Sikaria commented on HADOOP-14765:
---------------------------------------

What is the semantic expectation of unbuffer()? The Interface doc doesn't say much:
https://hadoop.apache.org/docs/r2.8.0/api/org/apache/hadoop/fs/CanUnbuffer.html

Is reducing the buffer equivalent to flushing everything we have in memory, or is there more to it? Note that the sockets themselves are managed by the connection pool in the underlying HttpUrlConnection used - and not sure what it would buy anybody to close those sockets if they are not used (and would probably hurt if the connections were to be opened again).

So for now we can do a simple implementation that does flush() or nothing, but just trying to understand what else is desired.



> AdlFsInputStream to implement CanUnbuffer
> -----------------------------------------
>
>                 Key: HADOOP-14765
>                 URL: https://issues.apache.org/jira/browse/HADOOP-14765
>             Project: Hadoop Common
>          Issue Type: Sub-task
>          Components: fs/adl
>    Affects Versions: 2.8.0
>            Reporter: John Zhuge
>            Priority: Minor
>
> HBase relies on FileSystems implementing CanUnbuffer.unbuffer() to force input streams to free up remote connections (HBASE-9393Link). This works for HDFS, but not elsewhere.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

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