[
https://issues.apache.org/jira/browse/HADOOP-3205?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12773199#action_12773199
]
Todd Lipcon commented on HADOOP-3205:
-------------------------------------
Looking at the source of BufferedInputStream (at
http://www.docjar.com/html/api/java/io/BufferedInputStream.java.html) it
actually seems like BufferedInputStream is already handling pass-through to the
underlying stream in the case that the read buffer is as large as its own
buffer. That was the crucial bit I was missing that explains why performing the
underlying reads in larger chunks would make a difference, even without
removing the BIS.
I'll give it a go and see if there is any discernible performance increase.
bq. Plus, looking at how long this jira has been open, it is no blocker
Of course :)
> FSInputChecker and FSOutputSummer should allow better access to user buffer
> ---------------------------------------------------------------------------
>
> Key: HADOOP-3205
> URL: https://issues.apache.org/jira/browse/HADOOP-3205
> Project: Hadoop Common
> Issue Type: Bug
> Components: fs
> Reporter: Raghu Angadi
> Assignee: Raghu Angadi
>
> Implementations of FSInputChecker and FSOutputSummer like DFS do not have
> access to full user buffer. At any time DFS can access only up to 512 bytes
> even though user usually reads with a much larger buffer (often controlled by
> io.file.buffer.size). This requires implementations to double buffer data if
> an implementation wants to read or write larger chunks of data from
> underlying storage.
> We could separate changes for FSInputChecker and FSOutputSummer into two
> separate jiras.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
