[
https://issues.apache.org/jira/browse/HDFS-2246?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13112187#comment-13112187
]
Suresh Srinivas commented on HDFS-2246:
---------------------------------------
Reposting a comment that was dropped due to Jira reindexing.
Comment from Milind on HDFS-7668:
bq. Suresh, I think you want a check to see if the same file representing a
block is accessible from the client as well as the datanode. (Checking that
both client and datanode is on the same host, is an overly restrictive check.
In some ways, even if both client and datanode are run on the same node, with
chrooted containers, this is an overly general check.) Right ?
Currently based on the local host check, we might qualify a read for short
circuit, on a multi container system. Short circuit read fails for such blocks
and would eventually try reading the block over datanode, which succeeds.
However the problem is, an unnecessary attempt to read locally is made, slowing
down reads.
I cannot think of a good solution. Should we compare the local hostname of the
client with that of the hostname of the datanode, and allow the short circuit
reads? The problem is, this might be too strict and may miss valid short
circuit reads.
> Shortcut a local client reads to a Datanodes files directly
> -----------------------------------------------------------
>
> Key: HDFS-2246
> URL: https://issues.apache.org/jira/browse/HDFS-2246
> Project: Hadoop HDFS
> Issue Type: Improvement
> Reporter: Sanjay Radia
> Attachments: 0001-HDFS-347.-Local-reads.patch, HDFS-2246.20s.1.patch,
> HDFS-2246.20s.2.txt, HDFS-2246.20s.3.txt, HDFS-2246.20s.4.txt,
> HDFS-2246.20s.patch, localReadShortcut20-security.2patch
>
>
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
