hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrew Wang (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HDFS-8598) Add and optimize for get LocatedFileStatus in DFSClient
Date Thu, 18 Jun 2015 16:12:00 GMT

    [ https://issues.apache.org/jira/browse/HDFS-8598?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14592038#comment-14592038
] 

Andrew Wang commented on HDFS-8598:
-----------------------------------

Sorry if I missed something, doesn't DFS#listLocatedStatus already batch up calls via the
DirListingIterator? It shouldn't be one RPC per file.

We also generally don't want listing APIs that aren't iterators. Forming up a big listing
can be very expensive on the NN, which impacts other clients.

> Add and optimize for get LocatedFileStatus  in DFSClient
> --------------------------------------------------------
>
>                 Key: HDFS-8598
>                 URL: https://issues.apache.org/jira/browse/HDFS-8598
>             Project: Hadoop HDFS
>          Issue Type: Improvement
>            Reporter: Yong Zhang
>            Assignee: Yong Zhang
>         Attachments: HDFS-8598.001.patch
>
>
> If we want to get all files block locations in one directory, we have to call getFileBlockLocations
for each file, it will take long time because of too many request. 
> LocatedFileStatus has block location, but we can find it also call getFileBlockLocations
 for each file in DFSClient. this jira is trying to optimize with only one RPC. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message