[
https://issues.apache.org/jira/browse/HADOOP-13530?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15430335#comment-15430335
]
Steve Loughran commented on HADOOP-13530:
-----------------------------------------
Although we are removing s3, and leaving s3n alone to avoid breaking it, s3a is
undergoing lots of work. However, this feature isn't needed there; it's config
option {{fs.s3a.buffer.dir}} already takes a list, using {{LocalDirAllocator}}
for the same QoS as HDFS itself: round robin allocation, though looking at the
code there, it's doing it wrong (round-robin isn't being set up right).
For this JIRA, closing as a wontfix. If you are still using ASF s3://, time to
move to Hadoop 2.7+ and embrace s3a
> Upgrade S3 fs.s3.buffer.dir to support multi directories
> --------------------------------------------------------
>
> Key: HADOOP-13530
> URL: https://issues.apache.org/jira/browse/HADOOP-13530
> Project: Hadoop Common
> Issue Type: Improvement
> Components: fs/s3
> Affects Versions: 2.4.0
> Reporter: Adrian Muraru
> Assignee: Ted Malaska
> Priority: Minor
>
> fs.s3.buffer.dir defines the tmp folder where files will be written to before
> getting sent to S3. Right now this is limited to a single folder which
> causes to major issues.
> 1. You need a drive with enough space to store all the tmp files at once
> 2. You are limited to the IO speeds of a single drive
> This is similar to HADOOP-10610 but applies to {{s3://}} hadoop block fs.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
