[
https://issues.apache.org/jira/browse/HDFS-5224?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chris Nauroth updated HDFS-5224:
--------------------------------
Attachment: HDFS-5224.2.patch
Here is version 2 of the patch.
bq. Is it acceptable to have the descriptor's internal representation still be
a String, but then turn it into a Path for the client in the getter?
I wound up not doing this, because it didn't seem to simplify anything in
practice. On the client side, it causes an extra trip through the {{Path}}
constructor. On the server side, code like {{CacheManager}} still ends up
needing to turn the {{Path}} back into a {{String}}.
The fundamental problem here is that the client wants a URI, and the NN wants
just a path. If we don't want to define different data types to represent that
difference, then we need to overload the meaning of something. Putting a
{{Path}} in the object and unpacking in a few places in the NN seems like the
least intrusive way to do that, so I'm OK with this. Let me know what you
think.
> Refactor PathBasedCache* methods to use a Path rather than a String
> -------------------------------------------------------------------
>
> Key: HDFS-5224
> URL: https://issues.apache.org/jira/browse/HDFS-5224
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: datanode, namenode
> Affects Versions: HDFS-4949
> Reporter: Andrew Wang
> Assignee: Chris Nauroth
> Attachments: HDFS-5224.1.patch, HDFS-5224.2.patch
>
>
> As discussed in HDFS-5213, we should refactor PathBasedCacheDirective and
> related methods in DistributedFileSystem to use a Path to represent paths to
> cache, rather than a String.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
