[
https://issues.apache.org/jira/browse/FLINK-2606?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14728711#comment-14728711
]
Matthias J. Sax commented on FLINK-2606:
----------------------------------------
It just came to my mind, that there exists a storm-hdfs module already:
https://github.com/apache/storm/tree/master/external/storm-hdfs
Thus, I am not sure if Flink should provide an own HdfsSpout. It's not the
purpose of the compatibility layer to provide spout/bolt implementations with
all kind of features. The currently included spouts/bolt are actually provided
for testing and showcase purpose only.
Maybe we should just close this ticket?
> Extend StormFileSpout to handle HDFS
> ------------------------------------
>
> Key: FLINK-2606
> URL: https://issues.apache.org/jira/browse/FLINK-2606
> Project: Flink
> Issue Type: New Feature
> Components: Storm Compatibility
> Reporter: Matthias J. Sax
> Priority: Minor
>
> Currently, `StormFileSpout` uses a simple `FileReader` internally to read
> local files. Thus, it cannot read data from HDFS.
> With this ticket, either `StormFileSpout` should be extended to be able to
> handle HDFS files or a new class (eg, `StormHdfsSpout`) should be introduced
> for it.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
