reef-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Markus Weimer (JIRA)" <>
Subject [jira] [Commented] (REEF-1892) HDFS File Copy only uses local HDFS
Date Sat, 14 Oct 2017 16:10:00 GMT


Markus Weimer commented on REEF-1892:


Would we add this interface to REEF.Network?

Probably {{REEF.IO}}, but that's the idea, yes.

Markus Weimer, do you have a sense of which public interfaces would need to change?

Thanks to the genius of Tang: None.

Instead, we'd define a new interface {{IURLFormatter}} together with a default implementation,
e.g. {{ToStringURLFormatter}} which merely calls {{URL.ToString()}}. The HDFS client implementation
then gets a constructor parameter of type {{IURLFormatter}} and uses that instance in lieu
of calling {{.ToString()}} itself.

On clusters which require special treatment for URLs, we'd create a specific implementation
of {{IURLFormatter}} and configure Tang to provide that as the constructor parameter to the
HDFS client.

> HDFS File Copy only uses local HDFS
> -----------------------------------
>                 Key: REEF-1892
>                 URL:
>             Project: REEF
>          Issue Type: Bug
>          Components: REEF.NET IO
>    Affects Versions: 0.17
>            Reporter: Rogan Carr
> In REEF-1827 [1], the URI used to specify remote and local files were changed to use
the "AbsolutePath". [2]
> This means that a file specified as "hdfs://my/file" becomes "/my/file" and the hdfs://
is assumed by the `dfs` command.
> This is fine if you are using vanilla HDFS, but for cases like Blob Storage in Azure,
there is a special prefix, `wasb://` that is used instead of `hdfs://`. This means that the
AbsolutePath method trims off the wasb, and this Copy() function instead attempts to download
the file from the local HDFS instead of WASB.
> We need to revisit this issue and keep the full path for copies while also keeping proper
casing in the path.

This message was sent by Atlassian JIRA

View raw message