It can be sequence file datastream and compressed stream as per format of your source of data.

Sincerely,

Robin Jain

On Sep 14, 2015, at 12:47 PM, Buntu Dev <buntudev@gmail.com> wrote:

Don't mean to hijack this thread but I've some issue along the same lines -- does the hdfs.fileType need to be set to DataStream even if the source (kafka in my case instead of spooldir) data is Avro? 

On Mon, Sep 14, 2015 at 12:35 PM, Robin Jain <er.robinjain@gmail.com> wrote:
Hi Darshan,

Define hdfs.fileType parameter in your config file. If its a simple data file put the value as DataStream

Sincerely,

Robin Jain

> On Sep 14, 2015, at 12:00 PM, Darshan Pandya <dpandya@zaloni.com> wrote:
>
> I am trying to use the spooling source and spool over some text data into HDFS using the following agent configuration.
> <image.png>
>
>
> In HDFS I get to see Timestamp and ASCII characters only. I expect to see the real data. Any suggestions ?
>
> Real data looks like

>
>
> What I seeing in HDFS is this
>
> <image.png>
> --
> Sincerely,
> Darshan