Brock,

Doesn't this mean it is limited to 32k?

https://github.com/apache/flume/blob/branch-0.9.5/flume-core/src/main/java/com/cloudera/flume/core/EventImpl.java#L39


On Tue, May 8, 2012 at 11:01 AM, Brock Noland <[email protected]> wrote:

> Hi,
>
> On Tue, May 8, 2012 at 8:51 AM, S Ahmed <[email protected]> wrote:
> > Could flume be used to store log data where payload sizes are 2-100K in
> > size? (with the average being 25KB).
>
> Yes Flume could handle that size.
>
> >
> > This data is collected from a website, so each request there will be a
> > payload that needs to be stored (it is in XML format) and then every x
> > seconds it should be persisted from memory to mysql/hdfs/disk.
> >
> > Is this an appropriate use case for flume?
>
> Sounds like it to me.
>
> >
> > Can flume persist to things other than hdfs like mysql?
>
> You'd have to write a  MySQLSink or JDBCSink but yes that should work.
>
> Brock
>
> --
> Apache MRUnit - Unit testing MapReduce -
> http://incubator.apache.org/mrunit/
>

Reply via email to