flume-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Otis Gospodnetic <otis_gospodne...@yahoo.com>
Subject Re: FlumeBase state, future, etc.
Date Tue, 27 Sep 2011 14:13:23 GMT
Hello,

Thanks Eric, this is valuable feedback!
We've included FlumeBase in our little ESP matrix here: http://blog.sematext.com/2011/09/26/event-stream-processor-matrix/

Now that I look ed FlumeBase on Github, it looks pretty "dead" - no commits in 3-4 months,
0 pull requests, just 1 issue...  Does anyone know is Aaron Kimball is still working on it?

Thanks,
Otis
----
Sematext :: http://sematext.com/ :: Solr - Lucene - Hadoop - HBase
Hadoop ecosystem search :: http://search-hadoop.com/


----- Original Message -----
> From: Eric Hauser <ewhauser@gmail.com>
> To: flume-user@incubator.apache.org; Otis Gospodnetic <otis_gospodnetic@yahoo.com>
> Cc: 
> Sent: Tuesday, September 20, 2011 9:58 AM
> Subject: Re: FlumeBase state, future, etc.
> 
> Otis,
> 
> I don't have answers to your questions, but I can share some of our
> experiences experimenting with FlumeBase:
> 
> 1. Beyond just the issues in FlumeBase itself, there are a number of
> issues with Flume 0.9.4 (FLUME-626, FLUME-706) that make it unusable
> with FlumeBase.  (I'd argue that Flume 0.9.4 is unusable by itself due
> to critical bugs, but that is another story)
> 2. We ran into a lot of issues trying to run Avro encoding events
> through FlumeBase.  Aaron was responsive in helping our engineer
> resolve the issues, but the error handling certainly did not give us a
> lot of confidence that FlumeBase was ready for primetime yet
> 3. Personally, I'd rather see the FlumeBase's execution engine stand
> on its own and not be coupled to Flume.  Since Kafka is a pub/sub
> model, integration there is much easier if FlumeBase's execution
> engine is embedded in any app.  You could probably do FlumeBase and
> Scribe together right now since Flume has a Scribe facade.
> 
> Although GPL and thus maybe not appropriate depending on your use
> case, Esper is much more mature and feature complete than FlumeBase.
> In my limited use, I have had some good experiences with it.
> 
> On Tue, Sep 20, 2011 at 12:00 AM, Otis Gospodnetic
> <otis_gospodnetic@yahoo.com> wrote:
>>  Hello,
>>  Sorry if this is not the ideal place to ask about FlumeBase, but I could 
> not
>>  find a better place and I know Aaron is on the list here.
>>  I'm evaluating stream even processors and of course looking at 
> FlumeBase,
>>  which is now at 0.2.0.
>>  Things I'm wondering about:
>>  * Is FlumeBase a 1-man-Aaron show?  Or maybe there is a plan to incubate at
>>  ASF in 2011/12?
>>  * Is there a new release coming up soon?  I think the last one was in June
>>  and it came with a big "not stable!" warning (see Documentation 
> for 0.2.0)
>>  * Is the plan to stay tightly integrated with Flume or allow integration
>>  with things like Kafka or Scribe or ...?
>>  Thanks,
>>  Otis
>>  ----
>>  Sematext :: http://sematext.com/ :: Solr - Lucene - Hadoop - HBase
>>  Hadoop ecosystem search :: http://search-hadoop.com/
>

Mime
View raw message