bigtop-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From jay vyas <jayunit100.apa...@gmail.com>
Subject Re: What will the next generation of bigtop look like?
Date Sat, 13 Dec 2014 04:06:30 GMT
Yeah, I think HDFS is pretty important, esp. if we want to use SolrCloud +
HBase, both of which need a good DFS
working under the hood.   So it sounds like the most important packages
I've heard so far are:

HDFS
yarn
Spark
HBase+Phoenix
Kafka
Solr

are the ones which seem to be most appealing.

On Fri, Dec 12, 2014 at 9:54 PM, Andrew Purtell <apurtell@apache.org> wrote:
>
> Well if HDFS is dropped, and all packages that depend on it, the
> usefulness of Bigtop for me drops to zero.
>
> On Fri, Dec 12, 2014 at 6:24 PM, jay vyas <jayunit100.apache@gmail.com>
> wrote:
>
>> youre right - this thread is just a discussion ... nothing at all has
>> been replaced, nor even proposed to be replaced.
>> I think the purpose of this thread is to discuss the **entire** bigtop
>> stack,  from a theoretical perspective... so there are no sacred cows :)...
>>
>>
>> On Fri, Dec 12, 2014 at 7:59 PM, Andrew Purtell <apurtell@apache.org>
>> wrote:
>>
>>> > thanks for the input maybe yarn and HDFS should continue to stick
>>> around
>>>
>>> Huh? When/where was HDFS not sticking around? To be replaced with what?
>>>
>>>
>>> On Thu, Dec 11, 2014 at 6:57 PM, jay vyas <jayunit100.apache@gmail.com>
>>> wrote:
>>>>
>>>> great feedback guys.  my thoughts:
>>>>
>>>> @andrew, yeah thats some very good points you make.  thanks for the
>>>> input maybe yarn and HDFS should continue to stick around.
>>>>
>>>> @RJ, i think python, and the mgmt tooling can be complimentary, but
>>>> more ** on top ** of bigtop, i.e. in a vendor product based on bigtop : or
>>>> a community repackaging of bigtop ---  managing them as part of bigtop
>>>> itself --  might be extra features are a little out of scope of bigtop ;
>>>> which is more around deployment and packaging / testing for the core of a
>>>> big data infrastructure, as opposed to an e2e solution.  but *at the least*
>>>> i think it makes sense to keep in mind the  ambaris and also the
>>>> ipythons/tableus  of the world when producing bigtop releases, b/c if those
>>>> represent use cases we can have good tests for (i.e. REST Apis and PySpark
>>>> tests and so on)...
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>> --
>>> Best regards,
>>>
>>>    - Andy
>>>
>>> Problems worthy of attack prove their worth by hitting back. - Piet Hein
>>> (via Tom White)
>>>
>>
>>
>>
>> --
>> jay vyas
>>
>
>
>
> --
> Best regards,
>
>    - Andy
>
> Problems worthy of attack prove their worth by hitting back. - Piet Hein
> (via Tom White)
>


-- 
jay vyas

Mime
View raw message