incubator-hama-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Edward J. Yoon" <edwardy...@apache.org>
Subject Re: Cutting a release 0.5-incubating
Date Tue, 06 Mar 2012 09:08:35 GMT
In my opinion, Hama seems reached to "consider-to-use" level, so now
quickly supporting various version of Hadoop for people who want to
use Hama with their existing cluster can be our top priority.

Of course, if something happened in the middle of Job processing, they
might **** us but better than no feedback.

On Tue, Mar 6, 2012 at 5:18 PM, Thomas Jungblut
<thomas.jungblut@googlemail.com> wrote:
> I have to admit that I never thought that Suraj is going to take over all
> the issues.
> https://issues.apache.org/jira/browse/HAMA-505
> 2/4 is done, and I guess there will be a fifth with the real rollback.
>
> I wanted to target June/Juli, which is 5-6 months since we actually planned
> to release 0.4.0.
>
> I think the harmony with our surroundings and environment is important.
>>
>
> Well actually you are right-
>
> Am 6. März 2012 09:14 schrieb Edward J. Yoon <edwardyoon@apache.org>:
>
>> If it takes 5 ~ 6 months[1], then we can release 0.5 in the autumn or
>> winter of this year.
>>
>> I'm not talking about Incubator Graduation. I think the harmony with
>> our surroundings and environment is important. Someone want to use
>> Graph APIs, other people want to use with Hadoop 0.20.205, In my case
>> I want to use Hama on Clouds, and Whirr/Bigtop teams are go faster
>> than us.
>>
>> 1. http://markmail.org/message/ziviktlm7yjihvew
>>
>> On Tue, Mar 6, 2012 at 5:00 PM, Thomas Jungblut
>> <thomas.jungblut@googlemail.com> wrote:
>> > Jukka told us recently that there is no technical requirement for
>> > graduation.
>> > However, the attention we get should be well used.
>> > We should target people that use Hama in production use and they really
>> > need some fault tolerance. (I plan it to use as well once we upgraded to
>> > YARN).
>> >
>> > Based on the new features we implemented in this few weeks, we could
>> slice
>> > a new release yes.
>> > But I don't see fault tolerance (at least checkpoint recovery) as a new
>> > feature rather than a must-have improvement.
>> >
>> > We have delayed the fault tolerance for now 2 releases, why delay it
>> > further?
>> >
>> > Am 6. März 2012 08:56 schrieb Edward J. Yoon <edwardyoon@apache.org>:
>> >
>> >> Agree.
>> >>
>> >> Sorry for my short mail, I was mean about new features. I think, we've
>> >> added enough new features (MSG Compressor, Avro RPC, and Graph API
>> >> package).
>> >>
>> >> Should we must add checkpoint recovery to 0.5 road-map?
>> >>
>> >> On Tue, Mar 6, 2012 at 4:43 PM, Thomas Jungblut
>> >> <thomas.jungblut@googlemail.com> wrote:
>> >> > Hy,
>> >> >
>> >> > we haven't released 0.4.0 at all yet, so why don't we wait for bug
>> >> reports
>> >> > or feedback?
>> >> >
>> >> > Am 6. März 2012 01:14 schrieb Edward J. Yoon <edwardyoon@apache.org>:
>> >> >
>> >> >> Hi,
>> >> >>
>> >> >> What do you think about cutting a release 0.5-incubating with MSG
>> >> >> Compressor, Avro RPC, and Graph API package new features and Support
>> >> >> Hadoop 1.0?
>> >> >>
>> >> >> --
>> >> >> Best Regards, Edward J. Yoon
>> >> >> @eddieyoon
>> >> >>
>> >> >
>> >> >
>> >> >
>> >> > --
>> >> > Thomas Jungblut
>> >> > Berlin <thomas.jungblut@gmail.com>
>> >>
>> >>
>> >>
>> >> --
>> >> Best Regards, Edward J. Yoon
>> >> @eddieyoon
>> >>
>> >
>> >
>> >
>> > --
>> > Thomas Jungblut
>> > Berlin <thomas.jungblut@gmail.com>
>>
>>
>>
>> --
>> Best Regards, Edward J. Yoon
>> @eddieyoon
>>
>
>
>
> --
> Thomas Jungblut
> Berlin <thomas.jungblut@gmail.com>



-- 
Best Regards, Edward J. Yoon
@eddieyoon

Mime
View raw message