You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hama.apache.org by "Edward J. Yoon" <ed...@udanax.org> on 2008/06/09 06:59:05 UTC
Dependencies management and Distributions
Hi all,
I think we need to focus on how we are going to distribute the Hama
and to manage the dependencies.
To access to a hadoop cluster, I would propose to roughly follow Apache PIG.
On the other hand, I started considering Maven with [HADOOP-3302
Support Maven-based builds].
What do you think?
--
Best regards,
Edward J. Yoon,
http://blog.udanax.org
Re: Dependencies management and Distributions
Posted by "Edward J. Yoon" <ed...@udanax.org>.
> I am not against HBase but I just think to be the least dependent is the best.
I agree with you, BTW, I think we can't avoid the communication
between machines while some matrix operations are performed. Then,
Hbase will be useful.
I'm not sure yet so require confirmation by a lot of test. Let's have
a try at it. :)
Regards, Edward
> 2008/6/9 pi song <pi...@gmail.com>:
> Again, my question is "Does HBase really add value to matrix processing?". I
> doubt that people wanting to run matrix processing on distributed nodes
> would really require low latency. AFAIK, you just use Hbase column family
> to do horizontal partitioning. This should be replaceable by a simple
> abstraction. I am not against HBase but I just think to be the least
> dependent is the best.
>
> Pi
>
> On Mon, Jun 9, 2008 at 6:46 PM, Edward J. Yoon <ed...@udanax.org> wrote:
>
>> No, problem.
>>
>> But, We can't remove it. :) So, You should be more careful.
>>
>> Regards, Edward
>>
>> 2008/6/9 Donguk Choi <al...@gmail.com>:
>> > Oops, I was copy&paste the message of edward from IRC. Sorry for my rash
>> > mailing.
>> >
>> >
>> > On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
>> >>
>> >> Again, We hope that the Hama can be adaptable to other distributed
>> >> computing environments (e.g. Hypertable, KFS, ..., etc). However,
>> >> Let's not rush; do it step by step at the moment.
>> >>
>> >> Regards, Edward
>> >>
>> >> On Mon, Jun 9, 2008 at 5:14 PM, Edward J. Yoon <ed...@udanax.org>
>> wrote:
>> >> > Hi π,
>> >> >
>> >> > Hama uses some external libraries (e.g. commons-*, junit, log4j)
>> >> > besides Hadoop and Hbase.
>> >> >
>> >> > Regards, Edward
>> >> >
>> >> > On Mon, Jun 9, 2008 at 5:00 PM, pi song <pi...@gmail.com> wrote:
>> >> >> What are your dependencies so far? Only Hadoop and HBase?
>> >> >>
>> >> >> On Mon, Jun 9, 2008 at 3:41 PM, Edward J. Yoon <ed...@udanax.org>
>> >> wrote:
>> >> >>
>> >> >>> > BTW, When HAMA's prototype source update to svn server?
>> >> >>>
>> >> >>> I think we need to get SVN accounts.
>> >> >>>
>> >> >>> On Mon, Jun 9, 2008 at 2:36 PM, Donguk Choi <al...@gmail.com>
>> >> wrote:
>> >> >>> > hi all.
>> >> >>> >
>> >> >>> > +1 about the following PIG. Then We can deal with the system
>> sharing
>> >> >>> problem
>> >> >>> > flexibly with the existing Hadoop users. BTW, When HAMA's
>> prototype
>> >> >>> source
>> >> >>> > update to svn server?
>> >> >>> >
>> >> >>> >
>> >> >>> >
>> >> >>> > On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
>> >> >>> >>
>> >> >>> >> IMO, I prefer the latter opinion to the former. But, Considering
>> the
>> >> >>> >> pre-existing hadoop cluster, The former is better than the
>> latter.
>> >> >>> >>
>> >> >>> >> > To access to a hadoop cluster, I would propose to roughly
>> follow
>> >> >>> Apache
>> >> >>> >> PIG.
>> >> >>> >>
>> >> >>> >> +1
>> >> >>> >>
>> >> >>> >> Edward
>> >> >>> >>
>> >> >>> >> On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <
>> edward@udanax.org>
>> >> >>> wrote:
>> >> >>> >> > Hi all,
>> >> >>> >> >
>> >> >>> >> > I think we need to focus on how we are going to distribute the
>> >> Hama
>> >> >>> >> > and to manage the dependencies.
>> >> >>> >> >
>> >> >>> >> > To access to a hadoop cluster, I would propose to roughly
>> follow
>> >> >>> Apache
>> >> >>> >> PIG.
>> >> >>> >> > On the other hand, I started considering Maven with
>> [HADOOP-3302
>> >> >>> >> > Support Maven-based builds].
>> >> >>> >> >
>> >> >>> >> > What do you think?
>> >> >>> >> > --
>> >> >>> >> > Best regards,
>> >> >>> >> > Edward J. Yoon,
>> >> >>> >> > http://blog.udanax.org
>> >> >>> >> >
>> >> >>> >>
>> >> >>> >>
>> >> >>> >>
>> >> >>> >> --
>> >> >>> >> Best regards,
>> >> >>> >> Edward J. Yoon,
>> >> >>> >> http://blog.udanax.org
>> >> >>> >>
>> >> >>> >
>> >> >>> >
>> >> >>> >
>> >> >>> > --
>> >> >>> > Best. Regards,
>> >> >>> > Donguk Choi
>> >> >>> > http://www.alloes.com
>> >> >>> >
>> >> >>>
>> >> >>>
>> >> >>>
>> >> >>> --
>> >> >>> Best regards,
>> >> >>> Edward J. Yoon,
>> >> >>> http://blog.udanax.org
>> >> >>>
>> >> >>
>> >> >
>> >> >
>> >> >
>> >> > --
>> >> > Best regards,
>> >> > Edward J. Yoon,
>> >> > http://blog.udanax.org
>> >> >
>> >>
>> >>
>> >>
>> >> --
>> >> Best regards,
>> >> Edward J. Yoon,
>> >> http://blog.udanax.org
>> >>
>> >
>> >
>> >
>> > --
>> > Best. Regards,
>> > Donguk Choi
>> > http://blog.naver.com/anurocky
>> >
>>
>>
>>
>> --
>> Best regards,
>> Edward J. Yoon,
>> http://blog.udanax.org
>>
>
--
Best regards,
Edward J. Yoon,
http://blog.udanax.org
Re: Dependencies management and Distributions
Posted by pi song <pi...@gmail.com>.
Again, my question is "Does HBase really add value to matrix processing?". I
doubt that people wanting to run matrix processing on distributed nodes
would really require low latency. AFAIK, you just use Hbase column family
to do horizontal partitioning. This should be replaceable by a simple
abstraction. I am not against HBase but I just think to be the least
dependent is the best.
Pi
On Mon, Jun 9, 2008 at 6:46 PM, Edward J. Yoon <ed...@udanax.org> wrote:
> No, problem.
>
> But, We can't remove it. :) So, You should be more careful.
>
> Regards, Edward
>
> 2008/6/9 Donguk Choi <al...@gmail.com>:
> > Oops, I was copy&paste the message of edward from IRC. Sorry for my rash
> > mailing.
> >
> >
> > On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
> >>
> >> Again, We hope that the Hama can be adaptable to other distributed
> >> computing environments (e.g. Hypertable, KFS, ..., etc). However,
> >> Let's not rush; do it step by step at the moment.
> >>
> >> Regards, Edward
> >>
> >> On Mon, Jun 9, 2008 at 5:14 PM, Edward J. Yoon <ed...@udanax.org>
> wrote:
> >> > Hi π,
> >> >
> >> > Hama uses some external libraries (e.g. commons-*, junit, log4j)
> >> > besides Hadoop and Hbase.
> >> >
> >> > Regards, Edward
> >> >
> >> > On Mon, Jun 9, 2008 at 5:00 PM, pi song <pi...@gmail.com> wrote:
> >> >> What are your dependencies so far? Only Hadoop and HBase?
> >> >>
> >> >> On Mon, Jun 9, 2008 at 3:41 PM, Edward J. Yoon <ed...@udanax.org>
> >> wrote:
> >> >>
> >> >>> > BTW, When HAMA's prototype source update to svn server?
> >> >>>
> >> >>> I think we need to get SVN accounts.
> >> >>>
> >> >>> On Mon, Jun 9, 2008 at 2:36 PM, Donguk Choi <al...@gmail.com>
> >> wrote:
> >> >>> > hi all.
> >> >>> >
> >> >>> > +1 about the following PIG. Then We can deal with the system
> sharing
> >> >>> problem
> >> >>> > flexibly with the existing Hadoop users. BTW, When HAMA's
> prototype
> >> >>> source
> >> >>> > update to svn server?
> >> >>> >
> >> >>> >
> >> >>> >
> >> >>> > On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
> >> >>> >>
> >> >>> >> IMO, I prefer the latter opinion to the former. But, Considering
> the
> >> >>> >> pre-existing hadoop cluster, The former is better than the
> latter.
> >> >>> >>
> >> >>> >> > To access to a hadoop cluster, I would propose to roughly
> follow
> >> >>> Apache
> >> >>> >> PIG.
> >> >>> >>
> >> >>> >> +1
> >> >>> >>
> >> >>> >> Edward
> >> >>> >>
> >> >>> >> On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <
> edward@udanax.org>
> >> >>> wrote:
> >> >>> >> > Hi all,
> >> >>> >> >
> >> >>> >> > I think we need to focus on how we are going to distribute the
> >> Hama
> >> >>> >> > and to manage the dependencies.
> >> >>> >> >
> >> >>> >> > To access to a hadoop cluster, I would propose to roughly
> follow
> >> >>> Apache
> >> >>> >> PIG.
> >> >>> >> > On the other hand, I started considering Maven with
> [HADOOP-3302
> >> >>> >> > Support Maven-based builds].
> >> >>> >> >
> >> >>> >> > What do you think?
> >> >>> >> > --
> >> >>> >> > Best regards,
> >> >>> >> > Edward J. Yoon,
> >> >>> >> > http://blog.udanax.org
> >> >>> >> >
> >> >>> >>
> >> >>> >>
> >> >>> >>
> >> >>> >> --
> >> >>> >> Best regards,
> >> >>> >> Edward J. Yoon,
> >> >>> >> http://blog.udanax.org
> >> >>> >>
> >> >>> >
> >> >>> >
> >> >>> >
> >> >>> > --
> >> >>> > Best. Regards,
> >> >>> > Donguk Choi
> >> >>> > http://www.alloes.com
> >> >>> >
> >> >>>
> >> >>>
> >> >>>
> >> >>> --
> >> >>> Best regards,
> >> >>> Edward J. Yoon,
> >> >>> http://blog.udanax.org
> >> >>>
> >> >>
> >> >
> >> >
> >> >
> >> > --
> >> > Best regards,
> >> > Edward J. Yoon,
> >> > http://blog.udanax.org
> >> >
> >>
> >>
> >>
> >> --
> >> Best regards,
> >> Edward J. Yoon,
> >> http://blog.udanax.org
> >>
> >
> >
> >
> > --
> > Best. Regards,
> > Donguk Choi
> > http://blog.naver.com/anurocky
> >
>
>
>
> --
> Best regards,
> Edward J. Yoon,
> http://blog.udanax.org
>
Re: Dependencies management and Distributions
Posted by "Edward J. Yoon" <ed...@udanax.org>.
No, problem.
But, We can't remove it. :) So, You should be more careful.
Regards, Edward
2008/6/9 Donguk Choi <al...@gmail.com>:
> Oops, I was copy&paste the message of edward from IRC. Sorry for my rash
> mailing.
>
>
> On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
>>
>> Again, We hope that the Hama can be adaptable to other distributed
>> computing environments (e.g. Hypertable, KFS, ..., etc). However,
>> Let's not rush; do it step by step at the moment.
>>
>> Regards, Edward
>>
>> On Mon, Jun 9, 2008 at 5:14 PM, Edward J. Yoon <ed...@udanax.org> wrote:
>> > Hi π,
>> >
>> > Hama uses some external libraries (e.g. commons-*, junit, log4j)
>> > besides Hadoop and Hbase.
>> >
>> > Regards, Edward
>> >
>> > On Mon, Jun 9, 2008 at 5:00 PM, pi song <pi...@gmail.com> wrote:
>> >> What are your dependencies so far? Only Hadoop and HBase?
>> >>
>> >> On Mon, Jun 9, 2008 at 3:41 PM, Edward J. Yoon <ed...@udanax.org>
>> wrote:
>> >>
>> >>> > BTW, When HAMA's prototype source update to svn server?
>> >>>
>> >>> I think we need to get SVN accounts.
>> >>>
>> >>> On Mon, Jun 9, 2008 at 2:36 PM, Donguk Choi <al...@gmail.com>
>> wrote:
>> >>> > hi all.
>> >>> >
>> >>> > +1 about the following PIG. Then We can deal with the system sharing
>> >>> problem
>> >>> > flexibly with the existing Hadoop users. BTW, When HAMA's prototype
>> >>> source
>> >>> > update to svn server?
>> >>> >
>> >>> >
>> >>> >
>> >>> > On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
>> >>> >>
>> >>> >> IMO, I prefer the latter opinion to the former. But, Considering the
>> >>> >> pre-existing hadoop cluster, The former is better than the latter.
>> >>> >>
>> >>> >> > To access to a hadoop cluster, I would propose to roughly follow
>> >>> Apache
>> >>> >> PIG.
>> >>> >>
>> >>> >> +1
>> >>> >>
>> >>> >> Edward
>> >>> >>
>> >>> >> On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <ed...@udanax.org>
>> >>> wrote:
>> >>> >> > Hi all,
>> >>> >> >
>> >>> >> > I think we need to focus on how we are going to distribute the
>> Hama
>> >>> >> > and to manage the dependencies.
>> >>> >> >
>> >>> >> > To access to a hadoop cluster, I would propose to roughly follow
>> >>> Apache
>> >>> >> PIG.
>> >>> >> > On the other hand, I started considering Maven with [HADOOP-3302
>> >>> >> > Support Maven-based builds].
>> >>> >> >
>> >>> >> > What do you think?
>> >>> >> > --
>> >>> >> > Best regards,
>> >>> >> > Edward J. Yoon,
>> >>> >> > http://blog.udanax.org
>> >>> >> >
>> >>> >>
>> >>> >>
>> >>> >>
>> >>> >> --
>> >>> >> Best regards,
>> >>> >> Edward J. Yoon,
>> >>> >> http://blog.udanax.org
>> >>> >>
>> >>> >
>> >>> >
>> >>> >
>> >>> > --
>> >>> > Best. Regards,
>> >>> > Donguk Choi
>> >>> > http://www.alloes.com
>> >>> >
>> >>>
>> >>>
>> >>>
>> >>> --
>> >>> Best regards,
>> >>> Edward J. Yoon,
>> >>> http://blog.udanax.org
>> >>>
>> >>
>> >
>> >
>> >
>> > --
>> > Best regards,
>> > Edward J. Yoon,
>> > http://blog.udanax.org
>> >
>>
>>
>>
>> --
>> Best regards,
>> Edward J. Yoon,
>> http://blog.udanax.org
>>
>
>
>
> --
> Best. Regards,
> Donguk Choi
> http://blog.naver.com/anurocky
>
--
Best regards,
Edward J. Yoon,
http://blog.udanax.org
Re: Dependencies management and Distributions
Posted by Donguk Choi <al...@gmail.com>.
Oops, I was copy&paste the message of edward from IRC. Sorry for my rash
mailing.
On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
>
> Again, We hope that the Hama can be adaptable to other distributed
> computing environments (e.g. Hypertable, KFS, ..., etc). However,
> Let's not rush; do it step by step at the moment.
>
> Regards, Edward
>
> On Mon, Jun 9, 2008 at 5:14 PM, Edward J. Yoon <ed...@udanax.org> wrote:
> > Hi π,
> >
> > Hama uses some external libraries (e.g. commons-*, junit, log4j)
> > besides Hadoop and Hbase.
> >
> > Regards, Edward
> >
> > On Mon, Jun 9, 2008 at 5:00 PM, pi song <pi...@gmail.com> wrote:
> >> What are your dependencies so far? Only Hadoop and HBase?
> >>
> >> On Mon, Jun 9, 2008 at 3:41 PM, Edward J. Yoon <ed...@udanax.org>
> wrote:
> >>
> >>> > BTW, When HAMA's prototype source update to svn server?
> >>>
> >>> I think we need to get SVN accounts.
> >>>
> >>> On Mon, Jun 9, 2008 at 2:36 PM, Donguk Choi <al...@gmail.com>
> wrote:
> >>> > hi all.
> >>> >
> >>> > +1 about the following PIG. Then We can deal with the system sharing
> >>> problem
> >>> > flexibly with the existing Hadoop users. BTW, When HAMA's prototype
> >>> source
> >>> > update to svn server?
> >>> >
> >>> >
> >>> >
> >>> > On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
> >>> >>
> >>> >> IMO, I prefer the latter opinion to the former. But, Considering the
> >>> >> pre-existing hadoop cluster, The former is better than the latter.
> >>> >>
> >>> >> > To access to a hadoop cluster, I would propose to roughly follow
> >>> Apache
> >>> >> PIG.
> >>> >>
> >>> >> +1
> >>> >>
> >>> >> Edward
> >>> >>
> >>> >> On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <ed...@udanax.org>
> >>> wrote:
> >>> >> > Hi all,
> >>> >> >
> >>> >> > I think we need to focus on how we are going to distribute the
> Hama
> >>> >> > and to manage the dependencies.
> >>> >> >
> >>> >> > To access to a hadoop cluster, I would propose to roughly follow
> >>> Apache
> >>> >> PIG.
> >>> >> > On the other hand, I started considering Maven with [HADOOP-3302
> >>> >> > Support Maven-based builds].
> >>> >> >
> >>> >> > What do you think?
> >>> >> > --
> >>> >> > Best regards,
> >>> >> > Edward J. Yoon,
> >>> >> > http://blog.udanax.org
> >>> >> >
> >>> >>
> >>> >>
> >>> >>
> >>> >> --
> >>> >> Best regards,
> >>> >> Edward J. Yoon,
> >>> >> http://blog.udanax.org
> >>> >>
> >>> >
> >>> >
> >>> >
> >>> > --
> >>> > Best. Regards,
> >>> > Donguk Choi
> >>> > http://www.alloes.com
> >>> >
> >>>
> >>>
> >>>
> >>> --
> >>> Best regards,
> >>> Edward J. Yoon,
> >>> http://blog.udanax.org
> >>>
> >>
> >
> >
> >
> > --
> > Best regards,
> > Edward J. Yoon,
> > http://blog.udanax.org
> >
>
>
>
> --
> Best regards,
> Edward J. Yoon,
> http://blog.udanax.org
>
--
Best. Regards,
Donguk Choi
http://blog.naver.com/anurocky
Re: Dependencies management and Distributions
Posted by "Edward J. Yoon" <ed...@udanax.org>.
Again, We hope that the Hama can be adaptable to other distributed
computing environments (e.g. Hypertable, KFS, ..., etc). However,
Let's not rush; do it step by step at the moment.
Regards, Edward
On Mon, Jun 9, 2008 at 5:14 PM, Edward J. Yoon <ed...@udanax.org> wrote:
> Hi π,
>
> Hama uses some external libraries (e.g. commons-*, junit, log4j)
> besides Hadoop and Hbase.
>
> Regards, Edward
>
> On Mon, Jun 9, 2008 at 5:00 PM, pi song <pi...@gmail.com> wrote:
>> What are your dependencies so far? Only Hadoop and HBase?
>>
>> On Mon, Jun 9, 2008 at 3:41 PM, Edward J. Yoon <ed...@udanax.org> wrote:
>>
>>> > BTW, When HAMA's prototype source update to svn server?
>>>
>>> I think we need to get SVN accounts.
>>>
>>> On Mon, Jun 9, 2008 at 2:36 PM, Donguk Choi <al...@gmail.com> wrote:
>>> > hi all.
>>> >
>>> > +1 about the following PIG. Then We can deal with the system sharing
>>> problem
>>> > flexibly with the existing Hadoop users. BTW, When HAMA's prototype
>>> source
>>> > update to svn server?
>>> >
>>> >
>>> >
>>> > On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
>>> >>
>>> >> IMO, I prefer the latter opinion to the former. But, Considering the
>>> >> pre-existing hadoop cluster, The former is better than the latter.
>>> >>
>>> >> > To access to a hadoop cluster, I would propose to roughly follow
>>> Apache
>>> >> PIG.
>>> >>
>>> >> +1
>>> >>
>>> >> Edward
>>> >>
>>> >> On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <ed...@udanax.org>
>>> wrote:
>>> >> > Hi all,
>>> >> >
>>> >> > I think we need to focus on how we are going to distribute the Hama
>>> >> > and to manage the dependencies.
>>> >> >
>>> >> > To access to a hadoop cluster, I would propose to roughly follow
>>> Apache
>>> >> PIG.
>>> >> > On the other hand, I started considering Maven with [HADOOP-3302
>>> >> > Support Maven-based builds].
>>> >> >
>>> >> > What do you think?
>>> >> > --
>>> >> > Best regards,
>>> >> > Edward J. Yoon,
>>> >> > http://blog.udanax.org
>>> >> >
>>> >>
>>> >>
>>> >>
>>> >> --
>>> >> Best regards,
>>> >> Edward J. Yoon,
>>> >> http://blog.udanax.org
>>> >>
>>> >
>>> >
>>> >
>>> > --
>>> > Best. Regards,
>>> > Donguk Choi
>>> > http://www.alloes.com
>>> >
>>>
>>>
>>>
>>> --
>>> Best regards,
>>> Edward J. Yoon,
>>> http://blog.udanax.org
>>>
>>
>
>
>
> --
> Best regards,
> Edward J. Yoon,
> http://blog.udanax.org
>
--
Best regards,
Edward J. Yoon,
http://blog.udanax.org
Re: Dependencies management and Distributions
Posted by Donguk Choi <al...@gmail.com>.
May be, On second thoughts We hope Hama can be adaptable to other
distributed computing environments (e.g. Hypertable, KFS, ...). However,
Let's don't rush; do it step by step. Now, Hard work in the maximum hadoop
and hbase set focus.
On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
>
> Hi π,
>
> Hama uses some external libraries (e.g. commons-*, junit, log4j)
> besides Hadoop and Hbase.
>
> Regards, Edward
>
> On Mon, Jun 9, 2008 at 5:00 PM, pi song <pi...@gmail.com> wrote:
> > What are your dependencies so far? Only Hadoop and HBase?
> >
> > On Mon, Jun 9, 2008 at 3:41 PM, Edward J. Yoon <ed...@udanax.org>
> wrote:
> >
> >> > BTW, When HAMA's prototype source update to svn server?
> >>
> >> I think we need to get SVN accounts.
> >>
> >> On Mon, Jun 9, 2008 at 2:36 PM, Donguk Choi <al...@gmail.com> wrote:
> >> > hi all.
> >> >
> >> > +1 about the following PIG. Then We can deal with the system sharing
> >> problem
> >> > flexibly with the existing Hadoop users. BTW, When HAMA's prototype
> >> source
> >> > update to svn server?
> >> >
> >> >
> >> >
> >> > On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
> >> >>
> >> >> IMO, I prefer the latter opinion to the former. But, Considering the
> >> >> pre-existing hadoop cluster, The former is better than the latter.
> >> >>
> >> >> > To access to a hadoop cluster, I would propose to roughly follow
> >> Apache
> >> >> PIG.
> >> >>
> >> >> +1
> >> >>
> >> >> Edward
> >> >>
> >> >> On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <ed...@udanax.org>
> >> wrote:
> >> >> > Hi all,
> >> >> >
> >> >> > I think we need to focus on how we are going to distribute the Hama
> >> >> > and to manage the dependencies.
> >> >> >
> >> >> > To access to a hadoop cluster, I would propose to roughly follow
> >> Apache
> >> >> PIG.
> >> >> > On the other hand, I started considering Maven with [HADOOP-3302
> >> >> > Support Maven-based builds].
> >> >> >
> >> >> > What do you think?
> >> >> > --
> >> >> > Best regards,
> >> >> > Edward J. Yoon,
> >> >> > http://blog.udanax.org
> >> >> >
> >> >>
> >> >>
> >> >>
> >> >> --
> >> >> Best regards,
> >> >> Edward J. Yoon,
> >> >> http://blog.udanax.org
> >> >>
> >> >
> >> >
> >> >
> >> > --
> >> > Best. Regards,
> >> > Donguk Choi
> >> > http://www.alloes.com
> >> >
> >>
> >>
> >>
> >> --
> >> Best regards,
> >> Edward J. Yoon,
> >> http://blog.udanax.org
> >>
> >
>
>
>
> --
> Best regards,
> Edward J. Yoon,
> http://blog.udanax.org
>
--
Best. Regards,
Donguk Choi
http://blog.naver.com/anurocky
Re: Dependencies management and Distributions
Posted by "Edward J. Yoon" <ed...@udanax.org>.
Hi π,
Hama uses some external libraries (e.g. commons-*, junit, log4j)
besides Hadoop and Hbase.
Regards, Edward
On Mon, Jun 9, 2008 at 5:00 PM, pi song <pi...@gmail.com> wrote:
> What are your dependencies so far? Only Hadoop and HBase?
>
> On Mon, Jun 9, 2008 at 3:41 PM, Edward J. Yoon <ed...@udanax.org> wrote:
>
>> > BTW, When HAMA's prototype source update to svn server?
>>
>> I think we need to get SVN accounts.
>>
>> On Mon, Jun 9, 2008 at 2:36 PM, Donguk Choi <al...@gmail.com> wrote:
>> > hi all.
>> >
>> > +1 about the following PIG. Then We can deal with the system sharing
>> problem
>> > flexibly with the existing Hadoop users. BTW, When HAMA's prototype
>> source
>> > update to svn server?
>> >
>> >
>> >
>> > On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
>> >>
>> >> IMO, I prefer the latter opinion to the former. But, Considering the
>> >> pre-existing hadoop cluster, The former is better than the latter.
>> >>
>> >> > To access to a hadoop cluster, I would propose to roughly follow
>> Apache
>> >> PIG.
>> >>
>> >> +1
>> >>
>> >> Edward
>> >>
>> >> On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <ed...@udanax.org>
>> wrote:
>> >> > Hi all,
>> >> >
>> >> > I think we need to focus on how we are going to distribute the Hama
>> >> > and to manage the dependencies.
>> >> >
>> >> > To access to a hadoop cluster, I would propose to roughly follow
>> Apache
>> >> PIG.
>> >> > On the other hand, I started considering Maven with [HADOOP-3302
>> >> > Support Maven-based builds].
>> >> >
>> >> > What do you think?
>> >> > --
>> >> > Best regards,
>> >> > Edward J. Yoon,
>> >> > http://blog.udanax.org
>> >> >
>> >>
>> >>
>> >>
>> >> --
>> >> Best regards,
>> >> Edward J. Yoon,
>> >> http://blog.udanax.org
>> >>
>> >
>> >
>> >
>> > --
>> > Best. Regards,
>> > Donguk Choi
>> > http://www.alloes.com
>> >
>>
>>
>>
>> --
>> Best regards,
>> Edward J. Yoon,
>> http://blog.udanax.org
>>
>
--
Best regards,
Edward J. Yoon,
http://blog.udanax.org
Re: Dependencies management and Distributions
Posted by pi song <pi...@gmail.com>.
What are your dependencies so far? Only Hadoop and HBase?
On Mon, Jun 9, 2008 at 3:41 PM, Edward J. Yoon <ed...@udanax.org> wrote:
> > BTW, When HAMA's prototype source update to svn server?
>
> I think we need to get SVN accounts.
>
> On Mon, Jun 9, 2008 at 2:36 PM, Donguk Choi <al...@gmail.com> wrote:
> > hi all.
> >
> > +1 about the following PIG. Then We can deal with the system sharing
> problem
> > flexibly with the existing Hadoop users. BTW, When HAMA's prototype
> source
> > update to svn server?
> >
> >
> >
> > On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
> >>
> >> IMO, I prefer the latter opinion to the former. But, Considering the
> >> pre-existing hadoop cluster, The former is better than the latter.
> >>
> >> > To access to a hadoop cluster, I would propose to roughly follow
> Apache
> >> PIG.
> >>
> >> +1
> >>
> >> Edward
> >>
> >> On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <ed...@udanax.org>
> wrote:
> >> > Hi all,
> >> >
> >> > I think we need to focus on how we are going to distribute the Hama
> >> > and to manage the dependencies.
> >> >
> >> > To access to a hadoop cluster, I would propose to roughly follow
> Apache
> >> PIG.
> >> > On the other hand, I started considering Maven with [HADOOP-3302
> >> > Support Maven-based builds].
> >> >
> >> > What do you think?
> >> > --
> >> > Best regards,
> >> > Edward J. Yoon,
> >> > http://blog.udanax.org
> >> >
> >>
> >>
> >>
> >> --
> >> Best regards,
> >> Edward J. Yoon,
> >> http://blog.udanax.org
> >>
> >
> >
> >
> > --
> > Best. Regards,
> > Donguk Choi
> > http://www.alloes.com
> >
>
>
>
> --
> Best regards,
> Edward J. Yoon,
> http://blog.udanax.org
>
Re: Dependencies management and Distributions
Posted by "Edward J. Yoon" <ed...@udanax.org>.
> BTW, When HAMA's prototype source update to svn server?
I think we need to get SVN accounts.
On Mon, Jun 9, 2008 at 2:36 PM, Donguk Choi <al...@gmail.com> wrote:
> hi all.
>
> +1 about the following PIG. Then We can deal with the system sharing problem
> flexibly with the existing Hadoop users. BTW, When HAMA's prototype source
> update to svn server?
>
>
>
> On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
>>
>> IMO, I prefer the latter opinion to the former. But, Considering the
>> pre-existing hadoop cluster, The former is better than the latter.
>>
>> > To access to a hadoop cluster, I would propose to roughly follow Apache
>> PIG.
>>
>> +1
>>
>> Edward
>>
>> On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <ed...@udanax.org> wrote:
>> > Hi all,
>> >
>> > I think we need to focus on how we are going to distribute the Hama
>> > and to manage the dependencies.
>> >
>> > To access to a hadoop cluster, I would propose to roughly follow Apache
>> PIG.
>> > On the other hand, I started considering Maven with [HADOOP-3302
>> > Support Maven-based builds].
>> >
>> > What do you think?
>> > --
>> > Best regards,
>> > Edward J. Yoon,
>> > http://blog.udanax.org
>> >
>>
>>
>>
>> --
>> Best regards,
>> Edward J. Yoon,
>> http://blog.udanax.org
>>
>
>
>
> --
> Best. Regards,
> Donguk Choi
> http://www.alloes.com
>
--
Best regards,
Edward J. Yoon,
http://blog.udanax.org
Re: Dependencies management and Distributions
Posted by Donguk Choi <al...@gmail.com>.
hi all.
+1 about the following PIG. Then We can deal with the system sharing problem
flexibly with the existing Hadoop users. BTW, When HAMA's prototype source
update to svn server?
On 6/9/08, Edward J. Yoon <ed...@udanax.org> wrote:
>
> IMO, I prefer the latter opinion to the former. But, Considering the
> pre-existing hadoop cluster, The former is better than the latter.
>
> > To access to a hadoop cluster, I would propose to roughly follow Apache
> PIG.
>
> +1
>
> Edward
>
> On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <ed...@udanax.org> wrote:
> > Hi all,
> >
> > I think we need to focus on how we are going to distribute the Hama
> > and to manage the dependencies.
> >
> > To access to a hadoop cluster, I would propose to roughly follow Apache
> PIG.
> > On the other hand, I started considering Maven with [HADOOP-3302
> > Support Maven-based builds].
> >
> > What do you think?
> > --
> > Best regards,
> > Edward J. Yoon,
> > http://blog.udanax.org
> >
>
>
>
> --
> Best regards,
> Edward J. Yoon,
> http://blog.udanax.org
>
--
Best. Regards,
Donguk Choi
http://www.alloes.com
Re: Dependencies management and Distributions
Posted by "Edward J. Yoon" <ed...@udanax.org>.
IMO, I prefer the latter opinion to the former. But, Considering the
pre-existing hadoop cluster, The former is better than the latter.
> To access to a hadoop cluster, I would propose to roughly follow Apache PIG.
+1
Edward
On Mon, Jun 9, 2008 at 1:59 PM, Edward J. Yoon <ed...@udanax.org> wrote:
> Hi all,
>
> I think we need to focus on how we are going to distribute the Hama
> and to manage the dependencies.
>
> To access to a hadoop cluster, I would propose to roughly follow Apache PIG.
> On the other hand, I started considering Maven with [HADOOP-3302
> Support Maven-based builds].
>
> What do you think?
> --
> Best regards,
> Edward J. Yoon,
> http://blog.udanax.org
>
--
Best regards,
Edward J. Yoon,
http://blog.udanax.org