Learning HBase

[Bing Li <lb...@gmail.com>]

Dear all,

I am a new user of HBase. I just finished reading the book, HBase the
Definitive Guide.

I am implementing a new page sharing system. HBase is a good choice to
process big data for me. The issue is that I still need much effort to
program my system and I need to focus on the logic of my system. At this
time, is it fine that I just take care of the Java client APIs for the
pseudo distributed mode? I plan to complete my system first and then
consider Map/Reduce integration and some other management techniques. Any
possible problems to do so?

Thanks so much!

Best regards,
Bing

Re: Learning HBase

[Ulrich Staudinger <us...@gmail.com>]

feedback appreciated
thanks


--
comnect on xing or linkedin

On 18.02.2012, at 19:25, Stack <st...@duboce.net> wrote:

> On Sat, Feb 18, 2012 at 8:24 AM, Ulrich Staudinger
> <us...@activequant.com> wrote:
>> It uses 0.90.4 and spawns a zooserver, regionserver and master in their own
>> thread, all running on localhost and with the data directory set the
>> application's root folder.
>>
>
> I like the ascii it outputs on startup Ulrich (and this bit '12/02/02
> 14:15:19 INFO server.AQ2Server: Starting mighty HBase ....').
>
> St.Ack

Re: Learning HBase

[Stack <st...@duboce.net>]

On Sat, Feb 18, 2012 at 8:24 AM, Ulrich Staudinger
<us...@activequant.com> wrote:
> It uses 0.90.4 and spawns a zooserver, regionserver and master in their own
> thread, all running on localhost and with the data directory set the
> application's root folder.
>

I like the ascii it outputs on startup Ulrich (and this bit '12/02/02
14:15:19 INFO server.AQ2Server: Starting mighty HBase ....').

St.Ack

Re: Learning HBase

[Ulrich Staudinger <us...@activequant.com>]

Hey Bing,

you might want to give my AQ2o master server a try, it's really great for
development, I use it everyday and a bunch of friends of mine, too.

It's download, double-click (or sh runserver.sh) and use. URL
http://aq2o.activequant.org/uberjar.html

It uses 0.90.4 and spawns a zooserver, regionserver and master in their own
thread, all running on localhost and with the data directory set the
application's root folder.

Don't use it in production, it's really there for development.

That server also contains other stuff in the same VM: namely a HSQLDB and a
SOAP server, which are not probably not relevant for you, but nice anyway
...

That should help you to get started a bit ... as said, download, start,
develop against it.



Cheers,
Ulrich





On Sat, Feb 18, 2012 at 3:36 PM, T Vinod Gupta <tv...@readypulse.com>wrote:

> yeah that should be fine. out of experience, when we started - we used
> standalone hbase! and then slowly migrated to distributed mode with a much
> more stable hbase version. and all this was done while we had beta
> customers along the way. If infrastructure costs are something you want to
> optimize on while you are developing the platform, then this is a good
> approach.
>
> thanks
>
> On Sat, Feb 18, 2012 at 6:29 AM, Bing Li <lb...@gmail.com> wrote:
>
> > Dear all,
> >
> > I am a new user of HBase. I just finished reading the book, HBase the
> > Definitive Guide.
> >
> > I am implementing a new page sharing system. HBase is a good choice to
> > process big data for me. The issue is that I still need much effort to
> > program my system and I need to focus on the logic of my system. At this
> > time, is it fine that I just take care of the Java client APIs for the
> > pseudo distributed mode? I plan to complete my system first and then
> > consider Map/Reduce integration and some other management techniques. Any
> > possible problems to do so?
> >
> > Thanks so much!
> >
> > Best regards,
> > Bing
> >
>



-- 
Ulrich Staudinger

<http://goog_958005736>http://www.activequant.com
Connect online: https://www.xing.com/profile/Ulrich_Staudinger

Re: Learning HBase

[T Vinod Gupta <tv...@readypulse.com>]

yeah that should be fine. out of experience, when we started - we used
standalone hbase! and then slowly migrated to distributed mode with a much
more stable hbase version. and all this was done while we had beta
customers along the way. If infrastructure costs are something you want to
optimize on while you are developing the platform, then this is a good
approach.

thanks

On Sat, Feb 18, 2012 at 6:29 AM, Bing Li <lb...@gmail.com> wrote:

> Dear all,
>
> I am a new user of HBase. I just finished reading the book, HBase the
> Definitive Guide.
>
> I am implementing a new page sharing system. HBase is a good choice to
> process big data for me. The issue is that I still need much effort to
> program my system and I need to focus on the logic of my system. At this
> time, is it fine that I just take care of the Java client APIs for the
> pseudo distributed mode? I plan to complete my system first and then
> consider Map/Reduce integration and some other management techniques. Any
> possible problems to do so?
>
> Thanks so much!
>
> Best regards,
> Bing
>