i see, i'm using hadoop to stores binary data (files) to have it high-avail & scalability & want also put all other data also there to avoid maintaince/admin effort (e.g database). as i also want to provide eg. preview for the data (if possible , like small text file) instead of downloading it to client, i think it's a good idea to use requence-file. maybe i've been thinking wrong.
is there any another way of doing it on hadoop ? thanx kim On Wed, Mar 21, 2012 at 12:07 AM, Ioan Eugen Stan <stan.ieu...@gmail.com>wrote: > 2012/3/20 kim young ill <khi...@googlemail.com>: > > > > > > On Tue, Mar 20, 2012 at 5:35 PM, Ioan Eugen Stan <stan.ieu...@gmail.com> > > wrote: > >> > >> Pe 20.03.2012 18:18, Manish Bhoge a scris: > >>> > >>> Kim, > >>> Interesting to know how to handle segment in sequence. Just want to > >>> understand the application of deleting the segment rather than the file > >>> itself in a sequence. > >>> Sorry for not providing solution here. But I would like to understand > the > >>> application at first place. > >>> Thank you, > >>> Manish. > >>> Sent from my BlackBerry, pls excuse typo > > > > > > > > Hi, > > > > just imagine one wants to save conversations, log-entries, > > blog-entries,emails... to files, one would also want to delete/change one > > entry in the middle. for some reason i want to avoid a database. each > > text-segment to a seperate file isnt a good idea, right ? > > > > thanx > > kim > > Depends on your use case. How much date are you thinking about? Hadoop > is justified only if you wish to do large scale processing, and a > single machine can't do the job. And it has some limitations regarding > what you can do. It's designed with write once, read many times ideas > in mind. > > -- > Ioan Eugen Stan > http://ieugen.blogspot.com/ >
