Hi Yogesh,

Just wanted to correct one point of yours:

On Thu, Dec 6, 2012 at 10:25 PM, yogesh dhari <[EMAIL PROTECTED]> wrote:
> Hadoop have single point of Failure but Cassandra doesn't..

In case you aren't aware yet, Hadoop (HDFS) has no single point of
failure anymore. The HDFS project implemented HA quite a while ago and
it works pretty well (and several of our customers run it too).

See http://hadoop.apache.org/docs/current/hadoop-yarn/hadoop-yarn-site/HDFSHighAvailability.html
for more. The release of 2.0.3, coming very soon, will also have a
non-NFS dependent form of HA support (trunk already has it).

--
Harsh J

Thanks a lot guys  :-)

Regards
Yogesh Kumar

From: [EMAIL PROTECTED]
Date: Thu, 6 Dec 2012 11:31:16 -0800
Subject: Re: Hadoop V/S Cassandra
To: [EMAIL PROTECTED]

Hi Yogesh,
As others have said Hadoop vs Cassandra is not a fair comparison. Although, HBase vs Cassandra is a fair comparison. You can have a look at this comparison: http://bigdatanoob.blogspot.com/2012/11/hbase-vs-cassandra.html

HTH,
Anil Gupta

On Thu, Dec 6, 2012 at 11:27 AM, Colin McCabe <[EMAIL PROTECTED]> wrote:
On Thu, Dec 6, 2012 at 8:55 AM, yogesh dhari <[EMAIL PROTECTED]> wrote:

> Hi all,

>

> Hadoop have different file system(HDFS) and Cassandra have different file

> system(CFS).

>

> As Hadoop have great Eco-System (Hive{Dataware House}, Hbase{Data Base} n

> etc..) and Cassandra(Database) it self providing its own file system

>

> Although  we can run Hadoop's Ecosystem on Cassandra (If I am right ? {like

> pig and hive})

>

> Hadoop have single point of Failure but Cassandra doesn't..

>

> Please put some light on both of these where Hadoop overtakes Cassandra an

> where Cassandra.

>

> And in which use cases we go for any them.

As Harsh commented, HDFS doesn't have a single point of failure any more.

Hadoop and Cassandara are not really competitors.  You can run Hadoop

on top of Cassandra if you want.

It's possible to compare Cassandra and HBase, and some people have

done so.  In general, I've been told that Cassandra is a better match

for situations where you write more than you read, and you don't care

about data consistency.  But that might be a better question for the

HBase and/or Cassandra mailing lists.

cheers,

Colin

--
Thanks & Regards,
Anil Gupta