hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ravi <ravikapoor...@gmail.com>
Subject Re: Can I use hadoop inside hbase?
Date Sun, 29 Nov 2009 23:57:01 GMT


I am using the windows distribution, it must have come with hadoop built 
in as I did not install hadoop separately. The distribution however does 
not have any hadoop commands or shell scripts.

So looks like this is a hbase windows distribution issue?


Andrew Purtell wrote:
> To run HBase you must have installed Hadoop underneath. The Hadoop top level command
line script 'bin/hadoop' is available in your Hadoop installation. Just put the Hadoop bin/
dir on the path along with the HBase bin/ dir. 
> 
> You can also run some Hadoop applications using the HBase command script, but that is
missing the shortcuts, e.g. instead of
> 
>     hadoop fs -copyFromLocal ...
> 
>   it is 
> 
>     hbase org.apache.hadoop.fs.FsShell -copyFromLocal ...
> 
> This will work for DFS but for other filesystems which need supporting jars in lib/,
your mileage will vary. Just put the Hadoop bin/ dir on the path along with the HBase bin/
dir and use the appropriate one for the circumstances. 
> 
>   - Andy
> 
> 
> 
> ________________________________
> From: Ravi <ravikapoor101@gmail.com>
> To: hbase-user@hadoop.apache.org
> Sent: Sun, November 29, 2009 8:28:58 AM
> Subject: Re: Can I use hadoop inside hbase?
> 
> Thanks for the quick clarification Jeff.
> 
> In hbase distribution, I am unable to find hadoop command to load files 
> into DFS. How do I add hadoop commands to hbase distribution?
> 
> 
> Jeff Zhang wrote:
>> Of course, you can. I run my pig script on the cluster and run hbase on the
>> same cluster, it works fine.
>>
>> Jeff Zhang
>>
>>
>> On Sat, Nov 28, 2009 at 9:45 PM, Ravi <ravikapoor101@gmail.com> wrote:
>>
>>> If I setup a hbase cluster, say with 10 nodes, can I also use the same
>>> cluster for my distributed program using the hadoop running underneath
>>> hbase?
>>>
>>> Essentially the thought is to use underlying hadoop for two purposes, data
>>> as well as logic.
>>>
>>> Any thoughts?
>>>
>>>
> 
> 
>       


Mime
View raw message