hadoop-common-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Yi-Kai Tsai <yi...@yahoo-inc.com>
Subject Re: save mp3 file Using Hadoop
Date Fri, 10 Oct 2008 09:56:01 GMT

Hadoop implement MapReduce by using HDFS,
yes, you could store things on HDFS
but I don't think you get any benefit on this use-case if you are not 
planing to "process" those data on top of hadoop/mapreduce.

> i am a beginner,i want to save a lot of Mp3 files into Hadoop and
> using Tomcat to service this files.
> 1.i want to using org.apache.hadoop.io.MapFile, key is mp3  name's
> MD5,value is Mp3 stream data. mp3 's filesize is nearly 3M. i split
> 20G per one MapFile.There are 30~50 MapFile.
> 2. i put these MapFile into  x linux server.
> 3. my user send Mp3' name request to Tomcat ,Tomcat using Mp3 name 's
> MD5 to find Mp3 file in one linux server .
> question:
> 1> is it right way to use hadoop?
> 2> if 20 user access one linux server in one time,maybe 20 Mp3 files
> load in memory,20* 3M memory is being eat? is it right? how to solve
> this problem ?

Yi-Kai Tsai (cuma) <yikai@yahoo-inc.com>

View raw message