Hi Amit, Gridmix generates its own data. You can get a trace by
1. Running a set of jobs on your own cluster 2. Copy the jobhistory files (usually in /mapred/history/done in HDFS) to local disk 3. Run Rumen on the history files (which is also easy http://hadoop.apache.org/mapreduce/docs/current/rumen.html ) OR if you have a compelling research requirement, from Yahoo's WebScope http://webscope.sandbox.yahoo.com/catalog.php?datatype=s I don't think you need users-list if you run without security. Hope this helps Ravi On Wed, Jun 13, 2012 at 8:28 AM, sangroya <sangroyaa...@gmail.com> wrote: > Hi, > > I want to use Gridmix3 benchmark with hadoop version 1.0.0. > > I am following this link. > > http://hadoop.apache.org/mapreduce/docs/current/gridmix.html > > Here, it is mentioned that, to run Gridmix3, I need > > data, users-list and traces > > Can anyone suggest me where do I get them. > > Is there any detailed documentation and tutorials on Gridmix3 benchmark. > Can > somebody point me to these links. > > 1) How to use Gridmix3 benchmark? > > 2) What are the inputs and outputs for Gridmix3 benchmark? e.g. What do we > mean by traces, data, and userlist. > > 3) How can we analyze the outputs of Gridmix3 benchmark? > > > > Thanks in advance, > > Amit > > > ----- > Sangroya > -- > View this message in context: > http://lucene.472066.n3.nabble.com/How-to-use-GridMix3-tp3989438.html > Sent from the Hadoop lucene-users mailing list archive at Nabble.com. >
