Thanks for this nice no-nonsense tutorial, Robin !...

2010-06-09T17:21:03.666+02:00

Thanks for this nice no-nonsense tutorial, Robin !!!!

* ERRATUM- php5-cli installation:
run ‘apt-get install php5-cli’ instead of ‘apt-get php5-cli’

* UPDATE- php5-cli installation:
i actually needed to do ‘apt-get upgrade’ and ‘apt-get update’ in order to get

all dependencies right.

* HINT- vmware file sharing with Ubuntu running on WinXP:
it is much easier to use WinSCP (SFTP session to the assigned IP address), than

to try to install VMWare Tools in command line and so on.

* UPDATE- before running $HADOOP_HOME/bin/hadoop:
i needed to set the JAVA_HOME (remove '#' and change location) in

$HADOOP_INSTALL/hadoop/conf/hadoop-env.sh . In my case to /usr/jre16 .

* UPDATE- MapReduce job with streams:
the file location contrib/hadoop-streaming.jar became contrib/streaming/hadoop-0.18.0-streaming.jar

I think hadoop's interface has changed a bit since...

2009-02-14T00:43:00.000+01:00

I think hadoop's interface has changed a bit since you posted this:

1) you need to include the -file option for mapper.php and reducer.php so that the streamer will package these files up and include them in the job
2) my image compressed the output by default. If this happens to you, try running the job with -jobconf mapred.output.compress=false

Comments on Lunchpauze: Writing A Hadoop MapReduce Program In PHP

Thanks for this nice no-nonsense tutorial, Robin !...

I think hadoop's interface has changed a bit since...