I'm puzzled about how the hadoop job tracker finds map/reduce classes.
I noticed a change when I went from running hadoop stand-alone to
running with a job/task tracker set of daemons active. In the later
case my hadoop request is not running. Digging further into the errors
it appears that the job tracker cannot find my classes. Specifically,
I get the error:
java.lang.ClassNotFoundException: MaxTemperatureMapper
In stand alone mode, my classes are found. With the job daemon
running, my classes are not found.
What am I missing?
I've tried testing with:
HADOOP_CLASSPATH
set to either
./
or the full path to my sample programs. Doesn't seem to matter either way.
Thanks,
David
