Pig 0.7 is compatible with hadoop 0.20.x, so I am not sure why you are
seeing this error.
You can try creating a pig jar without hadoop and specify the hadoop jar
used for your cluster in classpath -
svn co http://svn.apache.org/repos/asf/hadoop/pig/branches/branch-0.7
ls pig-withouthadoop.jar #use this jar
Please let us know if this fixed your problem.
On 8/16/10 1:52 AM, "Rico Bergmann" <[EMAIL PROTECTED]>
> I try to get the tutorial of pig0.7.0 running in mapreduce mode. But I
> always get IOExceptions. Looking into the HDFS-Logfiles I found a
> message "Incorrect header or version mismatch from ... got version2
> expected version 3". I don't know what to do. Is it a bug in the tutorial?
> For completness: I'm running Hadoop Version 0.20.2 in a clustered setup.
> Thanks, Rico.