on a hadoop cluster, yes. you will need to avoid setting the path variable and maintain the path yourself based on which version of pig you want to use.
on a hadoop cluster, yes. you will need to avoid setting the path variable and maintain the path yourself based on which version of pig you want to use.
Unless you use the local jar's for the pig there won't be any problem. If you are trying to use from hdp i.e. oozie(/hdp/apps/2.3.6.0-3796/pig/pig.tar.gz) then it will be problem.