I'm working on the 1.2 branch, and so far I see:

0) login as the admin/admin user. 
1) Job Tracker runs as "mapred"
2) Task Tracker runs as "mapred"
3) The map/reduce Service checks runs a map/reduce job as "ambari_qa"  UID. 

Both mapred & ambari_qa belong to hadoop group.

Could someone verify that this is anticipated?  IE Job + Task trackers 
processes run as one UID and the map reduce jobs as another.




-bc

Reply via email to