Hello,
I’m facing an issue whith sqoop when importing data from a Teradata database.
I’m using HW addon and HDP 2.2.
When I try to import data, all the map tasks generated by sqoop runs on the same node, whith obviously performance problems.
If I retry the sqoop import, map tasks can be on one another node, but just one node.
If I run a sqoop import from oracle (or any other map-reduce job), there is no problem map tasks are well distributed on all nodes.
My query : is there any specific location for map task with the hortonworks / teradata sqoop addon ? And if yes, how can I distribute the task on all the cluster ?
Thanks in advance,
Jeremie