- last edited on
I have a cluster of 3 datanodes and I'm launching a Hive query on MR.
The query is a join between two tambles which spawns 3 mappers.
- the 3 mappers are all running on the same node. Is this normal?
- the specified node is only using about 40$ of his CPU and the job takes ages to complete. Why? Considering the two tables I use are 3 blocks and 1 block respectively.
Any ideas? Thanks