Looking for some guidance here. If this question has already been answered please point me to response.
Working on a job that is not performing up to par. Noticed several spills in the map phase and the merge seems to taking a while. I see that io.sort.mb is the space allocated to the buffer, record and the data buffers. Given that my jvm for map tasks 700m and the space left after taking out the space used for buffers is 400m. What is stored in this 400m? Thanks, Ranjith