Hi Jayesh,
We presently don't have any notion of priorities in Accumulo RPCs. They
all fall into the same execution "bucket" inside of the TabletServer.
- Josh
Jayesh Patel wrote:
If I’m running a mapreduce job that is very expansive may be because it
scans through every row and every column-qualifier, it is likely going
to keep at least some of the tablet servers very busy and affect other
scans going on.
Is there a way to assign such a job a lower priority so it doesn’t
affect other scans that might be going on, may it be other mapreduce job
or accumulo scans.
How about some accumulo scans that we know are going to be stragglers
that we might want to run at lower priority?
Thank you,
Jayesh
