Hi Suresh,
It looks from the explain logging like everything is running
correctly. I'm guessing that the slowness is due to lack of
parallelism. By default we split the attribute table with one split
per attribute being indexed. How many of the records actually match
your query? If the number is large, then likely the entire
processing is being done on a single tablet server, so you're not
taking advantage of your full cluster. If this is indeed the
problem, you can manually add split points, or lower the split
threshold for the attribute table to spread the load some.
You can can check split points through the accumulo shell, or
through the accumulo monitor UI. You can also verify which tablet
servers are being hit by by your query through the monitor UI.
Hope that helps, if not circle back and let us know.
Thanks,
Emilio
On 04/26/2017 07:41 AM, Suresh
Prajapati wrote:
I am not able to understand why stats count is
taking such a long time though only 1 million record exists?.
What factors can affect this time?
What going miserably wrong?
_______________________________________________
geomesa-dev mailing list
geomesa-dev@xxxxxxxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.locationtech.org/mailman/listinfo/geomesa-dev
|