Tim Armstrong created IMPALA-7475:
-------------------------------------
Summary: Explicitly document perf consequences of setting mem_limit
Key: IMPALA-7475
URL: https://issues.apache.org/jira/browse/IMPALA-7475
Project: IMPALA
Issue Type: Improvement
Components: Docs
Reporter: Tim Armstrong
Assignee: Alex Rodoni
There are a few significant bits of information that the docs don't really
explicitly mention.
* In the mem_limit docs, we don't mention that Impala will employ various
strategies to keep the query's memory consumption under the mem_limit that
trade off execution speed for reduced memory consumption.
* Spill-to-disk is one such mechanism but we do other things like reducing
number of scanner threads and passing through rows from preaggregations.
[~arodoni_cloudera] is there a good place in the docs to add something like
this? I could draft something but not sure exactly where it would fit.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
