rmdmattingly opened a new pull request, #5654:
URL: https://github.com/apache/hbase/pull/5654

   See https://issues.apache.org/jira/browse/HBASE-27687
   
   We've found that result size is not a great indication of a request's 
workload for the server. For example, as is, one may set a user to have a 
100mb/machine throttle and then find that the given user is able to generate 
far more than 100mb of IO against any given machine if their result sizes are 
significantly smaller than the IO required to facilitate the request.
   
   This PR introduces `hbase.quota.use.block.bytes.scanned`, a configurable 
boolean (defaulted to false) which, if true, will cause quotas to consume read 
availability via block bytes scanned rather than result size. Extending the 
aforementioned example, this should help to ensure that a user throttled to 
100mb of IO/machine will not be able to exceed 100mb of IO/machine.
   
   @bbeaudreault @hgromer @eab148 @bozzkar


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to