Adds a new property "druid.broker.http.maxQueuedBytes" and a new context
parameter "maxQueuedBytes". Both represent a maximum number of bytes queued
per query before exerting backpressure on the channel to the data server.

I tested this by running on a modest cluster (6 historicals) and doing an 
unlimited,
unfiltered Scan query via SQL (`select * from tbl`). The query ran for about 
half an
hour before it timed out, and managed to fetch almost 20GB of results without
OOMing the broker. Without this patch, the same query quickly OOMed the broker.

Fixes #4933.

[ Full content available at: 
https://github.com/apache/incubator-druid/pull/6313 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to