Github user paul-rogers commented on a diff in the pull request:
https://github.com/apache/drill/pull/921#discussion_r150982788
--- Diff: exec/java-exec/src/main/resources/drill-module.conf ---
@@ -368,6 +368,13 @@ drill.exec: {
// planning and managing queries. Primarily for testing.
cpus_per_node: 0,
}
+ # Grace period is the amount of time where the drillbit accepts work
after
+ # the shutdown request is triggered. The primary use of grace period is
to
+ # avoid the race conditions caused by zookeeper delay in updating the
state
+ # information of the drillbit that is shutting down. So, it is advisable
+ # to have a grace period that is atleast twice the amount of zookeeper
+ # refresh time.
+ grace_period : 10000
--- End diff --
Units? Just add a note to the end of your comment to say the units are ms.
In the future, I've found it to be handy to put the units in the name:
`grace_period_ms` for example.
---
