Github user liyezhang556520 commented on a diff in the pull request:
https://github.com/apache/spark/pull/7753#discussion_r40339731
--- Diff:
core/src/main/scala/org/apache/spark/scheduler/EventLoggingListener.scala ---
@@ -152,8 +159,19 @@ private[spark] class EventLoggingListener(
}
}
+ // We log the event both when stage submitted and stage completed, and
after each logEvent call,
+ // replace the modifiedMetrics with the latestMetrics. In case the
stages submit and complete
+ // time might be interleaved. So as to make the result the same with the
running time.
+ private def logMetricsUpdateEvent() : Unit = {
+ modifiedMetrics.map(metrics => logEvent(metrics._2))
+ latestMetrics.map(metrics => modifiedMetrics.update(metrics._1,
metrics._2))
+ }
--- End diff --
> I don't understand the last two sentences of the comment -- can you
expand on that?
I'll update the code according to the design doc. I think the code is not
that correct. Please refer it in [design
doc](https://issues.apache.org/jira/secure/attachment/12762171/Tracking%20Spark%20Memory%20Usage%20-%20Phase%201.pdf)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
