Hao Hao has posted comments on this change. ( 
http://gerrit.cloudera.org:8080/8552 )

Change subject: KUDU-1454 [part 1]: update propagated timestamp on the driver
......................................................................


Patch Set 2:

(2 comments)

http://gerrit.cloudera.org:8080/#/c/8552/1/java/kudu-spark/src/main/scala/org/apache/kudu/spark/kudu/KuduContext.scala
File 
java/kudu-spark/src/main/scala/org/apache/kudu/spark/kudu/KuduContext.scala:

http://gerrit.cloudera.org:8080/#/c/8552/1/java/kudu-spark/src/main/scala/org/apache/kudu/spark/kudu/KuduContext.scala@86
PS1, Line 86:       
syncClient.updateLastPropagatedTimestamp(timestampAccumulator.value)
> it seems slightly odd that merge() has a side-effect here. Does this end up
My understanding is that merge() is only called on the driver according to 
reading of the source code: 
https://github.com/apache/spark/blob/master/core/src/main/scala/org/apache/spark/scheduler/DAGScheduler.scala#L1122.

I agree it looks odd to update TS here. But on a second thought, I decide to 
put it here since it is central place to do the TS update, other that go into 
different scan/write paths to do the update. Does that make sense?


http://gerrit.cloudera.org:8080/#/c/8552/1/java/kudu-spark/src/main/scala/org/apache/kudu/spark/kudu/KuduContext.scala@294
PS1, Line 294:     
timestampAccumulator.add(syncClient.getLastPropagatedTimestamp)
> should this be finally{} so that, even if there is an error, you can see al
I do not quite follow your second comment. My understanding is for each RDD, 
the kuduClient would be copied to the executors from the driver. And with the 
last-propagated timestamp set on driver, the next read/write on executors 
should also get this information?



--
To view, visit http://gerrit.cloudera.org:8080/8552
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings

Gerrit-Project: kudu
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: Id0a078ae8ebaa6a859be75c822879291192c5842
Gerrit-Change-Number: 8552
Gerrit-PatchSet: 2
Gerrit-Owner: Hao Hao <[email protected]>
Gerrit-Reviewer: Dan Burkert <[email protected]>
Gerrit-Reviewer: Hao Hao <[email protected]>
Gerrit-Reviewer: Kudu Jenkins
Gerrit-Reviewer: Todd Lipcon <[email protected]>
Gerrit-Comment-Date: Thu, 16 Nov 2017 00:13:44 +0000
Gerrit-HasComments: Yes

Reply via email to