spark-reviews mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From tdas <...@git.apache.org>
Subject [GitHub] spark pull request: [SPARK-7988][STREAMING] Round-robin scheduling...
Date Fri, 19 Jun 2015 23:46:00 GMT
Github user tdas commented on a diff in the pull request:

    https://github.com/apache/spark/pull/6607#discussion_r32878356
  
    --- Diff: streaming/src/main/scala/org/apache/spark/streaming/scheduler/ReceiverTracker.scala
---
    @@ -308,12 +336,25 @@ class ReceiverTracker(ssc: StreamingContext, skipReceiverLaunch:
Boolean = false
             supervisor.start()
             supervisor.awaitTermination()
           }
    +
           // Run the dummy Spark job to ensure that all slaves have registered.
           // This avoids all the receivers to be scheduled on the same node.
           if (!ssc.sparkContext.isLocal) {
             ssc.sparkContext.makeRDD(1 to 50, 50).map(x => (x, 1)).reduceByKey(_ + _,
20).collect()
           }
     
    +      // Get the list of executors and schedule receivers
    +      val executors = getExecutors(ssc)
    +      val locations = scheduleReceivers(receivers, executors)
    +      val tempRDD =
    +        if (locations(0) != null) {
    --- End diff --
    
    First of all, in Scala, we try to not rely on null rather use Option and None. Here a
suggestion, which I think is a cleaner design with clean semantics. The makeRDD is designed
to take a `sequence of (item, locations)`. If for a item, the location is a empty (not null,
just empty seq), then that automatically means there is no preferred location. That's intuitive.

    
    So the `scheduleReceiver` can designed as follows.
    1. `scheduleReceiver`  always returns a Array[ArrayBuffer[String]] where any of the buffers
can be empty, but there are no nulls. 
    2. the logic in this location becomes 
    ```
    if (sparkContext is local) {
       // make RDD
    } else {
       // schedule receivers
       // make RDD with returned result
    }
    ```
    if there were no executors at that point of time, all the buffers will be empty, 
    which is perfectly okay to pass on to makeRDD. The code stays simple with only condition,
and no matter what the executors are empty or not, it just works.
    
    How does this sound?



---
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 infrastructure@apache.org or file a JIRA ticket
with INFRA.
---

---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org


Mime
View raw message