[
https://issues.apache.org/jira/browse/BEAM-4783?focusedWorklogId=142171&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-142171
]
ASF GitHub Bot logged work on BEAM-4783:
----------------------------------------
Author: ASF GitHub Bot
Created on: 07/Sep/18 12:50
Start Date: 07/Sep/18 12:50
Worklog Time Spent: 10m
Work Description: kyle-winkelman commented on a change in pull request
#6181: [BEAM-4783] Add bundleSize for splitting BoundedSources.
URL: https://github.com/apache/beam/pull/6181#discussion_r215947597
##########
File path:
runners/spark/src/main/java/org/apache/beam/runners/spark/translation/GroupCombineFunctions.java
##########
@@ -52,13 +50,11 @@
.map(WindowingHelpers.unwindowFunction())
.mapToPair(TranslationUtils.toPairFunction())
.mapToPair(CoderHelpers.toByteFunction(keyCoder, wvCoder));
- // use a default parallelism HashPartitioner.
- Partitioner partitioner = new
HashPartitioner(rdd.rdd().sparkContext().defaultParallelism());
// using mapPartitions allows to preserve the partitioner
// and avoid unnecessary shuffle downstream.
return pairRDD
- .groupByKey(partitioner)
+ .groupByKey()
Review comment:
To me it appears the double shuffling was solved by changing map and
mapToPair to mapPartitions and mapPartitionsToPair, respectively. I don't think
the explicit declaration of the HashPartitioner has any effect.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
Issue Time Tracking
-------------------
Worklog Id: (was: 142171)
Time Spent: 1.5h (was: 1h 20m)
> Spark SourceRDD Not Designed With Dynamic Allocation In Mind
> ------------------------------------------------------------
>
> Key: BEAM-4783
> URL: https://issues.apache.org/jira/browse/BEAM-4783
> Project: Beam
> Issue Type: Improvement
> Components: runner-spark
> Affects Versions: 2.5.0
> Reporter: Kyle Winkelman
> Assignee: Jean-Baptiste Onofré
> Priority: Major
> Labels: newbie
> Time Spent: 1.5h
> Remaining Estimate: 0h
>
> When the spark-runner is used along with the configuration
> spark.dynamicAllocation.enabled=true the SourceRDD does not detect this. It
> then falls back to the value calculated in this description:
> // when running on YARN/SparkDeploy it's the result of max(totalCores,
> 2).
> // when running on Mesos it's 8.
> // when running local it's the total number of cores (local = 1,
> local[N] = N,
> // local[*] = estimation of the machine's cores).
> // ** the configuration "spark.default.parallelism" takes precedence
> over all of the above **
> So in most cases this default is quite small. This is an issue when using a
> very large input file as it will only get split in half.
> I believe that when Dynamic Allocation is enable the SourceRDD should use the
> DEFAULT_BUNDLE_SIZE and possibly expose a SparkPipelineOptions that allows
> you to change this DEFAULT_BUNDLE_SIZE.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
