[
https://issues.apache.org/jira/browse/SPARK-4630?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14283161#comment-14283161
]
Reynold Xin commented on SPARK-4630:
------------------------------------
The problem is that as the stage if producing its data, it needs to decide
where the data goes to (running the partitioner), and it becomes too late once
that is decided. If we delay it, it would help with collecting the information
and re-optimizing it, but it would be very expensive to run through the data
twice once to generate, and another to partition. Maybe we can come up with
smarter ways to deal with this.
If you look at the Flume Java implementation within Google, essentially they do
trial runs on sampled datasets, and then use that to infer the degree of
parallelism. In schema rdd, because we know the operations being done (e.g.
aggregate on dimensions with cardinality estimations), we can make much better
estimations.
> Dynamically determine optimal number of partitions
> --------------------------------------------------
>
> Key: SPARK-4630
> URL: https://issues.apache.org/jira/browse/SPARK-4630
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
> Reporter: Kostas Sakellis
> Assignee: Kostas Sakellis
>
> Partition sizes play a big part in how fast stages execute during a Spark
> job. There is a direct relationship between the size of partitions to the
> number of tasks - larger partitions, fewer tasks. For better performance,
> Spark has a sweet spot for how large partitions should be that get executed
> by a task. If partitions are too small, then the user pays a disproportionate
> cost in scheduling overhead. If the partitions are too large, then task
> execution slows down due to gc pressure and spilling to disk.
> To increase performance of jobs, users often hand optimize the number(size)
> of partitions that the next stage gets. Factors that come into play are:
> Incoming partition sizes from previous stage
> number of available executors
> available memory per executor (taking into account
> spark.shuffle.memoryFraction)
> Spark has access to this data and so should be able to automatically do the
> partition sizing for the user. This feature can be turned off/on with a
> configuration option.
> To make this happen, we propose modifying the DAGScheduler to take into
> account partition sizes upon stage completion. Before scheduling the next
> stage, the scheduler can examine the sizes of the partitions and determine
> the appropriate number tasks to create. Since this change requires
> non-trivial modifications to the DAGScheduler, a detailed design doc will be
> attached before proceeding with the work.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
