Github user srowen commented on a diff in the pull request:
https://github.com/apache/spark/pull/5074#discussion_r27163365
--- Diff: docs/programming-guide.md ---
@@ -1086,6 +1086,66 @@ for details.
</tr>
</table>
+### Shuffle operations
+
+Certain operations within Spark trigger an event known as the shuffle. The
shuffle is Spark's
+mechanism for re-distributing data so that is grouped differently across
partitions. This typically
+involves copying data across executors and machines, making the shuffle a
complex and
+costly operation.
+
+#### Background
+
+To understand what happens during the shuffle we can consider the example
of the
+[`reduceByKey`](#ReduceByLink) operation. The `reduceByKey` operation
generates a new RDD where all
+values for a single key are combined into a tuple - the key and the result
of executing a reduce
+function against all values associated with that key. The challenge is
that not all values for a
+single key necessarily reside on the same partition, or even the same
machine, but they must be
+co-located to compute the result.
+
+In Spark, data is generally not distributed across partitions to be in the
necessary place for a
+specific operation. During computations, a single task will operate on a
single partition - thus, to
+organize all the data for a single `reduceByKey` reduce task to execute,
Spark needs to perform an
+all-to-all operation. It must read from all partitions to find all the
values for all keys, and then
+organize those such that all values for any key lie within the same
partition - this is called the
+**shuffle**.
+
+Although the set of elements in each partition of newly shuffled data will
be deterministic, the
+ordering of these elements is not. If one desires predictably ordered data
following shuffle
+operations, [`mapPartitions`](#MapPartLink) can be used to sort each
partition or `sortBy` can be
+used to perform a global sort. A similar operation,
+[`repartitionAndSortWithinPartitions`](#Repartition2Link`) coupled with
`mapPartitions`,
+may be used to enact a Hadoop style shuffle.
+
+Operations which can cause a shuffle include **repartition** operations
like
+[`repartition`](#RepartitionLink), and [`coalesce`](#CoalesceLink),
**'byKey** operations
+(except for counting) like [`groupByKey`](#GroupByLink) and
[`reduceByKey`](#ReduceByLink), and
+**join** operations like [`cogroup`](#CogroupLink) and [`join`](#JoinLink).
+
+#### Performance Impact
+**Shuffle** is an expensive operation since it involves disk I/O, data
serialization, and
+network I/O. To organize data for the shuffle, Spark generates sets of
tasks - *map* tasks to
+organize the data, and a set of *reduce* tasks to aggregate it. This
nomenclature comes from
+MapReduce and does not directly relate to Spark's `map` and `reduce`
operations.
+
+Internally, results from individual map tasks are kept in memory until
they can't fit. Then, these
+are sorted based on the target partition and written to a single file. On
the reduce side, tasks
+read the relevant sorted blocks.
+
+Certain shuffle operations can consume significant amounts of heap memory
since they employ
+in-memory data structures to organize records before or after transferring
them. Specifically,
+`reduceByKey` and `aggregateByKey` create these structures on the map-side
and `'byKey` operations
--- End diff --
Last nit: I think map-side and reduce-side don't need a hyphen
---
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]
