diff options
Diffstat (limited to 'docs')
| -rw-r--r-- | docs/README.md | 8 | ||||
| -rwxr-xr-x | docs/_layouts/global.html | 8 | ||||
| -rw-r--r-- | docs/_plugins/copy_api_dirs.rb | 17 | ||||
| -rw-r--r-- | docs/api.md | 1 | ||||
| -rw-r--r-- | docs/configuration.md | 26 | ||||
| -rw-r--r-- | docs/ec2-scripts.md | 4 | ||||
| -rw-r--r-- | docs/index.md | 18 | ||||
| -rw-r--r-- | docs/java-programming-guide.md | 3 | ||||
| -rw-r--r-- | docs/python-programming-guide.md | 117 | ||||
| -rw-r--r-- | docs/quick-start.md | 50 | ||||
| -rw-r--r-- | docs/scala-programming-guide.md | 3 | ||||
| -rw-r--r-- | docs/spark-standalone.md | 43 | ||||
| -rw-r--r-- | docs/streaming-programming-guide.md | 246 |
13 files changed, 448 insertions, 96 deletions
diff --git a/docs/README.md b/docs/README.md index 092153070e..887f407f18 100644 --- a/docs/README.md +++ b/docs/README.md @@ -25,10 +25,12 @@ To mark a block of code in your markdown to be syntax highlighted by jekyll duri // supported languages too. {% endhighlight %} -## Scaladoc +## API Docs (Scaladoc and Epydoc) You can build just the Spark scaladoc by running `sbt/sbt doc` from the SPARK_PROJECT_ROOT directory. -When you run `jekyll` in the docs directory, it will also copy over the scala doc for the various Spark subprojects into the docs directory (and then also into the _site directory). We use a jekyll plugin to run `sbt/sbt doc` before building the site so if you haven't run it (recently) it may take some time as it generates all of the scaladoc. +Similarly, you can build just the PySpark epydoc by running `epydoc --config epydoc.conf` from the SPARK_PROJECT_ROOT/pyspark directory. -NOTE: To skip the step of building and copying over the scaladoc when you build the docs, run `SKIP_SCALADOC=1 jekyll`. +When you run `jekyll` in the docs directory, it will also copy over the scaladoc for the various Spark subprojects into the docs directory (and then also into the _site directory). We use a jekyll plugin to run `sbt/sbt doc` before building the site so if you haven't run it (recently) it may take some time as it generates all of the scaladoc. The jekyll plugin also generates the PySpark docs using [epydoc](http://epydoc.sourceforge.net/). + +NOTE: To skip the step of building and copying over the scaladoc when you build the docs, run `SKIP_SCALADOC=1 jekyll`. Similarly, `SKIP_EPYDOC=1 jekyll` will skip PySpark API doc generation. diff --git a/docs/_layouts/global.html b/docs/_layouts/global.html index a8be52f23e..94baa634aa 100755 --- a/docs/_layouts/global.html +++ b/docs/_layouts/global.html @@ -47,6 +47,7 @@ <li><a href="quick-start.html">Quick Start</a></li> <li><a href="scala-programming-guide.html">Scala</a></li> <li><a href="java-programming-guide.html">Java</a></li> + <li><a href="python-programming-guide.html">Python</a></li> <li><a href="streaming-programming-guide.html">Spark Streaming</a></li> </ul> </li> @@ -54,10 +55,9 @@ <li class="dropdown"> <a href="#" class="dropdown-toggle" data-toggle="dropdown">API (Scaladoc)<b class="caret"></b></a> <ul class="dropdown-menu"> - <li><a href="api/core/index.html">Spark</a></li> - <li><a href="api/examples/index.html">Spark Examples</a></li> - <li><a href="api/streaming/index.html">Spark Streaming</a></li> - <li><a href="api/bagel/index.html">Bagel</a></li> + <li><a href="api/core/index.html">Spark Scala/Java (Scaladoc)</a></li> + <li><a href="api/pyspark/index.html">Spark Python (Epydoc)</a></li> + <li><a href="api/streaming/index.html">Spark Streaming Scala/Java (Scaladoc) </a></li> </ul> </li> diff --git a/docs/_plugins/copy_api_dirs.rb b/docs/_plugins/copy_api_dirs.rb index 7654511eeb..e400dec619 100644 --- a/docs/_plugins/copy_api_dirs.rb +++ b/docs/_plugins/copy_api_dirs.rb @@ -28,3 +28,20 @@ if ENV['SKIP_SCALADOC'] != '1' cp_r(source + "/.", dest) end end + +if ENV['SKIP_EPYDOC'] != '1' + puts "Moving to python directory and building epydoc." + cd("../python") + puts `epydoc --config epydoc.conf` + + puts "Moving back into docs dir." + cd("../docs") + + puts "echo making directory pyspark" + mkdir_p "pyspark" + + puts "cp -r ../python/docs/. api/pyspark" + cp_r("../python/docs/.", "api/pyspark") + + cd("..") +end diff --git a/docs/api.md b/docs/api.md index 3a13e30f82..e86d07770a 100644 --- a/docs/api.md +++ b/docs/api.md @@ -9,3 +9,4 @@ Here you can find links to the Scaladoc generated for the Spark sbt subprojects. - [Spark Examples](api/examples/index.html) - [Spark Streaming](api/streaming/index.html) - [Bagel](api/bagel/index.html) +- [PySpark](api/pyspark/index.html) diff --git a/docs/configuration.md b/docs/configuration.md index 87cb4a6797..a7054b4321 100644 --- a/docs/configuration.md +++ b/docs/configuration.md @@ -198,25 +198,41 @@ Apart from these, the following properties are also available, and may be useful </td> </tr> <tr> + <td>spark.akka.frameSize</td> + <td>10</td> + <td> + Maximum message size to allow in "control plane" communication (for serialized tasks and task + results), in MB. Increase this if your tasks need to send back large results to the driver + (e.g. using <code>collect()</code> on a large dataset). + </td> +</tr> +<tr> <td>spark.akka.threads</td> <td>4</td> <td> Number of actor threads to use for communication. Can be useful to increase on large clusters - when the master has a lot of CPU cores. + when the driver has a lot of CPU cores. + </td> +</tr> +<tr> + <td>spark.akka.timeout</td> + <td>20</td> + <td> + Communication timeout between Spark nodes. </td> </tr> <tr> - <td>spark.master.host</td> + <td>spark.driver.host</td> <td>(local hostname)</td> <td> - Hostname or IP address for the master to listen on. + Hostname or IP address for the driver to listen on. </td> </tr> <tr> - <td>spark.master.port</td> + <td>spark.driver.port</td> <td>(random)</td> <td> - Port for the master to listen on. + Port for the driver to listen on. </td> </tr> <tr> diff --git a/docs/ec2-scripts.md b/docs/ec2-scripts.md index 6e1f7fd3b1..931b7a66bd 100644 --- a/docs/ec2-scripts.md +++ b/docs/ec2-scripts.md @@ -96,7 +96,9 @@ permissions on your private key file, you can run `launch` with the `spark-ec2` to attach a persistent EBS volume to each node for storing the persistent HDFS. - Finally, if you get errors while running your jobs, look at the slave's logs - for that job using the Mesos web UI (`http://<master-hostname>:8080`). + for that job inside of the Mesos work directory (/mnt/mesos-work). You can + also view the status of the cluster using the Mesos web UI + (`http://<master-hostname>:8080`). # Configuration diff --git a/docs/index.md b/docs/index.md index 560811ade8..c6ef507cb0 100644 --- a/docs/index.md +++ b/docs/index.md @@ -7,11 +7,11 @@ title: Spark Overview TODO(andyk): Rewrite to make the Java API a first class part of the story. {% endcomment %} -Spark is a MapReduce-like cluster computing framework designed for low-latency iterative jobs and interactive use from an -interpreter. It provides clean, language-integrated APIs in Scala and Java, with a rich array of parallel operators. Spark can -run on top of the [Apache Mesos](http://incubator.apache.org/mesos/) cluster manager, +Spark is a MapReduce-like cluster computing framework designed for low-latency iterative jobs and interactive use from an interpreter. +It provides clean, language-integrated APIs in [Scala](scala-programming-guide.html), [Java](java-programming-guide.html), and [Python](python-programming-guide.html), with a rich array of parallel operators. +Spark can run on top of the [Apache Mesos](http://incubator.apache.org/mesos/) cluster manager, [Hadoop YARN](http://hadoop.apache.org/docs/r2.0.1-alpha/hadoop-yarn/hadoop-yarn-site/YARN.html), -Amazon EC2, or without an independent resource manager ("standalone mode"). +Amazon EC2, or without an independent resource manager ("standalone mode"). # Downloading @@ -58,8 +58,15 @@ of `project/SparkBuild.scala`, then rebuilding Spark (`sbt/sbt clean compile`). * [Quick Start](quick-start.html): a quick introduction to the Spark API; start here! * [Spark Programming Guide](scala-programming-guide.html): an overview of Spark concepts, and details on the Scala API +* [Streaming Programming Guide](streaming-programming-guide.html): an API preview of Spark Streaming * [Java Programming Guide](java-programming-guide.html): using Spark from Java -* [Streaming Guide](streaming-programming-guide.html): an API preview of Spark Streaming +* [Python Programming Guide](python-programming-guide.html): using Spark from Python + +**API Docs:** + +* [Spark Java/Scala (Scaladoc)](api/core/index.html) +* [Spark Python (Epydoc)](api/pyspark/index.html) +* [Spark Streaming Java/Scala (Scaladoc)](api/streaming/index.html) **Deployment guides:** @@ -73,7 +80,6 @@ of `project/SparkBuild.scala`, then rebuilding Spark (`sbt/sbt clean compile`). * [Configuration](configuration.html): customize Spark via its configuration system * [Tuning Guide](tuning.html): best practices to optimize performance and memory use -* [API Docs (Scaladoc)](api/core/index.html) * [Bagel](bagel-programming-guide.html): an implementation of Google's Pregel on Spark * [Contributing to Spark](contributing-to-spark.html) diff --git a/docs/java-programming-guide.md b/docs/java-programming-guide.md index 188ca4995e..37a906ea1c 100644 --- a/docs/java-programming-guide.md +++ b/docs/java-programming-guide.md @@ -75,7 +75,8 @@ class has a single abstract method, `call()`, that must be implemented. ## Storage Levels RDD [storage level](scala-programming-guide.html#rdd-persistence) constants, such as `MEMORY_AND_DISK`, are -declared in the [spark.api.java.StorageLevels](api/core/index.html#spark.api.java.StorageLevels) class. +declared in the [spark.api.java.StorageLevels](api/core/index.html#spark.api.java.StorageLevels) class. To +define your own storage level, you can use StorageLevels.create(...). # Other Features diff --git a/docs/python-programming-guide.md b/docs/python-programming-guide.md new file mode 100644 index 0000000000..2012241a6a --- /dev/null +++ b/docs/python-programming-guide.md @@ -0,0 +1,117 @@ +--- +layout: global +title: Python Programming Guide +--- + + +The Spark Python API (PySpark) exposes most of the Spark features available in the Scala version to Python. +To learn the basics of Spark, we recommend reading through the +[Scala programming guide](scala-programming-guide.html) first; it should be +easy to follow even if you don't know Scala. +This guide will show how to use the Spark features described there in Python. + +# Key Differences in the Python API + +There are a few key differences between the Python and Scala APIs: + +* Python is dynamically typed, so RDDs can hold objects of different types. +* PySpark does not currently support the following Spark features: + - Special functions on RDDs of doubles, such as `mean` and `stdev` + - `lookup` + - `persist` at storage levels other than `MEMORY_ONLY` + - `sample` + - `sort` + +In PySpark, RDDs support the same methods as their Scala counterparts but take Python functions and return Python collection types. +Short functions can be passed to RDD methods using Python's [`lambda`](http://www.diveintopython.net/power_of_introspection/lambda_functions.html) syntax: + +{% highlight python %} +logData = sc.textFile(logFile).cache() +errors = logData.filter(lambda s: 'ERROR' in s.split()) +{% endhighlight %} + +You can also pass functions that are defined using the `def` keyword; this is useful for more complicated functions that cannot be expressed using `lambda`: + +{% highlight python %} +def is_error(line): + return 'ERROR' in line.split() +errors = logData.filter(is_error) +{% endhighlight %} + +Functions can access objects in enclosing scopes, although modifications to those objects within RDD methods will not be propagated to other tasks: + +{% highlight python %} +error_keywords = ["Exception", "Error"] +def is_error(line): + words = line.split() + return any(keyword in words for keyword in error_keywords) +errors = logData.filter(is_error) +{% endhighlight %} + +PySpark will automatically ship these functions to workers, along with any objects that they reference. +Instances of classes will be serialized and shipped to workers by PySpark, but classes themselves cannot be automatically distributed to workers. +The [Standalone Use](#standalone-use) section describes how to ship code dependencies to workers. + +# Installing and Configuring PySpark + +PySpark requires Python 2.6 or higher. +PySpark jobs are executed using a standard cPython interpreter in order to support Python modules that use C extensions. +We have not tested PySpark with Python 3 or with alternative Python interpreters, such as [PyPy](http://pypy.org/) or [Jython](http://www.jython.org/). +By default, PySpark's scripts will run programs using `python`; an alternate Python executable may be specified by setting the `PYSPARK_PYTHON` environment variable in `conf/spark-env.sh`. + +All of PySpark's library dependencies, including [Py4J](http://py4j.sourceforge.net/), are bundled with PySpark and automatically imported. + +Standalone PySpark jobs should be run using the `pyspark` script, which automatically configures the Java and Python environment using the settings in `conf/spark-env.sh`. +The script automatically adds the `pyspark` package to the `PYTHONPATH`. + + +# Interactive Use + +The `pyspark` script launches a Python interpreter that is configured to run PySpark jobs. To use `pyspark` interactively, first build Spark, then launch it directly from the command line without any options: + +{% highlight bash %} +$ sbt/sbt package +$ ./pyspark +{% endhighlight %} + +The Python shell can be used explore data interactively and is a simple way to learn the API: + +{% highlight python %} +>>> words = sc.textFile("/usr/share/dict/words") +>>> words.filter(lambda w: w.startswith("spar")).take(5) +[u'spar', u'sparable', u'sparada', u'sparadrap', u'sparagrass'] +>>> help(pyspark) # Show all pyspark functions +{% endhighlight %} + +By default, the `pyspark` shell creates SparkContext that runs jobs locally. +To connect to a non-local cluster, set the `MASTER` environment variable. +For example, to use the `pyspark` shell with a [standalone Spark cluster](spark-standalone.html): + +{% highlight bash %} +$ MASTER=spark://IP:PORT ./pyspark +{% endhighlight %} + + +# Standalone Use + +PySpark can also be used from standalone Python scripts by creating a SparkContext in your script and running the script using `pyspark`. +The Quick Start guide includes a [complete example](quick-start.html#a-standalone-job-in-python) of a standalone Python job. + +Code dependencies can be deployed by listing them in the `pyFiles` option in the SparkContext constructor: + +{% highlight python %} +from pyspark import SparkContext +sc = SparkContext("local", "Job Name", pyFiles=['MyFile.py', 'lib.zip', 'app.egg']) +{% endhighlight %} + +Files listed here will be added to the `PYTHONPATH` and shipped to remote worker machines. +Code dependencies can be added to an existing SparkContext using its `addPyFile()` method. + +# Where to Go from Here + +PySpark includes several sample programs using the Python API in `python/examples`. +You can run them by passing the files to the `pyspark` script -- for example `./pyspark python/examples/wordcount.py`. +Each example program prints usage help when run without any arguments. + +We currently provide [API documentation](api/pyspark/index.html) for the Python API as Epydoc. +Many of the RDD method descriptions contain [doctests](http://docs.python.org/2/library/doctest.html) that provide additional usage examples. diff --git a/docs/quick-start.md b/docs/quick-start.md index 177cb14551..a4c4c9a8fb 100644 --- a/docs/quick-start.md +++ b/docs/quick-start.md @@ -6,7 +6,8 @@ title: Quick Start * This will become a table of contents (this text will be scraped). {:toc} -This tutorial provides a quick introduction to using Spark. We will first introduce the API through Spark's interactive Scala shell (don't worry if you don't know Scala -- you will not need much for this), then show how to write standalone jobs in Scala and Java. See the [programming guide](scala-programming-guide.html) for a more complete reference. +This tutorial provides a quick introduction to using Spark. We will first introduce the API through Spark's interactive Scala shell (don't worry if you don't know Scala -- you will not need much for this), then show how to write standalone jobs in Scala, Java, and Python. +See the [programming guide](scala-programming-guide.html) for a more complete reference. To follow along with this guide, you only need to have successfully built Spark on one machine. Simply go into your Spark directory and run: @@ -200,6 +201,16 @@ To build the job, we also write a Maven `pom.xml` file that lists Spark as a dep <name>Simple Project</name> <packaging>jar</packaging> <version>1.0</version> + <repositories> + <repository> + <id>Spray.cc repository</id> + <url>http://repo.spray.cc</url> + </repository> + <repository> + <id>Typesafe repository</id> + <url>http://repo.typesafe.com/typesafe/releases</url> + </repository> + </repositories> <dependencies> <dependency> <!-- Spark dependency --> <groupId>org.spark-project</groupId> @@ -230,3 +241,40 @@ Lines with a: 8422, Lines with b: 1836 {% endhighlight %} This example only runs the job locally; for a tutorial on running jobs across several machines, see the [Standalone Mode](spark-standalone.html) documentation, and consider using a distributed input source, such as HDFS. + +# A Standalone Job In Python +Now we will show how to write a standalone job using the Python API (PySpark). + +As an example, we'll create a simple Spark job, `SimpleJob.py`: + +{% highlight python %} +"""SimpleJob.py""" +from pyspark import SparkContext + +logFile = "/var/log/syslog" # Should be some file on your system +sc = SparkContext("local", "Simple job") +logData = sc.textFile(logFile).cache() + +numAs = logData.filter(lambda s: 'a' in s).count() +numBs = logData.filter(lambda s: 'b' in s).count() + +print "Lines with a: %i, lines with b: %i" % (numAs, numBs) +{% endhighlight %} + + +This job simply counts the number of lines containing 'a' and the number containing 'b' in a system log file. +Like in the Scala and Java examples, we use a SparkContext to create RDDs. +We can pass Python functions to Spark, which are automatically serialized along with any variables that they reference. +For jobs that use custom classes or third-party libraries, we can add those code dependencies to SparkContext to ensure that they will be available on remote machines; this is described in more detail in the [Python programming guide](python-programming-guide). +`SimpleJob` is simple enough that we do not need to specify any code dependencies. + +We can run this job using the `pyspark` script: + +{% highlight python %} +$ cd $SPARK_HOME +$ ./pyspark SimpleJob.py +... +Lines with a: 8422, Lines with b: 1836 +{% endhighlight python %} + +This example only runs the job locally; for a tutorial on running jobs across several machines, see the [Standalone Mode](spark-standalone.html) documentation, and consider using a distributed input source, such as HDFS. diff --git a/docs/scala-programming-guide.md b/docs/scala-programming-guide.md index 7350eca837..301b330a79 100644 --- a/docs/scala-programming-guide.md +++ b/docs/scala-programming-guide.md @@ -301,7 +301,8 @@ We recommend going through the following process to select one: * Use the replicated storage levels if you want fast fault recovery (e.g. if using Spark to serve requests from a web application). *All* the storage levels provide full fault tolerance by recomputing lost data, but the replicated ones let you continue running tasks on the RDD without waiting to recompute a lost partition. - + +If you want to define your own storage level (say, with replication factor of 3 instead of 2), then use the function factor method `apply()` of the [`StorageLevel`](api/core/index.html#spark.storage.StorageLevel$) singleton object. # Shared Variables diff --git a/docs/spark-standalone.md b/docs/spark-standalone.md index e0ba7c35cb..bf296221b8 100644 --- a/docs/spark-standalone.md +++ b/docs/spark-standalone.md @@ -51,11 +51,11 @@ Finally, the following configuration options can be passed to the master and wor </tr> <tr> <td><code>-c CORES</code>, <code>--cores CORES</code></td> - <td>Number of CPU cores to use (default: all available); only on worker</td> + <td>Total CPU cores to allow Spark jobs to use on the machine (default: all available); only on worker</td> </tr> <tr> <td><code>-m MEM</code>, <code>--memory MEM</code></td> - <td>Amount of memory to use, in a format like 1000M or 2G (default: your machine's total RAM minus 1 GB); only on worker</td> + <td>Total amount of memory to allow Spark jobs to use on the machine, in a format like 1000M or 2G (default: your machine's total RAM minus 1 GB); only on worker</td> </tr> <tr> <td><code>-d DIR</code>, <code>--work-dir DIR</code></td> @@ -66,9 +66,20 @@ Finally, the following configuration options can be passed to the master and wor # Cluster Launch Scripts -To launch a Spark standalone cluster with the deploy scripts, you need to set up two files, `conf/spark-env.sh` and `conf/slaves`. The `conf/spark-env.sh` file lets you specify global settings for the master and slave instances, such as memory, or port numbers to bind to, while `conf/slaves` is a list of slave nodes. The system requires that all the slave machines have the same configuration files, so *copy these files to each machine*. +To launch a Spark standalone cluster with the deploy scripts, you need to create a file called `conf/slaves` in your Spark directory, which should contain the hostnames of all the machines where you would like to start Spark workers, one per line. The master machine must be able to access each of the slave machines via password-less `ssh` (using a private key). For testing, you can just put `localhost` in this file. -In `conf/spark-env.sh`, you can set the following parameters, in addition to the [standard Spark configuration settings](configuration.html): +Once you've set up this fine, you can launch or stop your cluster with the following shell scripts, based on Hadoop's deploy scripts, and available in `SPARK_HOME/bin`: + +- `bin/start-master.sh` - Starts a master instance on the machine the script is executed on. +- `bin/start-slaves.sh` - Starts a slave instance on each machine specified in the `conf/slaves` file. +- `bin/start-all.sh` - Starts both a master and a number of slaves as described above. +- `bin/stop-master.sh` - Stops the master that was started via the `bin/start-master.sh` script. +- `bin/stop-slaves.sh` - Stops the slave instances that were started via `bin/start-slaves.sh`. +- `bin/stop-all.sh` - Stops both the master and the slaves as described above. + +Note that these scripts must be executed on the machine you want to run the Spark master on, not your local machine. + +You can optionally configure the cluster further by setting environment variables in `conf/spark-env.sh`. Create this file by starting with the `conf/spark-env.sh.template`, and _copy it to all your worker machines_ for the settings to take effect. The following settings are available: <table class="table"> <tr><th style="width:21%">Environment Variable</th><th>Meaning</th></tr> @@ -89,35 +100,23 @@ In `conf/spark-env.sh`, you can set the following parameters, in addition to the <td>Start the Spark worker on a specific port (default: random)</td> </tr> <tr> + <td><code>SPARK_WORKER_DIR</code></td> + <td>Directory to run jobs in, which will include both logs and scratch space (default: SPARK_HOME/work)</td> + </tr> + <tr> <td><code>SPARK_WORKER_CORES</code></td> - <td>Number of cores to use (default: all available cores)</td> + <td>Total number of cores to allow Spark jobs to use on the machine (default: all available cores)</td> </tr> <tr> <td><code>SPARK_WORKER_MEMORY</code></td> - <td>How much memory to use, e.g. 1000M, 2G (default: total memory minus 1 GB)</td> + <td>Total amount of memory to allow Spark jobs to use on the machine, e.g. 1000M, 2G (default: total memory minus 1 GB); note that each job's <i>individual</i> memory is configured using <code>SPARK_MEM</code></td> </tr> <tr> <td><code>SPARK_WORKER_WEBUI_PORT</code></td> <td>Port for the worker web UI (default: 8081)</td> </tr> - <tr> - <td><code>SPARK_WORKER_DIR</code></td> - <td>Directory to run jobs in, which will include both logs and scratch space (default: SPARK_HOME/work)</td> - </tr> </table> -In `conf/slaves`, include a list of all machines where you would like to start a Spark worker, one per line. The master machine must be able to access each of the slave machines via password-less `ssh` (using a private key). For testing purposes, you can have a single `localhost` entry in the slaves file. - -Once you've set up these configuration files, you can launch or stop your cluster with the following shell scripts, based on Hadoop's deploy scripts, and available in `SPARK_HOME/bin`: - -- `bin/start-master.sh` - Starts a master instance on the machine the script is executed on. -- `bin/start-slaves.sh` - Starts a slave instance on each machine specified in the `conf/slaves` file. -- `bin/start-all.sh` - Starts both a master and a number of slaves as described above. -- `bin/stop-master.sh` - Stops the master that was started via the `bin/start-master.sh` script. -- `bin/stop-slaves.sh` - Stops the slave instances that were started via `bin/start-slaves.sh`. -- `bin/stop-all.sh` - Stops both the master and the slaves as described above. - -Note that the scripts must be executed on the machine you want to run the Spark master on, not your local machine. # Connecting a Job to the Cluster diff --git a/docs/streaming-programming-guide.md b/docs/streaming-programming-guide.md index b6da7af654..d408e80359 100644 --- a/docs/streaming-programming-guide.md +++ b/docs/streaming-programming-guide.md @@ -34,8 +34,8 @@ The StreamingContext is used to creating InputDStreams from input sources: {% highlight scala %} // Assuming ssc is the StreamingContext -ssc.networkStream(hostname, port) // Creates a stream that uses a TCP socket to read data from hostname:port -ssc.textFileStream(directory) // Creates a stream by monitoring and processing new files in a HDFS directory +ssc.networkStream(hostname, port) // Creates a stream that uses a TCP socket to read data from hostname:port +ssc.textFileStream(directory) // Creates a stream by monitoring and processing new files in a HDFS directory {% endhighlight %} A complete list of input sources is available in the [StreamingContext API documentation](api/streaming/index.html#spark.streaming.StreamingContext). Data received from these sources can be processed using DStream operations, which are explained next. @@ -50,18 +50,18 @@ Once an input DStream has been created, you can transform it using _DStream oper DStreams support many of the transformations available on normal Spark RDD's: <table class="table"> -<tr><th style="width:25%">Transformation</th><th>Meaning</th></tr> +<tr><th style="width:30%">Transformation</th><th>Meaning</th></tr> <tr> <td> <b>map</b>(<i>func</i>) </td> - <td> Returns a new DStream formed by passing each element of the source through a function <i>func</i>. </td> + <td> Returns a new DStream formed by passing each element of the source DStream through a function <i>func</i>. </td> </tr> <tr> <td> <b>filter</b>(<i>func</i>) </td> - <td> Returns a new stream formed by selecting those elements of the source on which <i>func</i> returns true. </td> + <td> Returns a new DStream formed by selecting those elements of the source DStream on which <i>func</i> returns true. </td> </tr> <tr> <td> <b>flatMap</b>(<i>func</i>) </td> - <td> Similar to map, but each input item can be mapped to 0 or more output items (so <i>func</i> should return a Seq rather than a single item). </td> + <td> Similar to map, but each input item can be mapped to 0 or more output items (so <i>func</i> should return a <code>Seq</code> rather than a single item). </td> </tr> <tr> <td> <b>mapPartitions</b>(<i>func</i>) </td> @@ -70,73 +70,92 @@ DStreams support many of the transformations available on normal Spark RDD's: </tr> <tr> <td> <b>union</b>(<i>otherStream</i>) </td> - <td> Return a new stream that contains the union of the elements in the source stream and the argument. </td> + <td> Return a new DStream that contains the union of the elements in the source DStream and the argument DStream. </td> +</tr> +<tr> + <td> <b>count</b>() </td> + <td> Returns a new DStream of single-element RDDs by counting the number of elements in each RDD of the source DStream. </td> +</tr> +<tr> + <td> <b>reduce</b>(<i>func</i>) </td> + <td> Returns a new DStream of single-element RDDs by aggregating the elements in each RDD of the source DStream using a function <i>func</i> (which takes two arguments and returns one). The function should be associative so that it can be computed in parallel. </td> +</tr> +<tr> + <td> <b>countByValue</b>() </td> + <td> When called on a DStream of elements of type K, returns a new DStream of (K, Long) pairs where the value of each key is its frequency in each RDD of the source DStream. </td> </tr> <tr> <td> <b>groupByKey</b>([<i>numTasks</i>]) </td> - <td> When called on a stream of (K, V) pairs, returns a stream of (K, Seq[V]) pairs. <br /> -<b>Note:</b> By default, this uses only 8 parallel tasks to do the grouping. You can pass an optional <code>numTasks</code> argument to set a different number of tasks. + <td> When called on a DStream of (K, V) pairs, returns a new DStream of (K, Seq[V]) pairs by grouping together all the values of each key in the RDDs of the source DStream. <br /> + <b>Note:</b> By default, this uses Spark's default number of parallel tasks (2 for local machine, 8 for a cluser) to do the grouping. You can pass an optional <code>numTasks</code> argument to set a different number of tasks. </td> </tr> <tr> <td> <b>reduceByKey</b>(<i>func</i>, [<i>numTasks</i>]) </td> - <td> When called on a stream of (K, V) pairs, returns a stream of (K, V) pairs where the values for each key are aggregated using the given reduce function. Like in <code>groupByKey</code>, the number of reduce tasks is configurable through an optional second argument. </td> + <td> When called on a DStream of (K, V) pairs, returns a new DStream of (K, V) pairs where the values for each key are aggregated using the given reduce function. Like in <code>groupByKey</code>, the number of reduce tasks is configurable through an optional second argument. </td> </tr> <tr> <td> <b>join</b>(<i>otherStream</i>, [<i>numTasks</i>]) </td> - <td> When called on streams of type (K, V) and (K, W), returns a stream of (K, (V, W)) pairs with all pairs of elements for each key. </td> + <td> When called on two DStreams of (K, V) and (K, W) pairs, returns a new DStream of (K, (V, W)) pairs with all pairs of elements for each key. </td> </tr> <tr> <td> <b>cogroup</b>(<i>otherStream</i>, [<i>numTasks</i>]) </td> - <td> When called on DStream of type (K, V) and (K, W), returns a DStream of (K, Seq[V], Seq[W]) tuples.</td> -</tr> -<tr> - <td> <b>reduce</b>(<i>func</i>) </td> - <td> Returns a new DStream of single-element RDDs by aggregating the elements of the stream using a function func (which takes two arguments and returns one). The function should be associative so that it can be computed correctly in parallel. </td> + <td> When called on DStream of (K, V) and (K, W) pairs, returns a new DStream of (K, Seq[V], Seq[W]) tuples.</td> </tr> <tr> <td> <b>transform</b>(<i>func</i>) </td> <td> Returns a new DStream by applying func (a RDD-to-RDD function) to every RDD of the stream. This can be used to do arbitrary RDD operations on the DStream. </td> </tr> +<tr> + <td> <b>updateStateByKey</b>(<i>func</i>) </td> + <td> Return a new "state" DStream where the state for each key is updated by applying the given function on the previous state of the key and the new values of each key. This can be used to track session state by using the session-id as the key and updating the session state as new data is received.</td> +</tr> + </table> -Spark Streaming features windowed computations, which allow you to report statistics over a sliding window of data. All window functions take a <i>windowDuration</i>, which represents the width of the window and a <i>slideTime</i>, which represents the frequency during which the window is calculated. +Spark Streaming features windowed computations, which allow you to apply transformations over a sliding window of data. All window functions take a <i>windowDuration</i>, which represents the width of the window and a <i>slideTime</i>, which represents the frequency during which the window is calculated. <table class="table"> -<tr><th style="width:25%">Transformation</th><th>Meaning</th></tr> +<tr><th style="width:30%">Transformation</th><th>Meaning</th></tr> <tr> - <td> <b>window</b>(<i>windowDuration</i>, </i>slideTime</i>) </td> - <td> Return a new stream which is computed based on windowed batches of the source stream. <i>windowDuration</i> is the width of the window and <i>slideTime</i> is the frequency during which the window is calculated. Both times must be multiples of the batch interval. + <td> <b>window</b>(<i>windowDuration</i>, </i>slideDuration</i>) </td> + <td> Return a new DStream which is computed based on windowed batches of the source DStream. <i>windowDuration</i> is the width of the window and <i>slideTime</i> is the frequency during which the window is calculated. Both times must be multiples of the batch interval. </td> </tr> <tr> - <td> <b>countByWindow</b>(<i>windowDuration</i>, </i>slideTime</i>) </td> + <td> <b>countByWindow</b>(<i>windowDuration</i>, </i>slideDuration</i>) </td> <td> Return a sliding count of elements in the stream. <i>windowDuration</i> and <i>slideDuration</i> are exactly as defined in <code>window()</code>. </td> </tr> <tr> - <td> <b>reduceByWindow</b>(<i>func</i>, <i>windowDuration</i>, </i>slideDuration</i>) </td> + <td> <b>reduceByWindow</b>(<i>func</i>, <i>windowDuration</i>, <i>slideDuration</i>) </td> <td> Return a new single-element stream, created by aggregating elements in the stream over a sliding interval using <i>func</i>. The function should be associative so that it can be computed correctly in parallel. <i>windowDuration</i> and <i>slideDuration</i> are exactly as defined in <code>window()</code>. </td> </tr> <tr> - <td> <b>groupByKeyAndWindow</b>(windowDuration, slideDuration, [<i>numTasks</i>]) + <td> <b>groupByKeyAndWindow</b>(<i>windowDuration</i>, <i>slideDuration</i>, [<i>numTasks</i>]) </td> - <td> When called on a stream of (K, V) pairs, returns a stream of (K, Seq[V]) pairs over a sliding window. <br /> -<b>Note:</b> By default, this uses only 8 parallel tasks to do the grouping. You can pass an optional <code>numTasks</code> argument to set a different number of tasks. <i>windowDuration</i> and <i>slideDuration</i> are exactly as defined in <code>window()</code>. -</td> + <td> When called on a DStream of (K, V) pairs, returns a new DStream of (K, Seq[V]) pairs by grouping together values of each key over batches in a sliding window. <br /> +<b>Note:</b> By default, this uses Spark's default number of parallel tasks (2 for local machine, 8 for a cluser) to do the grouping. You can pass an optional <code>numTasks</code> argument to set a different number of tasks.</td> </tr> <tr> - <td> <b>reduceByKeyAndWindow</b>(<i>func</i>, [<i>numTasks</i>]) </td> - <td> When called on a stream of (K, V) pairs, returns a stream of (K, V) pairs where the values for each key are aggregated using the given reduce function over batches within a sliding window. Like in <code>groupByKeyAndWindow</code>, the number of reduce tasks is configurable through an optional second argument. + <td> <b>reduceByKeyAndWindow</b>(<i>func</i>, <i>windowDuration</i>, <i>slideDuration</i>, [<i>numTasks</i>]) </td> + <td> When called on a DStream of (K, V) pairs, returns a new DStream of (K, V) pairs where the values for each key are aggregated using the given reduce function <i>func</i> over batches in a sliding window. Like in <code>groupByKeyAndWindow</code>, the number of reduce tasks is configurable through an optional second argument. <i>windowDuration</i> and <i>slideDuration</i> are exactly as defined in <code>window()</code>. </td> </tr> <tr> - <td> <b>countByKeyAndWindow</b>([<i>numTasks</i>]) </td> - <td> When called on a stream of (K, V) pairs, returns a stream of (K, Int) pairs where the values for each key are the count within a sliding window. Like in <code>countByKeyAndWindow</code>, the number of reduce tasks is configurable through an optional second argument. + <td> <b>reduceByKeyAndWindow</b>(<i>func</i>, <i>invFunc</i>, <i>windowDuration</i>, <i>slideDuration</i>, [<i>numTasks</i>]) </td> + <td> A more efficient version of the above <code>reduceByKeyAndWindow()</code> where the reduce value of each window is calculated + incrementally using the reduce values of the previous window. This is done by reducing the new data that enter the sliding window, and "inverse reducing" the old data that leave the window. An example would be that of "adding" and "subtracting" counts of keys as the window slides. However, it is applicable to only "invertible reduce functions", that is, those reduce functions which have a corresponding "inverse reduce" function (taken as parameter <i>invFunc</i>. Like in <code>groupByKeyAndWindow</code>, the number of reduce tasks is configurable through an optional second argument. <i>windowDuration</i> and <i>slideDuration</i> are exactly as defined in <code>window()</code>. -</td> +</td> +</tr> +<tr> + <td> <b>countByValueAndWindow</b>(<i>windowDuration</i>, <i>slideDuration</i>, [<i>numTasks</i>]) </td> + <td> When called on a DStream of (K, V) pairs, returns a new DStream of (K, Long) pairs where the value of each key is its frequency within a sliding window. Like in <code>groupByKeyAndWindow</code>, the number of reduce tasks is configurable through an optional second argument. + <i>windowDuration</i> and <i>slideDuration</i> are exactly as defined in <code>window()</code>. +</td> </tr> </table> @@ -147,7 +166,7 @@ A complete list of DStream operations is available in the API documentation of [ When an output operator is called, it triggers the computation of a stream. Currently the following output operators are defined: <table class="table"> -<tr><th style="width:25%">Operator</th><th>Meaning</th></tr> +<tr><th style="width:30%">Operator</th><th>Meaning</th></tr> <tr> <td> <b>foreach</b>(<i>func</i>) </td> <td> The fundamental output operator. Applies a function, <i>func</i>, to each RDD generated from the stream. This function should have side effects, such as printing output, saving the RDD to external files, or writing it over the network to an external system. </td> @@ -176,11 +195,6 @@ When an output operator is called, it triggers the computation of a stream. Curr </table> -## DStream Persistence -Similar to RDDs, DStreams also allow developers to persist the stream's data in memory. That is, using `persist()` method on a DStream would automatically persist every RDD of that DStream in memory. This is useful if the data in the DStream will be computed multiple times (e.g., multiple DStream operations on the same data). For window-based operations like `reduceByWindow` and `reduceByKeyAndWindow` and state-based operations like `updateStateByKey`, this is implicitly true. Hence, DStreams generated by window-based operations are automatically persisted in memory, without the developer calling `persist()`. - -Note that, unlike RDDs, the default persistence level of DStreams keeps the data serialized in memory. This is further discussed in the [Performance Tuning](#memory-tuning) section. More information on different persistence levels can be found in [Spark Programming Guide](scala-programming-guide.html#rdd-persistence). - # Starting the Streaming computation All the above DStream operations are completely lazy, that is, the operations will start executing only after the context is started by using {% highlight scala %} @@ -192,8 +206,8 @@ Conversely, the computation can be stopped by using ssc.stop() {% endhighlight %} -# Example - NetworkWordCount.scala -A good example to start off is the spark.streaming.examples.NetworkWordCount. This example counts the words received from a network server every second. Given below is the relevant sections of the source code. You can find the full source code in <Spark repo>/streaming/src/main/scala/spark/streaming/examples/WordCountNetwork.scala. +# Example +A simple example to start off is the [NetworkWordCount](https://github.com/mesos/spark/tree/master/examples/src/main/scala/spark/streaming/examples/NetworkWordCount.scala). This example counts the words received from a network server every second. Given below is the relevant sections of the source code. You can find the full source code in `<Spark repo>/streaming/src/main/scala/spark/streaming/examples/WordCountNetwork.scala` . {% highlight scala %} import spark.streaming.{Seconds, StreamingContext} @@ -260,6 +274,31 @@ Time: 1357008430000 ms </td> </table> +You can find more examples in `<Spark repo>/streaming/src/main/scala/spark/streaming/examples/`. They can be run in the similar manner using `./run spark.streaming.examples....` . Executing without any parameter would give the required parameter list. Further explanation to run them can be found in comments in the files. + +# DStream Persistence +Similar to RDDs, DStreams also allow developers to persist the stream's data in memory. That is, using `persist()` method on a DStream would automatically persist every RDD of that DStream in memory. This is useful if the data in the DStream will be computed multiple times (e.g., multiple operations on the same data). For window-based operations like `reduceByWindow` and `reduceByKeyAndWindow` and state-based operations like `updateStateByKey`, this is implicitly true. Hence, DStreams generated by window-based operations are automatically persisted in memory, without the developer calling `persist()`. + +For input streams that receive data from the network (that is, subclasses of NetworkInputDStream like FlumeInputDStream and KafkaInputDStream), the default persistence level is set to replicate the data to two nodes for fault-tolerance. + +Note that, unlike RDDs, the default persistence level of DStreams keeps the data serialized in memory. This is further discussed in the [Performance Tuning](#memory-tuning) section. More information on different persistence levels can be found in [Spark Programming Guide](scala-programming-guide.html#rdd-persistence). + +# RDD Checkpointing within DStreams +DStreams created by stateful operations like `updateStateByKey` require the RDDs in the DStream to be periodically saved to HDFS files for checkpointing. This is because, unless checkpointed, the lineage of operations of the state RDDs can increase indefinitely (since each RDD in the DStream depends on the previous RDD). This leads to two problems - (i) the size of Spark tasks increase proportionally with the RDD lineage leading higher task launch times, (ii) no limit on the amount of recomputation required on failure. Checkpointing RDDs at some interval by writing them to HDFS allows the lineage to be truncated. Note that checkpointing also incurs the cost of saving to HDFS which may cause the corresponding batch to take longer to process. Hence, the interval of checkpointing needs to be set carefully. At small batch sizes (say 1 second), checkpointing every batch may significantly reduce operation throughput. Conversely, checkpointing too slowly causes the lineage and task sizes to grow which may have detrimental effects. Typically, a checkpoint interval of 5 - 10 times of sliding interval of a DStream is good setting to try. + +To enable checkpointing, the developer has to provide the HDFS path to which RDD will be saved. This is done by using + +{% highlight scala %} +ssc.checkpoint(hdfsPath) // assuming ssc is the StreamingContext +{% endhighlight %} + +The interval of checkpointing of a DStream can be set by using + +{% highlight scala %} +dstream.checkpoint(checkpointInterval) // checkpointInterval must be a multiple of slide duration of dstream +{% endhighlight %} + +For DStreams that must be checkpointed (that is, DStreams created by `updateStateByKey` and `reduceByKeyAndWindow` with inverse function), the checkpoint interval of the DStream is by default set to a multiple of the DStream's sliding interval such that its at least 10 seconds. # Performance Tuning @@ -273,17 +312,21 @@ Getting the best performance of a Spark Streaming application on a cluster requi There are a number of optimizations that can be done in Spark to minimize the processing time of each batch. These have been discussed in detail in [Tuning Guide](tuning.html). This section highlights some of the most important ones. ### Level of Parallelism -Cluster resources maybe underutilized if the number of parallel tasks used in any stage of the computation is not high enough. For example, for distributed reduce operations like `reduceByKey` and `reduceByKeyAndWindow`, the default number of parallel tasks is 8. You can pass the level of parallelism as an argument (see the [`spark.PairDStreamFunctions`](api/streaming/index.html#spark.PairDStreamFunctions) documentation), or set the system property `spark.default.parallelism` to change the default. +Cluster resources maybe under-utilized if the number of parallel tasks used in any stage of the computation is not high enough. For example, for distributed reduce operations like `reduceByKey` and `reduceByKeyAndWindow`, the default number of parallel tasks is 8. You can pass the level of parallelism as an argument (see the [`spark.PairDStreamFunctions`](api/streaming/index.html#spark.PairDStreamFunctions) documentation), or set the system property `spark.default.parallelism` to change the default. ### Data Serialization The overhead of data serialization can be significant, especially when sub-second batch sizes are to be achieved. There are two aspects to it. -* Serialization of RDD data in Spark: Please refer to the detailed discussion on data serialization in the [Tuning Guide](tuning.html). However, note that unlike Spark, by default RDDs are persisted as serialized byte arrays to minimize pauses related to GC. -* Serialization of input data: To ingest external data into Spark, data received as bytes (say, from the network) needs to deserialized from bytes and re-serialized into Spark's serialization format. Hence, the deserialization overhead of input data may be a bottleneck. + +* **Serialization of RDD data in Spark**: Please refer to the detailed discussion on data serialization in the [Tuning Guide](tuning.html). However, note that unlike Spark, by default RDDs are persisted as serialized byte arrays to minimize pauses related to GC. + +* **Serialization of input data**: To ingest external data into Spark, data received as bytes (say, from the network) needs to deserialized from bytes and re-serialized into Spark's serialization format. Hence, the deserialization overhead of input data may be a bottleneck. ### Task Launching Overheads If the number of tasks launched per second is high (say, 50 or more per second), then the overhead of sending out tasks to the slaves maybe significant and will make it hard to achieve sub-second latencies. The overhead can be reduced by the following changes: -* Task Serialization: Using Kryo serialization for serializing tasks can reduced the task sizes, and therefore reduce the time taken to send them to the slaves. -* Execution mode: Running Spark in Standalone mode or coarse-grained Mesos mode leads to better task launch times than the fine-grained Mesos mode. Please refer to the [Running on Mesos guide](running-on-mesos.html) for more details. + +* **Task Serialization**: Using Kryo serialization for serializing tasks can reduced the task sizes, and therefore reduce the time taken to send them to the slaves. + +* **Execution mode**: Running Spark in Standalone mode or coarse-grained Mesos mode leads to better task launch times than the fine-grained Mesos mode. Please refer to the [Running on Mesos guide](running-on-mesos.html) for more details. These changes may reduce batch processing time by 100s of milliseconds, thus allowing sub-second batch size to be viable. ## Setting the Right Batch Size @@ -292,22 +335,121 @@ For a Spark Streaming application running on a cluster to be stable, the process A good approach to figure out the right batch size for your application is to test it with a conservative batch size (say, 5-10 seconds) and a low data rate. To verify whether the system is able to keep up with data rate, you can check the value of the end-to-end delay experienced by each processed batch (in the Spark master logs, find the line having the phrase "Total delay"). If the delay is maintained to be less than the batch size, then system is stable. Otherwise, if the delay is continuously increasing, it means that the system is unable to keep up and it therefore unstable. Once you have an idea of a stable configuration, you can try increasing the data rate and/or reducing the batch size. Note that momentary increase in the delay due to temporary data rate increases maybe fine as long as the delay reduces back to a low value (i.e., less than batch size). ## 24/7 Operation -By default, Spark does not forget any of the metadata (RDDs generated, stages processed, etc.). But for a Spark Streaming application to operate 24/7, it is necessary for Spark to do periodic cleanup of it metadata. This can be enabled by setting the Java system property `spark.cleaner.delay` to the number of minutes you want any metadata to persist. For example, setting `spark.cleaner.delay` to 10 would cause Spark periodically cleanup all metadata and persisted RDDs that are older than 10 minutes. Note, that this property needs to be set before the SparkContext is created. +By default, Spark does not forget any of the metadata (RDDs generated, stages processed, etc.). But for a Spark Streaming application to operate 24/7, it is necessary for Spark to do periodic cleanup of it metadata. This can be enabled by setting the Java system property `spark.cleaner.delay` to the number of seconds you want any metadata to persist. For example, setting `spark.cleaner.delay` to 600 would cause Spark periodically cleanup all metadata and persisted RDDs that are older than 10 minutes. Note, that this property needs to be set before the SparkContext is created. This value is closely tied with any window operation that is being used. Any window operation would require the input data to be persisted in memory for at least the duration of the window. Hence it is necessary to set the delay to at least the value of the largest window operation used in the Spark Streaming application. If this delay is set too low, the application will throw an exception saying so. ## Memory Tuning Tuning the memory usage and GC behavior of Spark applications have been discussed in great detail in the [Tuning Guide](tuning.html). It is recommended that you read that. In this section, we highlight a few customizations that are strongly recommended to minimize GC related pauses in Spark Streaming applications and achieving more consistent batch processing times. -* <b>Default persistence level of DStreams</b>: Unlike RDDs, the default persistence level of DStreams serializes the data in memory (that is, [StorageLevel.MEMORY_ONLY_SER](api/core/index.html#spark.storage.StorageLevel$) for DStream compared to [StorageLevel.MEMORY_ONLY](api/core/index.html#spark.storage.StorageLevel$) for RDDs). Even though keeping the data serialized incurs a higher serialization overheads, it significantly reduces GC pauses. +* **Default persistence level of DStreams**: Unlike RDDs, the default persistence level of DStreams serializes the data in memory (that is, [StorageLevel.MEMORY_ONLY_SER](api/core/index.html#spark.storage.StorageLevel$) for DStream compared to [StorageLevel.MEMORY_ONLY](api/core/index.html#spark.storage.StorageLevel$) for RDDs). Even though keeping the data serialized incurs a higher serialization overheads, it significantly reduces GC pauses. + +* **Concurrent garbage collector**: Using the concurrent mark-and-sweep GC further minimizes the variability of GC pauses. Even though concurrent GC is known to reduce the overall processing throughput of the system, its use is still recommended to achieve more consistent batch processing times. -* <b>Concurrent garbage collector</b>: Using the concurrent mark-and-sweep GC further minimizes the variability of GC pauses. Even though concurrent GC is known to reduce the overall processing throughput of the system, its use is still recommended to achieve more consistent batch processing times. +# Fault-tolerance Properties +There are two aspects to fault-tolerance - failure of a worker node and that of a driver node. In this section, we are going to discuss the fault-tolerance behavior and the semantics of the processed data. -# Master Fault-tolerance (Alpha) -TODO +## Failure of a Worker Node +In case of the worker node failure, none of the processed data will be lost because -* Checkpointing of DStream graph +1. All the input data is fault-tolerant (either the data is on HDFS, or it replicated Spark Streaming if received from the network) +1. All intermediate data is expressed as RDDs with their lineage to the input data, which allows Spark to recompute any part of the intermediate data is lost to worker node failure. + +If the worker node where a network data receiver is running fails, then the receiver will be restarted on a different node and it will continue to receive data. However, data that was accepted by the receiver but not yet replicated to other Spark nodes may be lost, which is a fraction of a second of data. + +Since all data is modeled as RDDs with their lineage of deterministic operations, any recomputation always leads to the same result. As a result, all DStream transformations are guaranteed to have _exactly-once_ semantics. That is, the final transformed result will be same even if there were was a worker node failure. However, output operations (like `foreach`) have _at-least once_ semantics, that is, the transformed data may get written to an external entity more than once in the event of a worker failure. While this is acceptable for saving to HDFS using the `saveAs*Files` operations (as the file will simply get over-written by the same data), additional transactions-like mechanisms may be necessary to achieve exactly-once semantics for output operations. + +## Failure of a Driver Node +A system that is required to operate 24/7 needs to be able tolerate the failure of the drive node as well. Spark Streaming does this by saving the state of the DStream computation periodically to a HDFS file, that can be used to restart the streaming computation in the event of a failure of the driver node. To elaborate, the following state is periodically saved to a file. + +1. The DStream operator graph (input streams, output streams, etc.) +1. The configuration of each DStream (checkpoint interval, etc.) +1. The RDD checkpoint files of each DStream + +All this is periodically saved in the file `<checkpoint directory>/graph` where `<checkpoint directory>` is the HDFS path set using `ssc.checkpoint(...)` as described earlier. To recover, a new Streaming Context can be created with this directory by using + +{% highlight scala %} +val ssc = new StreamingContext(checkpointDirectory) +{% endhighlight %} + +Calling `ssc.start()` on this new context will restart the receivers and the stream computations. + +In case of stateful operations (that is, `updateStateByKey` and `reduceByKeyAndWindow` with inverse function), the intermediate data at the time of failure also needs to be recomputed.This requires two things - (i) the RDD checkpoints and (ii) the data received since the checkpoints. In the current _alpha_ release, the input data received from the network is not saved durably across driver failures (the data is only replicated in memory of the worker processes and gets lost when the driver fails). Only with file input streams (where the data is already durably stored) is the recovery from driver failure complete and all intermediate data is recomputed. In a future release, this will be true for all input streams. Note that for non-stateful operations, with _all_ input streams, the system will recover and continue receiving and processing new data. + +To understand the behavior of the system under driver failure, lets consider what will happen with a file input stream Specifically, in the case of the file input stream, it will correctly identify new files that were created while the driver was down and process them in the same way as it would have if the driver had not failed. To explain further in the case of file input stream, we shall use an example. Lets say, files are being generated every second, and a Spark Streaming program reads every new file and output the number of lines in the file. This is what the sequence of outputs would be with and without a driver failure. + +<table class="table"> + <!-- Results table headers --> + <tr> + <th> Time </th> + <th> Number of lines in input file </th> + <th> Output without driver failure </th> + <th> Output with driver failure </th> + </tr> + <tr> + <td>1</td> + <td>10</td> + <td>10</td> + <td>10</td> + </tr> + <tr> + <td>2</td> + <td>20</td> + <td>20</td> + <td>20</td> + </tr> + <tr> + <td>3</td> + <td>30</td> + <td>30</td> + <td>30</td> + </tr> + <tr> + <td>4</td> + <td>40</td> + <td>40</td> + <td>[DRIVER FAILS]<br />no output</td> + </tr> + <tr> + <td>5</td> + <td>50</td> + <td>50</td> + <td>no output</td> + </tr> + <tr> + <td>6</td> + <td>60</td> + <td>60</td> + <td>no output</td> + </tr> + <tr> + <td>7</td> + <td>70</td> + <td>70</td> + <td>[DRIVER RECOVERS]<br />40, 50, 60, 70</td> + </tr> + <tr> + <td>8</td> + <td>80</td> + <td>80</td> + <td>80</td> + </tr> + <tr> + <td>9</td> + <td>90</td> + <td>90</td> + <td>90</td> + </tr> + <tr> + <td>10</td> + <td>100</td> + <td>100</td> + <td>100</td> + </tr> +</table> -* Recovery from master faults +If the driver had crashed in the middle of the processing of time 3, then it will process time 3 and output 30 after recovery. -* Current state and future directions
\ No newline at end of file +# Where to Go from Here +* Documentation - [Scala and Java](api/streaming/index.html) +* More examples - [Scala](https://github.com/mesos/spark/tree/master/examples/src/main/scala/spark/streaming/examples) and [Java](https://github.com/mesos/spark/tree/master/examples/src/main/java/spark/streaming/examples)
\ No newline at end of file |