You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@flink.apache.org by se...@apache.org on 2016/12/14 14:10:36 UTC
[2/4] flink git commit: [FLINK-5258] [docs] Reorganize the docs to
improve navigation and reduce duplication
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/internals/add_operator.md
----------------------------------------------------------------------
diff --git a/docs/internals/add_operator.md b/docs/internals/add_operator.md
deleted file mode 100644
index c7e26c1..0000000
--- a/docs/internals/add_operator.md
+++ /dev/null
@@ -1,253 +0,0 @@
----
-title: "How to add a new Operator"
-nav-title: "How-To: Add an Operator"
-nav-parent_id: internals
-nav-pos: 6
----
-<!--
-Licensed to the Apache Software Foundation (ASF) under one
-or more contributor license agreements. See the NOTICE file
-distributed with this work for additional information
-regarding copyright ownership. The ASF licenses this file
-to you under the Apache License, Version 2.0 (the
-"License"); you may not use this file except in compliance
-with the License. You may obtain a copy of the License at
-
- http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing,
-software distributed under the License is distributed on an
-"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
-KIND, either express or implied. See the License for the
-specific language governing permissions and limitations
-under the License.
--->
-
-Operators in the Java API can be added in multiple different ways:
-
-1. On the DataSet, as a specialization/combination of existing operators
-2. As a custom extension operator
-3. As a new runtime operator
-
-The first two approaches are typically more lightweight and easier to implement. Sometimes,
-new functionality does require a new runtime operator, or it is much more efficient to
-
-* This will be replaced by the TOC
-{:toc}
-
-## Implementing a new Operator on DataSet
-
-Many operators can be implemented as a specialization of another operator, or by means of a UDF.
-
-The simplest example are the `sum()`, `min()`, and `max()` functions on the
-{% gh_link /flink-java/src/main/java/org/apache/flink/api/java/DataSet.java "DataSet" %}.
-These functions simply call other operations with some pre-defined parameters:
-
-~~~java
-public AggregateOperator<T> sum (int field) {
- return this.aggregate (Aggregations.SUM, field);
-}
-
-~~~
-
-Some operations can be implemented as compositions of multiple other operators. An example is to implement a
-*count()* function through a combination of *map* and *aggregate*.
-
-A simple way to do this is to define a function on the {% gh_link /flink-java/src/main/java/org/apache/flink/api/java/DataSet.java "DataSet" %} that calls *map(...)* and *reduce(...)* in turn:
-
-~~~java
-public DataSet<Long> count() {
- return this.map(new MapFunction<T, Long>() {
- public Long map(T value) {
- return 1L;
- }
- })
- .reduce(new ReduceFunction<Long>() {
- public Long reduce(Long val1, Long val1) {
- return val1 + val2;
- }
- });
-}
-~~~
-
-To define a new operator without altering the DataSet class is possible by putting the functions as static members
-into another class. The example of the *count()* operator would look the following way:
-
-~~~java
-public static <T>DataSet<Long> count(DataSet<T> data) {
- return data.map(...).reduce(...);
-}
-~~~
-
-### More Complex Operators
-
-A more complex example of an operation via specialization is the {% gh_link /flink-java/src/main/java/org/apache/flink/api/java/operators/AggregateOperator.java "Aggregation Operation" %} in the Java API. It is implemented by means of a *GroupReduce* UDF.
-
-The Aggregate Operation comes with its own operator in the *Java API*, but translates itself into a {% gh_link /flink-core/src/main/java/org/apache/flink/api/common/operators/base/GroupReduceOperatorBase.java "GroupReduceOperatorBase" %} in the *Common API*.
-The Java API aggregation operator is only a builder that takes the types of aggregations and the field positions, and used that information to
-parameterize the GroupReduce UDF that performs the aggregations.
-
-Because the operation is translated into a GroupReduce operation, it appears as a GroupReduceOperator in the optimizer and runtime.
-
-
-
-## Implementing a Custom Extension Operator
-
-The DataSet offers a method for custom operators: `DataSet<X> runOperation(CustomUnaryOperation<T, X> operation)`.
-The *CustomUnaryOperation* interface defines operators by means of the two functions:
-
-~~~ java
-
-void setInput(DataSet<IN> inputData);
-
-DataSet<OUT> createResult();
-~~~
-
-The {% gh_link /flink-libraries/flink-gelly/src/main/java/org/apache/flink/graph/pregel/VertexCentricIteration.java "VertexCentricIteration" %} operator is implemented that way. Below is an example how to implement the *count()* operator that way.
-
-~~~ java
-public class Counter<T> implements CustomUnaryOperation<T, Long> {
-
- private DataSet<T> input;
-
- public void setInput(DataSet<IN> inputData) { this.input = inputData; }
-
- public DataSet<Long> createResult() {
- return input.map(...).reduce(...);
- }
-}
-~~~
-
-The CountOperator can be called in the following way:
-
-~~~ java
-DataSet<String> lines = ...;
-DataSet<Long> count = lines.runOperation(new Counter<String>());
-~~~
-
-
-## Implementing a new Runtime Operator
-
-Adding an new runtime operator requires changes throughout the entire stack, from the API to the runtime:
-
-- *Java API*
-- *Common API*
-- *Optimizer*
-- *Runtime*
-
-We start the description bottom up, at the example of the *mapPartition()* function, which is like a *map*
-function, but invoked only once per parallel partition.
-
-**Runtime**
-
-Runtime Operators are implemented using the {% gh_link /flink-runtime/src/main/java/org/apache/flink/runtime/operators/Driver.java "Driver" %} interface. The interface defines the methods that describe the operator towards the runtime. The {% gh_link /flink-runtime/src/main/java/org/apache/flink/runtime/operators/MapDriver.java "MapDriver" %} serves as a simple example of how those operators work.
-
-The runtime works with the `MutableObjectIterator`, which describes data streams with the ability to reuse objects, to reduce pressure on the garbage collector.
-
-An implementation of the central `run()` method for the *mapPartition* operator could look the following way:
-
-~~~ java
-public void run() throws Exception {
- final MutableObjectIterator<IN> input = this.taskContext.getInput(0);
- final MapPartitionFunction<IN, OUT> function = this.taskContext.getStub();
- final Collector<OUT> output = this.taskContext.getOutputCollector();
- final TypeSerializer<IN> serializer = this.taskContext.getInputSerializer(0);
-
- // we assume that the UDF takes a java.util.Iterator, so we wrap the MutableObjectIterator
- Iterator<IN> iterator = new MutableToRegularIteratorWrapper(input, serializer);
-
- function.mapPartition(iterator, output);
-}
-~~~
-
-To increase efficiency, it is often beneficial to implement a *chained* version of an operator. Chained
-operators run in the same thread as their preceding operator, and work with nested function calls.
-This is very efficient, because it saves serialization/deserialization overhead.
-
-To learn how to implement a chained operator, take a look at the {% gh_link /flink-runtime/src/main/java/org/apache/flink/runtime/operators/MapDriver.java "MapDriver" %} (regular) and the
-{% gh_link /flink-runtime/src/main/java/org/apache/flink/runtime/operators/chaining/ChainedMapDriver.java "ChainedMapDriver" %} (chained variant).
-
-
-### Optimizer/Compiler
-
-This section does a minimal discussion of the important steps to add an operator. Please see the {% gh_link /flink-optimizer/src/main/java/org/apache/flink/optimizer/Optimizer.java "Optimizer" %} for more details on how the optimizer works.
-To allow the optimizer to include a new operator in its planning, it needs a bit of information about it; in particular, the following information:
-
-- *{% gh_link /flink-runtime/src/main/java/org/apache/flink/runtime/operators/DriverStrategy.java "DriverStrategy" %}*: The operation needs to be added to the Enum, to make it available to the optimizer. The parameters to the Enum entry define which class implements the runtime operator, its chained version, whether the operator accumulates records (and needs memory for that), and whether it requires a comparator (works on keys). For our example, we can add the entry
-~~~ java
-MAP_PARTITION(MapPartitionDriver.class, null /* or chained variant */, PIPELINED, false)
-~~~
-
-- *Cost function*: The class {% gh_link /flink-optimizer/src/main/java/org/apache/flink/optimizer/costs/CostEstimator.java "CostEstimator" %} needs to know how expensive the operation is to the system. The costs here refer to the non-UDF part of the operator. Since the operator does essentially no work (it forwards the record stream to the UDF), the costs are zero. We change the `costOperator(...)` method by adding the *MAP_PARTITION* constant to the switch statement similar to the *MAP* constant such that no cost is accounted for it.
-
-- *OperatorDescriptor*: The operator descriptors define how an operation needs to be treated by the optimizer. They describe how the operation requires the input data to be (e.g., sorted or partitioned) and that way allows the optimizer to optimize the data movement, sorting, grouping in a global fashion. They do that by describing which {% gh_link /flink-optimizer/src/main/java/org/apache/flink/optimizer/dataproperties/RequestedGlobalProperties.java "RequestedGlobalProperties" %} (partitioning, replication, etc) and which {% gh_link /flink-optimizer/src/main/java/org/apache/flink/optimizer/dataproperties/RequestedLocalProperties.java "RequestedLocalProperties" %} (sorting, grouping, uniqueness) the operator has, as well as how the operator affects the existing {% gh_link /flink-optimizer/src/main/java/org/apache/flink/optimizer/dataproperties/GlobalProperties.java "GlobalProperties" %} and {% gh_link /flink-optimizer/src/main/java/org/apache/flink/optimizer/dataproperties/LocalProp
erties.java "LocalProperties" %}. In addition, it defines a few utility methods, for example to instantiate an operator candidate.
-Since the *mapPartition()* function is very simple (no requirements on partitioning/grouping), the descriptor is very simple. Other operators have more complex requirements, for example the {% gh_link /flink-optimizer/src/main/java/org/apache/flink/optimizer/operators/HashJoinBuildFirstProperties.java "Hash Join 1" %}, {% gh_link /flink-optimizer/src/main/java/org/apache/flink/optimizer/operators/HashJoinBuildSecondProperties.java "Hash Join 2" %}, {% gh_link flink-optimizer/src/main/java/org/apache/flink/optimizer/operators/AbstractSortMergeJoinDescriptor.java "SortMerge Join" %}).
-The code sample below explains (with comments) how to create a descriptor for the *MapPartitionOperator*
-
- ~~~ java
- public DriverStrategy getStrategy() {
- return MAP_PARTITION;
- }
-
- // Instantiate the operator with the strategy over the input given in the form of the Channel
- public SingleInputPlanNode instantiate(Channel in, SingleInputNode node) {
- return new SingleInputPlanNode(node, "MapPartition", in, MAP_PARTITION);
- }
-
- // The operation accepts data with default global properties (arbitrary distribution)
- protected List<RequestedGlobalProperties> createPossibleGlobalProperties() {
- return Collections.singletonList(new RequestedGlobalProperties());
- }
-
- // The operation can accept data with any local properties. No grouping/sorting is necessary
- protected List<RequestedLocalProperties> createPossibleLocalProperties() {
- return Collections.singletonList(new RequestedLocalProperties());
- }
-
- // the operation itself does not affect the existing global properties.
- // The effect of the UDF's semantics// are evaluated separately (by interpreting the
- // semantic assertions)
- public GlobalProperties computeGlobalProperties(GlobalProperties gProps) {
- return gProps;
- }
-
- // since the operation can mess up all order, grouping, uniqueness, we cannot make any statements
- // about how local properties are preserved
- public LocalProperties computeLocalProperties(LocalProperties lProps) {
- return LocalProperties.EMPTY;
- }
- ~~~
-
-- *OptimizerNode*: The optimizer node is the place where all comes together. It creates the list of *OperatorDescriptors*, implements the result data set size estimation, and assigns a name to the operation. It is a relatively small class and can be more or less copied again from the {% gh_link /flink-optimizer/src/main/java/org/apache/flink/optimizer/dag/MapNode.java "MapNode" %}.
-
-
-### Common API
-
-To make the operation available to the higher-level APIs, it needs to be added to the Common API. The simplest way to do this is to add a
-base operator. Create a class `MapPartitionOperatorBase`, after the pattern of the {% gh_link /flink-core/src/main/java/org/apache/flink/api/common/operators/base/MapOperatorBase.java "MapOperatorBase" %}.
-
-In addition, the optimizer needs to know how to create an OptimizerNode from the OperatorBase. This happens in the class
-`GraphCreatingVisitor` in the {% gh_link /flink-optimizer/src/main/java/org/apache/flink/optimizer/Optimizer.java "Optimizer" %}.
-
-*Note:* A pending idea is to allow to skip this step by unifying the OptimizerNode and the Common API operator. They essentially fulfill the
-same function. The Common API operator exists only in order for the `flink-java` and `flink-scala` packages to not have a dependency on the
-optimizer.
-
-
-### Java API
-
-Create a Java API operator that is constructed in the same way as the {% gh_link /flink-java/src/main/java/org/apache/flink/api/java/operators/MapOperator.java "MapOperator" %}. The core method is the `translateToDataFlow(...)` method, which creates the Common API operator for the Java API operator.
-
-The final step is to add a function to the `DataSet` class:
-
-~~~ java
-public <R> DataSet<R> mapPartition(MapPartitionFunction<T, R> function) {
- return new MapPartitionOperator<T, R>(this, function);
-}
-~~~
-
----
-
-*This documentation is maintained by the contributors of the individual components.
-We kindly ask anyone that adds and changes components to eventually provide a patch
-or pull request that updates these documents as well.*
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/internals/components.md
----------------------------------------------------------------------
diff --git a/docs/internals/components.md b/docs/internals/components.md
new file mode 100644
index 0000000..b750ffd
--- /dev/null
+++ b/docs/internals/components.md
@@ -0,0 +1,59 @@
+---
+title: "Component Stack"
+nav-parent_id: internals
+nav-pos: 1
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+As a software stack, Flink is a layered system. The different layers of the stack build on
+top of each other and raise the abstraction level of the program representations they accept:
+
+- The **runtime** layer receives a program in the form of a *JobGraph*. A JobGraph is a generic parallel
+data flow with arbitrary tasks that consume and produce data streams.
+
+- Both the **DataStream API** and the **DataSet API** generate JobGraphs through separate compilation
+processes. The DataSet API uses an optimizer to determine the optimal plan for the program, while
+the DataStream API uses a stream builder.
+
+- The JobGraph is executed according to a variety of deployment options available in Flink (e.g., local,
+remote, YARN, etc)
+
+- Libraries and APIs that are bundled with Flink generate DataSet or DataStream API programs. These are
+Table for queries on logical tables, FlinkML for Machine Learning, and Gelly for graph processing.
+
+You can click on the components in the figure to learn more.
+
+<center>
+ <img src="{{ site.baseurl }}/fig/stack.png" width="700px" alt="Apache Flink: Stack" usemap="#overview-stack">
+</center>
+
+<map name="overview-stack">
+<area id="lib-datastream-cep" title="CEP: Complex Event Processing" href="{{ site.baseurl }}/dev/libs/cep.html" shape="rect" coords="63,0,143,177" />
+<area id="lib-datastream-table" title="Table: Relational DataStreams" href="{{ site.baseurl }}/dev/table_api.html" shape="rect" coords="143,0,223,177" />
+<area id="lib-dataset-ml" title="FlinkML: Machine Learning" href="{{ site.baseurl }}/dev/libs/ml/index.html" shape="rect" coords="382,2,462,176" />
+<area id="lib-dataset-gelly" title="Gelly: Graph Processing" href="{{ site.baseurl }}/dev/libs/gelly/index.html" shape="rect" coords="461,0,541,177" />
+<area id="lib-dataset-table" title="Table API and SQL" href="{{ site.baseurl }}/dev/table_api.html" shape="rect" coords="544,0,624,177" />
+<area id="datastream" title="DataStream API" href="{{ site.baseurl }}/dev/datastream_api.html" shape="rect" coords="64,177,379,255" />
+<area id="dataset" title="DataSet API" href="{{ site.baseurl }}/dev/batch/index.html" shape="rect" coords="382,177,697,255" />
+<area id="runtime" title="Runtime" href="{{ site.baseurl }}/concepts/runtime.html" shape="rect" coords="63,257,700,335" />
+<area id="local" title="Local" href="{{ site.baseurl }}/quickstart/setup_quickstart.html" shape="rect" coords="62,337,275,414" />
+<area id="cluster" title="Cluster" href="{{ site.baseurl }}/setup/cluster_setup.html" shape="rect" coords="273,336,486,413" />
+<area id="cloud" title="Cloud" href="{{ site.baseurl }}/setup/gce_setup.html" shape="rect" coords="485,336,700,414" />
+</map>
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/internals/general_arch.md
----------------------------------------------------------------------
diff --git a/docs/internals/general_arch.md b/docs/internals/general_arch.md
deleted file mode 100644
index e3e8c4d..0000000
--- a/docs/internals/general_arch.md
+++ /dev/null
@@ -1,101 +0,0 @@
----
-title: "General Architecture and Process Model"
-nav-title: Architecture and Process Model
-nav-parent_id: internals
-nav-pos: 2
----
-<!--
-Licensed to the Apache Software Foundation (ASF) under one
-or more contributor license agreements. See the NOTICE file
-distributed with this work for additional information
-regarding copyright ownership. The ASF licenses this file
-to you under the Apache License, Version 2.0 (the
-"License"); you may not use this file except in compliance
-with the License. You may obtain a copy of the License at
-
- http://www.apache.org/licenses/LICENSE-2.0
-
-Unless required by applicable law or agreed to in writing,
-software distributed under the License is distributed on an
-"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
-KIND, either express or implied. See the License for the
-specific language governing permissions and limitations
-under the License.
--->
-
-* This will be replaced by the TOC
-{:toc}
-
-## The Processes
-
-When the Flink system is started, it bring up the *JobManager* and one or more *TaskManagers*. The JobManager
-is the coordinator of the Flink system, while the TaskManagers are the workers that execute parts of the
-parallel programs. When starting the system in *local* mode, a single JobManager and TaskManager are brought
-up within the same JVM.
-
-When a program is submitted, a client is created that performs the pre-processing and turns the program
-into the parallel data flow form that is executed by the JobManager and TaskManagers. The figure below
-illustrates the different actors in the system and their interactions.
-
-<div style="text-align: center;">
-<img src="{{ site.baseurl }}/fig/process_model.svg" width="100%" alt="Flink Process Model">
-</div>
-
-## Component Stack
-
-As a software stack, Flink is a layered system. The different layers of the stack build on
-top of each other and raise the abstraction level of the program representations they accept:
-
-- The **runtime** layer receives a program in the form of a *JobGraph*. A JobGraph is a generic parallel
-data flow with arbitrary tasks that consume and produce data streams.
-
-- Both the **DataStream API** and the **DataSet API** generate JobGraphs through separate compilation
-processes. The DataSet API uses an optimizer to determine the optimal plan for the program, while
-the DataStream API uses a stream builder.
-
-- The JobGraph is executed according to a variety of deployment options available in Flink (e.g., local,
-remote, YARN, etc)
-
-- Libraries and APIs that are bundled with Flink generate DataSet or DataStream API programs. These are
-Table for queries on logical tables, FlinkML for Machine Learning, and Gelly for graph processing.
-
-You can click on the components in the figure to learn more.
-
-<center>
- <img src="{{ site.baseurl }}/fig/stack.png" width="700px" alt="Apache Flink: Stack" usemap="#overview-stack">
-</center>
-
-<map name="overview-stack">
-<area id="lib-datastream-cep" title="CEP: Complex Event Processing" href="{{ site.baseurl }}/dev/libs/cep.html" shape="rect" coords="63,0,143,177" />
-<area id="lib-datastream-table" title="Table: Relational DataStreams" href="{{ site.baseurl }}/dev/table_api.html" shape="rect" coords="143,0,223,177" />
-<area id="lib-dataset-ml" title="FlinkML: Machine Learning" href="{{ site.baseurl }}/dev/libs/ml/index.html" shape="rect" coords="382,2,462,176" />
-<area id="lib-dataset-gelly" title="Gelly: Graph Processing" href="{{ site.baseurl }}/dev/libs/gelly/index.html" shape="rect" coords="461,0,541,177" />
-<area id="lib-dataset-table" title="Table API and SQL" href="{{ site.baseurl }}/dev/table_api.html" shape="rect" coords="544,0,624,177" />
-<area id="datastream" title="DataStream API" href="{{ site.baseurl }}/dev/datastream_api.html" shape="rect" coords="64,177,379,255" />
-<area id="dataset" title="DataSet API" href="{{ site.baseurl }}/dev/batch/index.html" shape="rect" coords="382,177,697,255" />
-<area id="runtime" title="Runtime" href="{{ site.baseurl }}/internals/general_arch.html" shape="rect" coords="63,257,700,335" />
-<area id="local" title="Local" href="{{ site.baseurl }}/setup/local_setup.html" shape="rect" coords="62,337,275,414" />
-<area id="cluster" title="Cluster" href="{{ site.baseurl }}/setup/cluster_setup.html" shape="rect" coords="273,336,486,413" />
-<area id="cloud" title="Cloud" href="{{ site.baseurl }}/setup/gce_setup.html" shape="rect" coords="485,336,700,414" />
-</map>
-
-## Projects and Dependencies
-
-The Flink system code is divided into multiple sub-projects. The goal is to reduce the number of
-dependencies that a project implementing a Flink program needs, as well as to faciltate easier testing
-of smaller sub-modules.
-
-The individual projects and their dependencies are shown in the figure below.
-
-<div style="text-align: center;">
-<img src="{{ site.baseurl }}/fig/projects_dependencies.svg" alt="The Flink sub-projects and their dependencies" height="600px" style="text-align: center;"/>
-</div>
-
-In addition to the projects listed in the figure above, Flink currently contains the following sub-projects:
-
-- `flink-dist`: The *distribution* project. It defines how to assemble the compiled code, scripts, and other resources
-into the final folder structure that is ready to use.
-
-- `flink-quickstart`: Scripts, maven archetypes, and example programs for the quickstarts and tutorials.
-
-- `flink-contrib`: A series of projects that are in an early version and useful tools contributed by users. The code for the latter is maintained mainly by external contributors. The requirements for code being accepted into `flink-contrib` are lower compared to the rest of the code.
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/internals/ide_setup.md
----------------------------------------------------------------------
diff --git a/docs/internals/ide_setup.md b/docs/internals/ide_setup.md
index bd931dd..1c514d0 100644
--- a/docs/internals/ide_setup.md
+++ b/docs/internals/ide_setup.md
@@ -1,7 +1,7 @@
---
title: "IDE Setup"
-nav-parent_id: internals
-nav-pos: 1
+nav-parent_id: start
+nav-pos: 3
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
@@ -25,6 +25,32 @@ under the License.
* Replaced by the TOC
{:toc}
+## IntelliJ IDEA
+
+A brief guide on how to set up IntelliJ IDEA IDE for development of the Flink core.
+As Eclipse is known to have issues with mixed Scala and Java projects, more and more contributers are migrating to IntelliJ IDEA.
+
+The following documentation describes the steps to setup IntelliJ IDEA 14.0.3 (https://www.jetbrains.com/idea/download/) with the Flink sources.
+
+Prior to doing anything, make sure that the Flink project is built at least once from the terminal:
+`mvn clean package -DskipTests`
+
+### Installing the Scala plugin
+1. Go to IntelliJ plugins settings (File -> Settings -> Plugins) and click on "Install Jetbrains plugin...".
+2. Select and install the "Scala" plugin.
+3. Restart IntelliJ
+
+### Installing the Scala compiler plugin
+1. Go to IntelliJ scala compiler settings (File -> Settings -> Build, Execution, Deployment -> Compiler -> Scala Compiler) and click on "Install Jetbrains plugin...".
+2. Click on the green plus icon on the right to add a compiler plugin
+3. Point to the paradise jar: ~/.m2/repository/org/scalamacros/paradise_2.10.4/2.0.1/paradise_2.10.4-2.0.1.jar If there is no such file, this means that you should build Flink from the terminal as explained above.
+
+### Importing Flink
+1. Start IntelliJ IDEA and choose "Import Project"
+2. Select the root folder of the Flink repository
+3. Choose "Import project from external model" and select "Maven"
+4. Leave the default options and finish the import.
+
## Eclipse
A brief guide how to set up Eclipse for development of the Flink core.
@@ -101,28 +127,3 @@ Uncheck the box labeled "Resolve dependencies from Workspace projects", click "A
See [this post](http://stackoverflow.com/questions/25391207/how-do-i-add-execution-environment-1-8-to-eclipse-luna)
for details.
-## IntelliJ IDEA
-
-A brief guide on how to set up IntelliJ IDEA IDE for development of the Flink core.
-As Eclipse is known to have issues with mixed Scala and Java projects, more and more contributers are migrating to IntelliJ IDEA.
-
-The following documentation describes the steps to setup IntelliJ IDEA 14.0.3 (https://www.jetbrains.com/idea/download/) with the Flink sources.
-
-Prior to doing anything, make sure that the Flink project is built at least once from the terminal:
-`mvn clean package -DskipTests`
-
-### Installing the Scala plugin
-1. Go to IntelliJ plugins settings (File -> Settings -> Plugins) and click on "Install Jetbrains plugin...".
-2. Select and install the "Scala" plugin.
-3. Restart IntelliJ
-
-### Installing the Scala compiler plugin
-1. Go to IntelliJ scala compiler settings (File -> Settings -> Build, Execution, Deployment -> Compiler -> Scala Compiler) and click on "Install Jetbrains plugin...".
-2. Click on the green plus icon on the right to add a compiler plugin
-3. Point to the paradise jar: ~/.m2/repository/org/scalamacros/paradise_2.10.4/2.0.1/paradise_2.10.4-2.0.1.jar If there is no such file, this means that you should build Flink from the terminal as explained above.
-
-### Importing Flink
-1. Start IntelliJ IDEA and choose "Import Project"
-2. Select the root folder of the Flink repository
-3. Choose "Import project from external model" and select "Maven"
-4. Leave the default options and finish the import.
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/internals/index.md
----------------------------------------------------------------------
diff --git a/docs/internals/index.md b/docs/internals/index.md
index fd0af4a..8b3184c 100644
--- a/docs/internals/index.md
+++ b/docs/internals/index.md
@@ -1,9 +1,10 @@
---
title: "Internals"
nav-id: internals
-nav-pos: 5
-nav-title: '<i class="fa fa-book" aria-hidden="true"></i> Internals'
+nav-pos: 8
+nav-title: '<i class="fa fa-book title dessert" aria-hidden="true"></i> Internals'
nav-parent_id: root
+section-break: true
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/internals/state_backends.md
----------------------------------------------------------------------
diff --git a/docs/internals/state_backends.md b/docs/internals/state_backends.md
index e9f9fd8..f6a4cc7 100644
--- a/docs/internals/state_backends.md
+++ b/docs/internals/state_backends.md
@@ -38,21 +38,21 @@ There are two basic state backends: `Keyed State` and `Operator State`.
#### Keyed State
-*Keyed State* is always relative to keys and can only be used in functions and operator on a `KeyedStream`.
+*Keyed State* is always relative to keys and can only be used in functions and operators on a `KeyedStream`.
Examples of keyed state are the `ValueState` or `ListState` that one can create in a function on a `KeyedStream`, as
-well at the state of a keyed window operator.
+well as the state of a keyed window operator.
-Keyed State is organized in so called *Key Groups*. Key Groups are the unit in which keyed state can be redistributed and
+Keyed State is organized in so called *Key Groups*. Key Groups are the unit by which keyed state can be redistributed and
there are as many key groups as the defined maximum parallelism.
-During execution, each parallel instance of an operator gets one or more key groups.
+During execution each parallel instance of an operator gets one or more key groups.
#### Operator State
-*Operator State* is state per parallel subtask. It subsume the `Checkpointed` interface in Flink 1.0 and Flink 1.1.
+*Operator State* is state per parallel subtask. It subsumes the `Checkpointed` interface in Flink 1.0 and Flink 1.1.
The new `CheckpointedFunction` interface is basically a shortcut (syntactic sugar) for the Operator State.
Operator State needs special re-distribution schemes when parallelism is changed. There can be different variations of such
-schemes, out of which the following are currently defined:
+schemes; the following are currently defined:
- **List-style redistribution:** Each operator returns a List of state elements. The whole state is logically a concatenation of
all lists. On restore/redistribution, the list is evenly divided into as many sublists as there are parallel operators.
@@ -63,22 +63,21 @@ schemes, out of which the following are currently defined:
*Keyed State* and *Operator State* exist in two forms: *managed* and *raw*.
-*Managed State* is represented in data structured controlled by the Flink runtime, such as internal hash tables, or RocksDB.
-Examples are the "ValueState", "ListState", etc. Flink's runtime encodes the states and writes them into the checkpoints.
+*Managed State* is represented in data structures controlled by the Flink runtime, such as internal hash tables, or RocksDB.
+Examples are "ValueState", "ListState", etc. Flink's runtime encodes the states and writes them into the checkpoints.
-*Raw State* is state that users and operators keep in their own data structures. Upon checkpoints, they only write a sequence or bytes into
+*Raw State* is state that users and operators keep in their own data structures. When checkpointed, they only write a sequence of bytes into
the checkpoint. Flink knows nothing about the state's data structures and sees only the raw bytes.
## Checkpointing Procedure
-When operator snapshots are takes, there are two parts: The **synchronous** and the **asynchronous** part.
+When operator snapshots are taken, there are two parts: the **synchronous** and the **asynchronous** parts.
-Operators and state backends provide their snapshots as a Java `FutureTask`. That task contains the state where tte *synchronous* part
+Operators and state backends provide their snapshots as a Java `FutureTask`. That task contains the state where the *synchronous* part
is completed and the *asynchronous* part is pending. The asynchronous part is then executed by a background thread for that checkpoint.
-Operators that checkpoint purely synchronous return an already completed `FutureTask`.
+Operators that checkpoint purely synchronously return an already completed `FutureTask`.
If an asynchronous operation needs to be performed, it is executed in the `run()` method of that `FutureTask`.
-The tasks are canceleable, in order to release streams and other resource consuming handles.
-
+The tasks are cancelable, in order to release streams and other resource consuming handles.
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/monitoring/best_practices.md
----------------------------------------------------------------------
diff --git a/docs/monitoring/best_practices.md b/docs/monitoring/best_practices.md
index 6a58a78..0bd362e 100644
--- a/docs/monitoring/best_practices.md
+++ b/docs/monitoring/best_practices.md
@@ -1,7 +1,7 @@
---
title: "Best Practices"
-nav-parent_id: monitoring
-nav-pos: 5
+nav-parent_id: dev
+nav-pos: 90
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
@@ -180,96 +180,6 @@ public static class CustomType extends Tuple11<String, String, ..., String> {
}
~~~
-
-## Register a custom serializer for your Flink program
-
-If you use a custom type in your Flink program which cannot be serialized by the
-Flink type serializer, Flink falls back to using the generic Kryo
-serializer. You may register your own serializer or a serialization system like
-Google Protobuf or Apache Thrift with Kryo. To do that, simply register the type
-class and the serializer in the `ExecutionConfig` of your Flink program.
-
-
-{% highlight java %}
-final ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
-
-// register the class of the serializer as serializer for a type
-env.getConfig().registerTypeWithKryoSerializer(MyCustomType.class, MyCustomSerializer.class);
-
-// register an instance as serializer for a type
-MySerializer mySerializer = new MySerializer();
-env.getConfig().registerTypeWithKryoSerializer(MyCustomType.class, mySerializer);
-{% endhighlight %}
-
-Note that your custom serializer has to extend Kryo's Serializer class. In the
-case of Google Protobuf or Apache Thrift, this has already been done for
-you:
-
-{% highlight java %}
-
-final ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
-
-// register the Google Protobuf serializer with Kryo
-env.getConfig().registerTypeWithKryoSerializer(MyCustomType.class, ProtobufSerializer.class);
-
-// register the serializer included with Apache Thrift as the standard serializer
-// TBaseSerializer states it should be initialized as a default Kryo serializer
-env.getConfig().addDefaultKryoSerializer(MyCustomType.class, TBaseSerializer.class);
-
-{% endhighlight %}
-
-For the above example to work, you need to include the necessary dependencies in
-your Maven project file (pom.xml). In the dependency section, add the following
-for Apache Thrift:
-
-{% highlight xml %}
-
-<dependency>
- <groupId>com.twitter</groupId>
- <artifactId>chill-thrift</artifactId>
- <version>0.5.2</version>
-</dependency>
-<!-- libthrift is required by chill-thrift -->
-<dependency>
- <groupId>org.apache.thrift</groupId>
- <artifactId>libthrift</artifactId>
- <version>0.6.1</version>
- <exclusions>
- <exclusion>
- <groupId>javax.servlet</groupId>
- <artifactId>servlet-api</artifactId>
- </exclusion>
- <exclusion>
- <groupId>org.apache.httpcomponents</groupId>
- <artifactId>httpclient</artifactId>
- </exclusion>
- </exclusions>
-</dependency>
-
-{% endhighlight %}
-
-For Google Protobuf you need the following Maven dependency:
-
-{% highlight xml %}
-
-<dependency>
- <groupId>com.twitter</groupId>
- <artifactId>chill-protobuf</artifactId>
- <version>0.5.2</version>
-</dependency>
-<!-- We need protobuf for chill-protobuf -->
-<dependency>
- <groupId>com.google.protobuf</groupId>
- <artifactId>protobuf-java</artifactId>
- <version>2.5.0</version>
-</dependency>
-
-{% endhighlight %}
-
-
-Please adjust the versions of both libraries as needed.
-
-
## Using Logback instead of Log4j
**Note: This tutorial is applicable starting from Flink 0.10**
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/monitoring/index.md
----------------------------------------------------------------------
diff --git a/docs/monitoring/index.md b/docs/monitoring/index.md
index c2582db..36399fd 100644
--- a/docs/monitoring/index.md
+++ b/docs/monitoring/index.md
@@ -1,9 +1,9 @@
---
-title: "Monitoring & Debugging"
-nav-title: '<i class="fa fa-bug" aria-hidden="true"></i> Monitoring & Debugging'
+title: "Debugging and Monitoring"
+nav-title: '<i class="fa fa-bug title maindish" aria-hidden="true"></i> Debugging & Monitoring'
nav-id: monitoring
nav-parent_id: root
-nav-pos: 4
+nav-pos: 7
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/page/css/flink.css
----------------------------------------------------------------------
diff --git a/docs/page/css/flink.css b/docs/page/css/flink.css
index 5ad0bc0..b6d4947 100644
--- a/docs/page/css/flink.css
+++ b/docs/page/css/flink.css
@@ -192,6 +192,27 @@ svg.graph text {
Side navigation
=============================================================================*/
+#sidenav a i.title {
+ min-width: 21px;
+}
+
+#sidenav a i.title.appetizer {
+ color: #FBB142;
+}
+
+#sidenav a i.title.maindish {
+ color: #7E4F89;
+}
+
+#sidenav a i.title.dessert {
+ color: #E6526F;
+}
+
+#sidenav hr.section-break {
+ margin-bottom: 5px;
+ border-top: 0px;
+}
+
#sidenav, #sidenav ul {
list-style: none;
display: block;
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/quickstart/java_api_quickstart.md
----------------------------------------------------------------------
diff --git a/docs/quickstart/java_api_quickstart.md b/docs/quickstart/java_api_quickstart.md
index 1c7bd8a..c8f21d8 100644
--- a/docs/quickstart/java_api_quickstart.md
+++ b/docs/quickstart/java_api_quickstart.md
@@ -1,8 +1,8 @@
---
-title: "Quickstart: Java API"
-nav-title: Java API
-nav-parent_id: quickstarts
-nav-pos: 3
+title: "Sample Project using the Java API"
+nav-title: Sample Project in Java
+nav-parent_id: start
+nav-pos: 0
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
@@ -64,8 +64,7 @@ There will be a new directory in your working directory. If you've used the _cur
The sample project is a __Maven project__, which contains four classes. _StreamingJob_ and _BatchJob_ are basic skeleton programs, _SocketTextStreamWordCount_ is a working streaming example and _WordCountJob_ is a working batch example. Please note that the _main_ method of all classes allow you to start Flink in a development/testing mode.
-We recommend you __import this project into your IDE__ to develop and test it. If you use Eclipse, the [m2e plugin](http://www.eclipse.org/m2e/) allows to [import Maven projects](http://books.sonatype.com/m2eclipse-book/reference/creating-sect-importing-projects.html#fig-creating-import). Some Eclipse bundles include that plugin by default, others require you to install it manually. The IntelliJ IDE also supports Maven projects out of the box.
-
+We recommend you [import this project into your IDE]({{ site.baseurl }}/internals/ide_setup) to develop and test it.
A note to Mac OS X users: The default JVM heapsize for Java is too small for Flink. You have to manually increase it. Choose "Run Configurations" -> Arguments and write into the "VM Arguments" box: "-Xmx800m" in Eclipse.
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/quickstart/run_example_quickstart.md
----------------------------------------------------------------------
diff --git a/docs/quickstart/run_example_quickstart.md b/docs/quickstart/run_example_quickstart.md
index a64e87f..9ce4c84 100644
--- a/docs/quickstart/run_example_quickstart.md
+++ b/docs/quickstart/run_example_quickstart.md
@@ -1,8 +1,8 @@
---
title: "Monitoring the Wikipedia Edit Stream"
nav-title: Monitoring Wikipedia Edits
-nav-parent_id: quickstarts
-nav-pos: 2
+nav-parent_id: examples
+nav-pos: 10
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
@@ -278,8 +278,8 @@ was produced.
This should get you started with writing your own Flink programs. To learn more
you can check out our guides
-about [basic concepts]({{ site.baseurl }}/apis/common/index.html) and the
-[DataStream API]({{ site.baseurl }}/apis/streaming/index.html). Stick
+about [basic concepts]({{ site.baseurl }}/dev/api_concepts) and the
+[DataStream API]({{ site.baseurl }}/dev/datastream_api). Stick
around for the bonus exercise if you want to learn about setting up a Flink cluster on
your own machine and writing results to [Kafka](http://kafka.apache.org).
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/quickstart/scala_api_quickstart.md
----------------------------------------------------------------------
diff --git a/docs/quickstart/scala_api_quickstart.md b/docs/quickstart/scala_api_quickstart.md
index 22562f4..388f0a7 100644
--- a/docs/quickstart/scala_api_quickstart.md
+++ b/docs/quickstart/scala_api_quickstart.md
@@ -1,8 +1,8 @@
---
-title: "Quickstart: Scala API"
-nav-title: Scala API
-nav-parent_id: quickstarts
-nav-pos: 4
+title: "Sample Project using the Scala API"
+nav-title: Sample Project in Scala
+nav-parent_id: start
+nav-pos: 1
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/quickstart/setup_quickstart.md
----------------------------------------------------------------------
diff --git a/docs/quickstart/setup_quickstart.md b/docs/quickstart/setup_quickstart.md
index 3dd0950..6aab686 100644
--- a/docs/quickstart/setup_quickstart.md
+++ b/docs/quickstart/setup_quickstart.md
@@ -1,8 +1,8 @@
---
-title: "Quickstart: Setup"
-nav-title: Setup
-nav-parent_id: quickstarts
-nav-pos: 1
+title: "Quickstart"
+nav-title: '<i class="fa fa-power-off title appetizer" aria-hidden="true"></i> Quickstart'
+nav-parent_id: root
+nav-pos: 2
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
@@ -28,13 +28,29 @@ under the License.
Get a Flink example program up and running in a few simple steps.
-## Setup: Download and Start
+## Setup: Download and Start Flink
-Flink runs on __Linux, Mac OS X, and Windows__. To be able to run Flink, the only requirement is to have a working __Java 7.x__ (or higher) installation. Windows users, please take a look at the [Flink on Windows]({{ site.baseurl }}/setup/local_setup.html#flink-on-windows) guide which describes how to run Flink on Windows for local setups.
+Flink runs on __Linux, Mac OS X, and Windows__. To be able to run Flink, the only requirement is to have a working __Java 7.x__ (or higher) installation. Windows users, please take a look at the [Flink on Windows]({{ site.baseurl }}/setup/flink_on_windows) guide which describes how to run Flink on Windows for local setups.
+
+You can check the correct installation of Java by issuing the following command:
+
+~~~bash
+java -version
+~~~
+
+If you have Java 8, the output will look something like this:
+
+~~~bash
+java version "1.8.0_111"
+Java(TM) SE Runtime Environment (build 1.8.0_111-b14)
+Java HotSpot(TM) 64-Bit Server VM (build 25.111-b14, mixed mode)
+~~~
### Download
-Download a binary from the [downloads page](http://flink.apache.org/downloads.html). You can pick any Hadoop/Scala combination you like.
+Download a binary from the [downloads page](http://flink.apache.org/downloads.html). You can pick
+any Hadoop/Scala combination you like. If you plan to just use the local file system, any Hadoop
+version will work fine.
### Start a Local Flink Cluster
@@ -53,9 +69,135 @@ Check the __JobManager's web frontend__ at [http://localhost:8081](http://localh
<a href="{{ site.baseurl }}/page/img/quickstart-setup/jobmanager-1.png" ><img class="img-responsive" src="{{ site.baseurl }}/page/img/quickstart-setup/jobmanager-1.png" alt="JobManager: Overview"/></a>
-## Run Example
+You can also verify that the system is running by checking the log files in the `logs` directory:
+
+~~~bash
+$ tail log/flink-*-jobmanager-*.log
+INFO ... - Starting JobManager
+INFO ... - Starting JobManager web frontend
+INFO ... - Web frontend listening at 127.0.0.1:8081
+INFO ... - Registered TaskManager at 127.0.0.1 (akka://flink/user/taskmanager)
+~~~
-Now, we are going to run the [SocketWindowWordCount example](https://github.com/apache/flink/blob/master/flink-examples/flink-examples-streaming/src/main/java/org/apache/flink/streaming/examples/socket/SocketWindowWordCount.java) and read text from a socket and count the number of distinct words.
+## Read the Code
+
+You can find the complete source code for this SocketWindowWordCount example in [scala](https://github.com/apache/flink/blob/master/flink-examples/flink-examples-streaming/src/main/scala/org/apache/flink/streaming/scala/examples/socket/SocketWindowWordCount.scala) and [java](https://github.com/apache/flink/blob/master/flink-examples/flink-examples-streaming/src/main/java/org/apache/flink/streaming/examples/socket/SocketWindowWordCount.java) on GitHub.
+
+<div class="codetabs" markdown="1">
+<div data-lang="scala" markdown="1">
+{% highlight scala %}
+object SocketWindowWordCount {
+
+ def main(args: Array[String]) : Unit = {
+
+ // the port to connect to
+ val port: Int = try {
+ ParameterTool.fromArgs(args).getInt("port")
+ } catch {
+ case e: Exception => {
+ System.err.println("No port specified. Please run 'SocketWindowWordCount --port <port>'")
+ return
+ }
+ }
+
+ // get the execution environment
+ val env: StreamExecutionEnvironment = StreamExecutionEnvironment.getExecutionEnvironment
+
+ // get input data by connecting to the socket
+ val text = env.socketTextStream("localhost", port, '\n')
+
+ // parse the data, group it, window it, and aggregate the counts
+ val windowCounts = text
+ .flatMap { w => w.split("\\s") }
+ .map { w => WordWithCount(w, 1) }
+ .keyBy("word")
+ .timeWindow(Time.seconds(5), Time.seconds(1))
+ .sum("count")
+
+ // print the results with a single thread, rather than in parallel
+ windowCounts.print().setParallelism(1)
+
+ env.execute("Socket Window WordCount")
+ }
+
+ // Data type for words with count
+ case class WordWithCount(word: String, count: Long)
+}
+{% endhighlight %}
+</div>
+<div data-lang="java" markdown="1">
+{% highlight java %}
+public class SocketWindowWordCount {
+
+ public static void main(String[] args) throws Exception {
+
+ // the port to connect to
+ final int port;
+ try {
+ final ParameterTool params = ParameterTool.fromArgs(args);
+ port = params.getInt("port");
+ } catch (Exception e) {
+ System.err.println("No port specified. Please run 'SocketWindowWordCount --port <port>'");
+ return;
+ }
+
+ // get the execution environment
+ final StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
+
+ // get input data by connecting to the socket
+ DataStream<String> text = env.socketTextStream("localhost", port, "\n");
+
+ // parse the data, group it, window it, and aggregate the counts
+ DataStream<WordWithCount> windowCounts = text
+ .flatMap(new FlatMapFunction<String, WordWithCount>() {
+ @Override
+ public void flatMap(String value, Collector<WordWithCount> out) {
+ for (String word : value.split("\\s")) {
+ out.collect(new WordWithCount(word, 1L));
+ }
+ }
+ })
+ .keyBy("word")
+ .timeWindow(Time.seconds(5), Time.seconds(1))
+ .reduce(new ReduceFunction<WordWithCount>() {
+ @Override
+ public WordWithCount reduce(WordWithCount a, WordWithCount b) {
+ return new WordWithCount(a.word, a.count + b.count);
+ }
+ });
+
+ // print the results with a single thread, rather than in parallel
+ windowCounts.print().setParallelism(1);
+
+ env.execute("Socket Window WordCount");
+ }
+
+ // Data type for words with count
+ public static class WordWithCount {
+
+ public String word;
+ public long count;
+
+ public WordWithCount() {}
+
+ public WordWithCount(String word, long count) {
+ this.word = word;
+ this.count = count;
+ }
+
+ @Override
+ public String toString() {
+ return word + " : " + count;
+ }
+ }
+}
+{% endhighlight %}
+</div>
+</div>
+
+## Run the Example
+
+Now, we are going to run this Flink application. It will read text from a socket and once a second print the number of occurances of each distinct word during the previous 5 seconds.
* First of all, we use **netcat** to start local server via
@@ -119,67 +261,4 @@ Now, we are going to run the [SocketWindowWordCount example](https://github.com/
## Next Steps
-Check out the [step-by-step example](run_example_quickstart.html) in order to get a first feel of Flink's programming APIs. When you are done with that, go ahead and read the [streaming guide]({{ site.baseurl }}/dev/datastream_api.html).
-
-### Cluster Setup
-
-__Running Flink on a cluster__ is as easy as running it locally. Having __passwordless SSH__ and
-__the same directory structure__ on all your cluster nodes lets you use our scripts to control
-everything.
-
-1. Copy the unpacked __flink__ directory from the downloaded archive to the same file system path
-on each node of your setup.
-2. Choose a __master node__ (JobManager) and set the `jobmanager.rpc.address` key in
-`conf/flink-conf.yaml` to its IP or hostname. Make sure that all nodes in your cluster have the same
-`jobmanager.rpc.address` configured.
-3. Add the IPs or hostnames (one per line) of all __worker nodes__ (TaskManager) to the slaves files
-in `conf/slaves`.
-
-You can now __start the cluster__ at your master node with `bin/start-cluster.sh`.
-
-The following __example__ illustrates the setup with three nodes (with IP addresses from _10.0.0.1_
-to _10.0.0.3_ and hostnames _master_, _worker1_, _worker2_) and shows the contents of the
-configuration files, which need to be accessible at the same path on all machines:
-
-<div class="row">
- <div class="col-md-6 text-center">
- <img src="{{ site.baseurl }}/page/img/quickstart_cluster.png" style="width: 85%">
- </div>
-<div class="col-md-6">
- <div class="row">
- <p class="lead text-center">
- /path/to/<strong>flink/conf/<br>flink-conf.yaml</strong>
- <pre>jobmanager.rpc.address: 10.0.0.1</pre>
- </p>
- </div>
-<div class="row" style="margin-top: 1em;">
- <p class="lead text-center">
- /path/to/<strong>flink/<br>conf/slaves</strong>
- <pre>
-10.0.0.2
-10.0.0.3</pre>
- </p>
-</div>
-</div>
-</div>
-
-Have a look at the [Configuration]({{ site.baseurl }}/setup/config.html) section of the documentation to see other available configuration options.
-For Flink to run efficiently, a few configuration values need to be set.
-
-In particular,
-
- * the amount of available memory per TaskManager (`taskmanager.heap.mb`),
- * the number of available CPUs per machine (`taskmanager.numberOfTaskSlots`),
- * the total number of CPUs in the cluster (`parallelism.default`) and
- * the temporary directories (`taskmanager.tmp.dirs`)
-
-
-are very important configuration values.
-
-### Flink on YARN
-
-You can easily deploy Flink on your existing __YARN cluster__.
-
-1. Download the __Flink Hadoop2 package__ from the [downloads page](https://flink.apache.org/downloads.html)
-2. Make sure your __HADOOP_HOME__ (or _YARN_CONF_DIR_ or _HADOOP_CONF_DIR_) __environment variable__ is set to read your YARN and HDFS configuration.
-3. Run the __YARN client__ with: `./bin/yarn-session.sh`. You can run the client with options `-n 10 -tm 8192` to allocate 10 TaskManagers with 8GB of memory each.
+Check out some more [examples]({{ site.baseurl }}/examples) to get a better feel for Flink's programming APIs. When you are done with that, go ahead and read the [streaming guide]({{ site.baseurl }}/dev/datastream_api.html).
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/redirects/concepts.md
----------------------------------------------------------------------
diff --git a/docs/redirects/concepts.md b/docs/redirects/concepts.md
index 82790e8..5a96bf4 100644
--- a/docs/redirects/concepts.md
+++ b/docs/redirects/concepts.md
@@ -1,7 +1,7 @@
---
title: "Concepts"
layout: redirect
-redirect: /concepts/index.html
+redirect: /concepts/programming-model.html
permalink: /concepts/concepts.html
---
<!--
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/setup/building.md
----------------------------------------------------------------------
diff --git a/docs/setup/building.md b/docs/setup/building.md
index 8c60997..8a71fc6 100644
--- a/docs/setup/building.md
+++ b/docs/setup/building.md
@@ -1,7 +1,7 @@
---
-title: Build Flink from Source
-nav-parent_id: setup
-nav-pos: 1
+title: Building Flink from Source
+nav-parent_id: start
+nav-pos: 20
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/setup/checkpoints.md
----------------------------------------------------------------------
diff --git a/docs/setup/checkpoints.md b/docs/setup/checkpoints.md
new file mode 100644
index 0000000..ea2f406
--- /dev/null
+++ b/docs/setup/checkpoints.md
@@ -0,0 +1,57 @@
+---
+title: "Checkpoints"
+nav-parent_id: setup
+nav-pos: 7
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+
+* toc
+{:toc}
+
+## Overview
+
+TBD
+
+
+## Externalized Checkpoints
+
+You can configure periodic checkpoints to be persisted externally. Externalized checkpoints write their meta data out to persistent storage and are *not* automatically cleaned up when the job fails. This way, you will have a checkpoint around to resume from if your job fails.
+
+```java
+CheckpoingConfig config = env.getCheckpointConfig();
+config.enableExternalizedCheckpoints(ExternalizedCheckpointCleanup.RETAIN_ON_CANCELLATION);
+```
+
+The `ExternalizedCheckpointCleanup` mode configures what happens with externalized checkpoints when you cancel the job:
+
+- **`ExternalizedCheckpointCleanup.RETAIN_ON_CANCELLATION`**: Retain the externalized checkpoint when the job is cancelled. Note that you have to manually clean up the checkpoint state after cancellation in this case.
+
+- **`ExternalizedCheckpointCleanup.DELETE_ON_CANCELLATION`**: Delete the externalized checkpoint when the job is cancelled. The checkpoint state will only be available if the job fails.
+
+The **target directory** for the checkpoint is determined from the default checkpoint directory configuration. This is configured via the configuration key `state.checkpoints.dir`, which should point to the desired target directory:
+
+```
+state.checkpoints.dir: hdfs:///checkpoints/
+```
+
+This directory will then contain the checkpoint meta data required to restore the checkpoint. The actual checkpoint files will still be available in their configured directory. You currently can only set this via the configuration files.
+
+Follow the [savepoint guide]({{ site.baseurl }}/setup/cli.html#savepoints) when you want to resume from a specific checkpoint.
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/setup/cluster_setup.md
----------------------------------------------------------------------
diff --git a/docs/setup/cluster_setup.md b/docs/setup/cluster_setup.md
index 77e1a3c..7d3684f 100644
--- a/docs/setup/cluster_setup.md
+++ b/docs/setup/cluster_setup.md
@@ -39,6 +39,10 @@ Flink runs on all *UNIX-like environments*, e.g. **Linux**, **Mac OS X**, and **
If your cluster does not fulfill these software requirements you will need to install/upgrade it.
+Having __passwordless SSH__ and
+__the same directory structure__ on all your cluster nodes will allow you to use our scripts to control
+everything.
+
{% top %}
### `JAVA_HOME` Configuration
@@ -51,7 +55,7 @@ You can set this variable in `conf/flink-conf.yaml` via the `env.java.home` key.
## Flink Setup
-Go to the [downloads page]({{ site.download_url }}) and get the ready to run package. Make sure to pick the Flink package **matching your Hadoop version**. If you don't plan to use Hadoop, pick any version.
+Go to the [downloads page]({{ site.download_url }}) and get the ready-to-run package. Make sure to pick the Flink package **matching your Hadoop version**. If you don't plan to use Hadoop, pick any version.
After downloading the latest release, copy the archive to your master node and extract it:
@@ -64,29 +68,45 @@ cd flink-*
After having extracted the system files, you need to configure Flink for the cluster by editing *conf/flink-conf.yaml*.
-Set the `jobmanager.rpc.address` key to point to your master node. Furthermode define the maximum amount of main memory the JVM is allowed to allocate on each node by setting the `jobmanager.heap.mb` and `taskmanager.heap.mb` keys.
-
-The value is given in MB. If some worker nodes have more main memory which you want to allocate to the Flink system you can overwrite the default value by setting an environment variable `FLINK_TM_HEAP` on the respective node.
-
-Finally you must provide a list of all nodes in your cluster which shall be used as worker nodes. Therefore, similar to the HDFS configuration, edit the file *conf/slaves* and enter the IP/host name of each worker node. Each worker node will later run a TaskManager.
-
-Each entry must be separated by a new line, as in the following example:
-
-~~~
-192.168.0.100
-192.168.0.101
-.
-.
-.
-192.168.0.150
-~~~
-
-The Flink directory must be available on every worker under the same path. You can use a shared NSF directory, or copy the entire Flink directory to every worker node.
+Set the `jobmanager.rpc.address` key to point to your master node. You should also define the maximum amount of main memory the JVM is allowed to allocate on each node by setting the `jobmanager.heap.mb` and `taskmanager.heap.mb` keys.
+
+These values are given in MB. If some worker nodes have more main memory which you want to allocate to the Flink system you can overwrite the default value by setting the environment variable `FLINK_TM_HEAP` on those specific nodes.
+
+Finally, you must provide a list of all nodes in your cluster which shall be used as worker nodes. Therefore, similar to the HDFS configuration, edit the file *conf/slaves* and enter the IP/host name of each worker node. Each worker node will later run a TaskManager.
+
+The following example illustrates the setup with three nodes (with IP addresses from _10.0.0.1_
+to _10.0.0.3_ and hostnames _master_, _worker1_, _worker2_) and shows the contents of the
+configuration files (which need to be accessible at the same path on all machines):
+
+<div class="row">
+ <div class="col-md-6 text-center">
+ <img src="{{ site.baseurl }}/page/img/quickstart_cluster.png" style="width: 60%">
+ </div>
+<div class="col-md-6">
+ <div class="row">
+ <p class="lead text-center">
+ /path/to/<strong>flink/conf/<br>flink-conf.yaml</strong>
+ <pre>jobmanager.rpc.address: 10.0.0.1</pre>
+ </p>
+ </div>
+<div class="row" style="margin-top: 1em;">
+ <p class="lead text-center">
+ /path/to/<strong>flink/<br>conf/slaves</strong>
+ <pre>
+10.0.0.2
+10.0.0.3</pre>
+ </p>
+</div>
+</div>
+</div>
+
+The Flink directory must be available on every worker under the same path. You can use a shared NFS directory, or copy the entire Flink directory to every worker node.
Please see the [configuration page](config.html) for details and additional configuration options.
In particular,
+ * the amount of available memory per JobManager (`jobmanager.heap.mb`),
* the amount of available memory per TaskManager (`taskmanager.heap.mb`),
* the number of available CPUs per machine (`taskmanager.numberOfTaskSlots`),
* the total number of CPUs in the cluster (`parallelism.default`) and
@@ -126,6 +146,6 @@ bin/jobmanager.sh (start cluster)|stop|stop-all
bin/taskmanager.sh start|stop|stop-all
~~~
-Make sure to call these scripts on the hosts, on which you want to start/stop the respective instance.
+Make sure to call these scripts on the hosts on which you want to start/stop the respective instance.
{% top %}
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/setup/config.md
----------------------------------------------------------------------
diff --git a/docs/setup/config.md b/docs/setup/config.md
index c24065a..89e8207 100644
--- a/docs/setup/config.md
+++ b/docs/setup/config.md
@@ -1,7 +1,8 @@
---
title: "Configuration"
+nav-id: "config"
nav-parent_id: setup
-nav-pos: 3
+nav-pos: 1
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
@@ -22,7 +23,9 @@ specific language governing permissions and limitations
under the License.
-->
-The default configuration parameters allow Flink to run out-of-the-box in single node setups.
+**For single-node setups Flink is ready to go out of the box and you don't need to change the default configuration to get started.**
+
+The out of the box configuration will use your default Java installation. You can manually set the environment variable `JAVA_HOME` or the configuration key `env.java.home` in `conf/flink-conf.yaml` if you want to manually override the Java runtime to use.
This page lists the most common options that are typically needed to set up a well performing (distributed) installation. In addition a full list of all available configuration parameters is listed here.
@@ -56,7 +59,7 @@ The configuration files for the TaskManagers can be different, Flink does not as
- `taskmanager.numberOfTaskSlots`: The number of parallel operator or user function instances that a single TaskManager can run (DEFAULT: 1). If this value is larger than 1, a single TaskManager takes multiple instances of a function or operator. That way, the TaskManager can utilize multiple CPU cores, but at the same time, the available memory is divided between the different operator or function instances. This value is typically proportional to the number of physical CPU cores that the TaskManager's machine has (e.g., equal to the number of cores, or half the number of cores). [More about task slots](config.html#configuring-taskmanager-processing-slots).
- `parallelism.default`: The default parallelism to use for programs that have no parallelism specified. (DEFAULT: 1). For setups that have no concurrent jobs running, setting this value to NumTaskManagers * NumSlotsPerTaskManager will cause the system to use all available execution resources for the program's execution. **Note**: The default parallelism can be overwriten for an entire job by calling `setParallelism(int parallelism)` on the `ExecutionEnvironment` or by passing `-p <parallelism>` to the Flink Command-line frontend. It can be overwritten for single transformations by calling `setParallelism(int
-parallelism)` on an operator. See the [Basic API Concepts]({{site.baseurl}}/dev/api_concepts.html#parallel-execution) for more information about the parallelism.
+parallelism)` on an operator. See [Parallel Execution]({{site.baseurl}}/dev/parallel) for more information about parallelism.
- `fs.default-scheme`: The default filesystem scheme to be used, with the necessary authority to contact, e.g. the host:port of the NameNode in the case of HDFS (if needed).
By default, this is set to `file:///` which points to the local filesystem. This means that the local
@@ -534,6 +537,6 @@ Flink executes a program in parallel by splitting it into subtasks and schedulin
Each Flink TaskManager provides processing slots in the cluster. The number of slots is typically proportional to the number of available CPU cores __of each__ TaskManager. As a general recommendation, the number of available CPU cores is a good default for `taskmanager.numberOfTaskSlots`.
-When starting a Flink application, users can supply the default number of slots to use for that job. The command line value therefore is called `-p` (for parallelism). In addition, it is possible to [set the number of slots in the programming APIs]({{site.baseurl}}/dev/api_concepts.html#parallel-execution) for the whole application and individual operators.
+When starting a Flink application, users can supply the default number of slots to use for that job. The command line value therefore is called `-p` (for parallelism). In addition, it is possible to [set the number of slots in the programming APIs]({{site.baseurl}}/dev/parallel) for the whole application and for individual operators.
<img src="{{ site.baseurl }}/fig/slots_parallelism.svg" class="img-responsive" />
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/setup/fault_tolerance.md
----------------------------------------------------------------------
diff --git a/docs/setup/fault_tolerance.md b/docs/setup/fault_tolerance.md
index cef746e..500e91a 100644
--- a/docs/setup/fault_tolerance.md
+++ b/docs/setup/fault_tolerance.md
@@ -1,8 +1,7 @@
---
-title: "Failure & Recovery Model"
-nav-id: fault_tolerance
-nav-parent_id: setup
-nav-pos: 4
+title: "Restart Strategies"
+nav-parent_id: execution
+nav-pos: 50
---
<!--
Licensed to the Apache Software Foundation (ASF) under one
@@ -23,222 +22,6 @@ specific language governing permissions and limitations
under the License.
-->
-Flink's fault tolerance mechanism recovers programs in the presence of failures and
-continues to execute them. Such failures include machine hardware failures, network failures,
-transient program failures, etc.
-
-* This will be replaced by the TOC
-{:toc}
-
-
-Streaming Fault Tolerance
--------------------------
-
-Flink has a checkpointing mechanism that recovers streaming jobs after failures. The checkpointing mechanism requires a *persistent* (or *durable*) source that
-can be asked for prior records again (Apache Kafka is a good example of such a source).
-
-The checkpointing mechanism stores the progress in the data sources and data sinks, the state of windows, as well as the user-defined state (see [Working with State]({{ site.baseurl }}/dev/state.html)) consistently to provide *exactly once* processing semantics. Where the checkpoints are stored (e.g., JobManager memory, file system, database) depends on the configured [state backend]({{ site.baseurl }}/dev/state_backends.html).
-
-The [docs on streaming fault tolerance]({{ site.baseurl }}/internals/stream_checkpointing.html) describe in detail the technique behind Flink's streaming fault tolerance mechanism.
-
-To enable checkpointing, call `enableCheckpointing(n)` on the `StreamExecutionEnvironment`, where *n* is the checkpoint interval in milliseconds.
-
-Other parameters for checkpointing include:
-
-- *exactly-once vs. at-least-once*: You can optionally pass a mode to the `enableCheckpointing(n)` method to choose between the two guarantee levels.
- Exactly-once is preferrable for most applications. At-least-once may be relevant for certain super-low-latency (consistently few milliseconds) applications.
-
-- *number of concurrent checkpoints*: By default, the system will not trigger another checkpoint while one is still in progress. This ensures that the topology does not spend too much time on checkpoints and not make progress with processing the streams. It is possible to allow for multiple overlapping checkpoints, which is interesting for pipelines that have a certain processing delay (for example because the functions call external services that need some time to respond) but that still want to do very frequent checkpoints (100s of milliseconds) to re-process very little upon failures.
-
-- *checkpoint timeout*: The time after which a checkpoint-in-progress is aborted, if it did not complete by then.
-
-<div class="codetabs" markdown="1">
-<div data-lang="java" markdown="1">
-{% highlight java %}
-StreamExecutionEnvironment env = StreamExecutionEnvironment.getExecutionEnvironment();
-
-// start a checkpoint every 1000 ms
-env.enableCheckpointing(1000);
-
-// advanced options:
-
-// set mode to exactly-once (this is the default)
-env.getCheckpointConfig().setCheckpointingMode(CheckpointingMode.EXACTLY_ONCE);
-
-// checkpoints have to complete within one minute, or are discarded
-env.getCheckpointConfig().setCheckpointTimeout(60000);
-
-// allow only one checkpoint to be in progress at the same time
-env.getCheckpointConfig().setMaxConcurrentCheckpoints(1);
-{% endhighlight %}
-</div>
-<div data-lang="scala" markdown="1">
-{% highlight scala %}
-val env = StreamExecutionEnvironment.getExecutionEnvironment()
-
-// start a checkpoint every 1000 ms
-env.enableCheckpointing(1000)
-
-// advanced options:
-
-// set mode to exactly-once (this is the default)
-env.getCheckpointConfig.setCheckpointingMode(CheckpointingMode.EXACTLY_ONCE)
-
-// checkpoints have to complete within one minute, or are discarded
-env.getCheckpointConfig.setCheckpointTimeout(60000)
-
-// allow only one checkpoint to be in progress at the same time
-env.getCheckpointConfig.setMaxConcurrentCheckpoints(1)
-{% endhighlight %}
-</div>
-</div>
-
-{% top %}
-
-### Externalized Checkpoints
-
-You can configure periodic checkpoints to be persisted externally. Externalized checkpoints write their meta data out to persistent storage and are *not* automatically cleaned up when the job fails. This way, you will have a checkpoint around to resume from if your job fails.
-
-```java
-CheckpoingConfig config = env.getCheckpointConfig();
-config.enableExternalizedCheckpoints(ExternalizedCheckpointCleanup.RETAIN_ON_CANCELLATION);
-```
-
-The `ExternalizedCheckpointCleanup` mode configures what happens with externalized checkpoints when you cancel the job:
-
-- **`ExternalizedCheckpointCleanup.RETAIN_ON_CANCELLATION`**: Retain the externalized checkpoint when the job is cancelled. Note that you have to manually clean up the checkpoint state after cancellation in this case.
-
-- **`ExternalizedCheckpointCleanup.DELETE_ON_CANCELLATION`**: Delete the externalized checkpoint when the job is cancelled. The checkpoint state will only be available if the job fails.
-
-The **target directory** for the checkpoint is determined from the default checkpoint directory configuration. This is configured via the configuration key `state.checkpoints.dir`, which should point to the desired target directory:
-
-```
-state.checkpoints.dir: hdfs:///checkpoints/
-```
-
-This directory will then contain the checkpoint meta data required to restore the checkpoint. The actual checkpoint files will still be available in their configured directory. You currently can only set this via the configuration files.
-
-Follow the [savepoint guide]({{ site.baseurl }}/setup/cli.html#savepoints) when you want to resume from a specific checkpoint.
-
-### Fault Tolerance Guarantees of Data Sources and Sinks
-
-Flink can guarantee exactly-once state updates to user-defined state only when the source participates in the
-snapshotting mechanism. The following table lists the state update guarantees of Flink coupled with the bundled connectors.
-
-Please read the documentation of each connector to understand the details of the fault tolerance guarantees.
-
-<table class="table table-bordered">
- <thead>
- <tr>
- <th class="text-left" style="width: 25%">Source</th>
- <th class="text-left" style="width: 25%">Guarantees</th>
- <th class="text-left">Notes</th>
- </tr>
- </thead>
- <tbody>
- <tr>
- <td>Apache Kafka</td>
- <td>exactly once</td>
- <td>Use the appropriate Kafka connector for your version</td>
- </tr>
- <tr>
- <td>AWS Kinesis Streams</td>
- <td>exactly once</td>
- <td></td>
- </tr>
- <tr>
- <td>RabbitMQ</td>
- <td>at most once (v 0.10) / exactly once (v 1.0) </td>
- <td></td>
- </tr>
- <tr>
- <td>Twitter Streaming API</td>
- <td>at most once</td>
- <td></td>
- </tr>
- <tr>
- <td>Collections</td>
- <td>exactly once</td>
- <td></td>
- </tr>
- <tr>
- <td>Files</td>
- <td>exactly once</td>
- <td></td>
- </tr>
- <tr>
- <td>Sockets</td>
- <td>at most once</td>
- <td></td>
- </tr>
- </tbody>
-</table>
-
-To guarantee end-to-end exactly-once record delivery (in addition to exactly-once state semantics), the data sink needs
-to take part in the checkpointing mechanism. The following table lists the delivery guarantees (assuming exactly-once
-state updates) of Flink coupled with bundled sinks:
-
-<table class="table table-bordered">
- <thead>
- <tr>
- <th class="text-left" style="width: 25%">Sink</th>
- <th class="text-left" style="width: 25%">Guarantees</th>
- <th class="text-left">Notes</th>
- </tr>
- </thead>
- <tbody>
- <tr>
- <td>HDFS rolling sink</td>
- <td>exactly once</td>
- <td>Implementation depends on Hadoop version</td>
- </tr>
- <tr>
- <td>Elasticsearch</td>
- <td>at least once</td>
- <td></td>
- </tr>
- <tr>
- <td>Kafka producer</td>
- <td>at least once</td>
- <td></td>
- </tr>
- <tr>
- <td>Cassandra sink</td>
- <td>at least once / exactly once</td>
- <td>exactly once only for idempotent updates</td>
- </tr>
- <tr>
- <td>AWS Kinesis Streams</td>
- <td>at least once</td>
- <td></td>
- </tr>
- <tr>
- <td>File sinks</td>
- <td>at least once</td>
- <td></td>
- </tr>
- <tr>
- <td>Socket sinks</td>
- <td>at least once</td>
- <td></td>
- </tr>
- <tr>
- <td>Standard output</td>
- <td>at least once</td>
- <td></td>
- </tr>
- <tr>
- <td>Redis sink</td>
- <td>at least once</td>
- <td></td>
- </tr>
- </tbody>
-</table>
-
-{% top %}
-
-## Restart Strategies
-
Flink supports different restart strategies which control how the jobs are restarted in case of a failure.
The cluster can be started with a default restart strategy which is always used when no job specific restart strategy has been defined.
In case that the job is submitted with a restart strategy, this strategy overrides the cluster's default setting.
@@ -246,7 +29,6 @@ In case that the job is submitted with a restart strategy, this strategy overrid
The default restart strategy is set via Flink's configuration file `flink-conf.yaml`.
The configuration parameter *restart-strategy* defines which strategy is taken.
Per default, the no-restart strategy is used.
-When checkpointing is activated and no restart strategy has been configured, the job will be restarted infinitely often.
See the following list of available restart strategies to learn what values are supported.
Each restart strategy comes with its own set of parameters which control its behaviour.
@@ -306,7 +88,7 @@ env.setRestartStrategy(RestartStrategies.fixedDelayRestart(
{% top %}
-### Fixed Delay Restart Strategy
+## Fixed Delay Restart Strategy
The fixed delay restart strategy attempts a given number of times to restart the job.
If the maximum number of attempts is exceeded, the job eventually fails.
@@ -329,12 +111,12 @@ restart-strategy: fixed-delay
<tbody>
<tr>
<td><it>restart-strategy.fixed-delay.attempts</it></td>
- <td>The number of times that Flink retries the execution before the job is declared as failed</td>
+ <td>Number of restart attempts</td>
<td>1</td>
</tr>
<tr>
<td><it>restart-strategy.fixed-delay.delay</it></td>
- <td>Delay between two consecutive restart attempts. Delaying the retry means that after a failed execution, the re-execution does not start immediately, but only after a certain delay. Delaying the retries can be helpful when the program interacts with external systems where for example connections or pending transactions should reach a timeout before re-execution is attempted.</td>
+ <td>Delay between two consecutive restart attempts</td>
<td><it>akka.ask.timeout</it></td>
</tr>
</tbody>
@@ -368,9 +150,23 @@ env.setRestartStrategy(RestartStrategies.fixedDelayRestart(
</div>
</div>
+### Restart Attempts
+
+The number of times that Flink retries the execution before the job is declared as failed is configurable via the *restart-strategy.fixed-delay.attempts* parameter.
+
+The default value is **1**.
+
+### Retry Delays
+
+Execution retries can be configured to be delayed. Delaying the retry means that after a failed execution, the re-execution does not start immediately, but only after a certain delay.
+
+Delaying the retries can be helpful when the program interacts with external systems where for example connections or pending transactions should reach a timeout before re-execution is attempted.
+
+The default value is the value of *akka.ask.timeout*.
+
{% top %}
-### Failure Rate Restart Strategy
+## Failure Rate Restart Strategy
The failure rate restart strategy restarts job after failure, but when `failure rate` (failures per time interval) is exceeded, the job eventually fails.
In-between two consecutive restart attempts, the restart strategy waits a fixed amount of time.
@@ -441,7 +237,7 @@ env.setRestartStrategy(RestartStrategies.failureRateRestart(
{% top %}
-### No Restart Strategy
+## No Restart Strategy
The job fails directly and no restart is attempted.
@@ -466,7 +262,7 @@ env.setRestartStrategy(RestartStrategies.noRestart())
</div>
</div>
-### Fallback Restart Strategy
+## Fallback Restart Strategy
The cluster defined restart strategy is used.
This helpful for streaming programs which enable checkpointing.
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/setup/flink_on_windows.md
----------------------------------------------------------------------
diff --git a/docs/setup/flink_on_windows.md b/docs/setup/flink_on_windows.md
new file mode 100644
index 0000000..2cbc163
--- /dev/null
+++ b/docs/setup/flink_on_windows.md
@@ -0,0 +1,86 @@
+---
+title: "Running Flink on Windows"
+nav-parent_id: start
+nav-pos: 12
+---
+<!--
+Licensed to the Apache Software Foundation (ASF) under one
+or more contributor license agreements. See the NOTICE file
+distributed with this work for additional information
+regarding copyright ownership. The ASF licenses this file
+to you under the Apache License, Version 2.0 (the
+"License"); you may not use this file except in compliance
+with the License. You may obtain a copy of the License at
+
+ http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing,
+software distributed under the License is distributed on an
+"AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+KIND, either express or implied. See the License for the
+specific language governing permissions and limitations
+under the License.
+-->
+
+If you want to run Flink locally on a Windows machine you need to [download](http://flink.apache.org/downloads.html) and unpack the binary Flink distribution. After that you can either use the **Windows Batch** file (`.bat`), or use **Cygwin** to run the Flink Jobmanager.
+
+## Starting with Windows Batch Files
+
+To start Flink in local mode from the *Windows Batch*, open the command window, navigate to the `bin/` directory of Flink and run `start-local.bat`.
+
+Note: The ``bin`` folder of your Java Runtime Environment must be included in Window's ``%PATH%`` variable. Follow this [guide](http://www.java.com/en/download/help/path.xml) to add Java to the ``%PATH%`` variable.
+
+~~~bash
+$ cd flink
+$ cd bin
+$ start-local.bat
+Starting Flink job manager. Web interface by default on http://localhost:8081/.
+Do not close this batch window. Stop job manager by pressing Ctrl+C.
+~~~
+
+After that, you need to open a second terminal to run jobs using `flink.bat`.
+
+{% top %}
+
+## Starting with Cygwin and Unix Scripts
+
+With *Cygwin* you need to start the Cygwin Terminal, navigate to your Flink directory and run the `start-local.sh` script:
+
+~~~bash
+$ cd flink
+$ bin/start-local.sh
+Starting jobmanager.
+~~~
+
+{% top %}
+
+## Installing Flink from Git
+
+If you are installing Flink from the git repository and you are using the Windows git shell, Cygwin can produce a failure similiar to this one:
+
+~~~bash
+c:/flink/bin/start-local.sh: line 30: $'\r': command not found
+~~~
+
+This error occurs because git is automatically transforming UNIX line endings to Windows style line endings when running in Windows. The problem is that Cygwin can only deal with UNIX style line endings. The solution is to adjust the Cygwin settings to deal with the correct line endings by following these three steps:
+
+1. Start a Cygwin shell.
+
+2. Determine your home directory by entering
+
+ ~~~bash
+ cd; pwd
+ ~~~
+
+ This will return a path under the Cygwin root path.
+
+3. Using NotePad, WordPad or a different text editor open the file `.bash_profile` in the home directory and append the following: (If the file does not exist you will have to create it)
+
+~~~bash
+export SHELLOPTS
+set -o igncr
+~~~
+
+Save the file and open a new bash shell.
+
+{% top %}
http://git-wip-us.apache.org/repos/asf/flink/blob/79d7e301/docs/setup/index.md
----------------------------------------------------------------------
diff --git a/docs/setup/index.md b/docs/setup/index.md
index bf34fe0..bc8a9dd 100644
--- a/docs/setup/index.md
+++ b/docs/setup/index.md
@@ -1,9 +1,9 @@
---
-title: "Setup & Operations"
+title: "Deployment & Operations"
nav-id: "setup"
-nav-title: '<i class="fa fa-cogs" aria-hidden="true"></i> Setup & Operations'
+nav-title: '<i class="fa fa-sliders title maindish" aria-hidden="true"></i> Deployment & Operations'
nav-parent_id: root
-nav-pos: 2
+nav-pos: 6
---
<!--
Licensed to the Apache Software Foundation (ASF) under one