Skip to content

Commit

Permalink
[FLINK-3575] [docs] Update 'Working With State' section in docs
Browse files Browse the repository at this point in the history 
This closes apache#1760
  • Loading branch information
@aljoscha @StephanEwen
aljoscha authored and StephanEwen committed Mar 4, 2016
1 parent b10bd92 commit e650d13
Showing 1 changed file with 157 additions and 62 deletions.
219 changes: 157 additions & 62 deletions docs/apis/streaming/state.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -26,110 +26,139 @@ under the License.

All transformations in Flink may look like functions (in the functional processing terminology), but
are in fact stateful operators. You can make *every* transformation (`map`, `filter`, etc) stateful
by declaring local variables or using Flink's state interface. You can register any local variable
by using Flink's state interface or checkpointing instance fields of your function. You can register
any instance field
as ***managed*** state by implementing an interface. In this case, and also in the case of using
Flink's native state interface, Flink will automatically take consistent snapshots of your state
periodically, and restore its value in the case of a failure.

The end effect is that updates to any form of state are the same under failure-free execution and
execution under failures.

First, we look at how to make local variables consistent under failures, and then we look at
First, we look at how to make instance fields consistent under failures, and then we look at
Flink's state interface.

By default state checkpoints will be stored in-memory at the JobManager. For proper persistence of large
state, Flink supports storing the checkpoints on file systems (HDFS, S3, or any mounted POSIX file system),
which can be configured in the `flink-conf.yaml` or via `StreamExecutionEnvironment.setStateBackend(…)`.
See [state backends]({{ site.baseurl }}/apis/streaming/state_backends.html) for information
about the available state backends and how to configure them.

* ToC
{:toc}

## Using the Key/Value State Interface

## Checkpointing Local Variables
The Key/Value state interface provides access to different types of state that are all scoped to
the key of the current input element. This means that this type of state can only be used
on a `KeyedStream`, which can be created via `stream.keyBy(…)`.

Local variables can be checkpointed by using the `Checkpointed` interface.
Now, we will first look at the different types of state available and then we will see
how they can be used in a program. The available state primitives are:

When the user-defined function implements the `Checkpointed` interface, the `snapshotState(…)` and `restoreState(…)`
methods will be executed to draw and restore function state.
* `ValueState<T>`: This keeps a value that can be updated and
retrieved (scoped to key of the input element, mentioned above, so there will possibly be one value
for each key that the operation sees). The value can be set using `update(T)` and retrieved using
`T value()`.

In addition to that, user functions can also implement the `CheckpointNotifier` interface to receive notifications on
completed checkpoints via the `notifyCheckpointComplete(long checkpointId)` method.
Note that there is no guarantee for the user function to receive a notification if a failure happens between
checkpoint completion and notification. The notifications should hence be treated in a way that notifications from
later checkpoints can subsume missing notifications.
* `ListState<T>`: This keeps a list of elements. You can append elements and retrieve an `Iterable`
over all currently stored elements. Elements are added using `add(T)`, the Iterable can
be retrieved using `Iterable<T> get()`.

For example the same counting, reduce function shown for `OperatorState`s by using the `Checkpointed` interface instead:
* `ReducingState<T>`: This keeps a single value that represents the aggregation of all values
added to the state. The interface is the same as for `ListState` but elements added using
`add(T)` are reduced to an aggregate using a specified `ReduceFunction`.

{% highlight java %}
public class CounterSum extends ReduceFunction<Long>, Checkpointed<Long> {
All types of state also have a method `clear()` that clears the state for the currently
active key (i.e. the key of the input element).

// persistent counter
private long counter = 0;
It is important to keep in mind that these state objects are only used for interfacing
with state. The state is not necessarily stored inside but might reside on disk or somewhere else.
The second thing to keep in mind is that the value you get from the state
depend on the key of the input element. So the value you get in one invocation of your
user function can be different from the one you get in another invocation if the key of
the element is different.

@Override
public Long reduce(Long value1, Long value2) {
counter++;
return value1 + value2;
}
To get a state handle you have to create a `StateDescriptor` this holds the name of the state
(as we will later see you can create several states, and they have to have unique names so
that you can reference them), the type of the values that the state holds and possibly
a user-specified function, such as a `ReduceFunction`. Depending on what type of state you
want to retrieve you create one of `ValueStateDescriptor`, `ListStateDescriptor` or
`ReducingStateDescriptor`.

// regularly persists state during normal operation
@Override
public Serializable snapshotState(long checkpointId, long checkpointTimestamp) {
return counter;
}
State is accessed using the `RuntimeContext`, so it is only possible in *rich functions*.
Please see [here]({{ site.baseurl }}/apis/common/#specifying-transformation-functions) for
information about that but we will also see an example shortly. The `RuntimeContext` that
is available in a `RichFunction` has these methods for accessing state:

// restores state on recovery from failure
@Override
public void restoreState(Long state) {
counter = state;
}
}
{% endhighlight %}
* `ValueState<T> getState(ValueStateDescriptor<T>)`
* `ReducingState<T> getReducingState(ReducingStateDescriptor<T>)`
* `ListState<T> getListState(ListStateDescriptor<T>)`

## Using the Key/Value State Interface
This is an example `FlatMapFunction` that shows how all of the parts fit together:

The state interface gives access to key/value states, which are a collection of key/value pairs.
Because the state is partitioned by the keys (distributed accross workers), it can only be used
on the `KeyedStream`, created via `stream.keyBy(…)` (which means also that it is usable in all
types of functions on keyed windows).
{% highlight java %}
public class CountWindowAverage extends RichFlatMapFunction<Tuple2<Long, Long>, Tuple2<Long, Long>> {

The handle to the state can be obtained from the function's `RuntimeContext`. The state handle will
then give access to the value mapped under the key of the current record or window - each key consequently
has its own value.
/**
* The ValueState handle. The first field is the count, the second field a running sum.
*/
private transient ValueState<Tuple2<Long, Long>> sum;

The following code sample shows how to use the key/value state inside a reduce function.
When creating the state handle, one needs to supply a name for that state (a function can have multiple states
of different types), the type of the state (used to create efficient serializers), and the default value (returned
as a value for keys that do not yet have a value associated).
@Override
public void flatMap(Tuple2<Long, Long> input, Collector<Tuple2<Long, Long>> out) throws Exception {

{% highlight java %}
public class CounterSum extends RichReduceFunction<Long> {
// access the state value
Tuple2<Long, Long> currentSum = sum.value();

/** The state handle */
private OperatorState<Long> counter;
// update the count
currentSum.f0 += 1;

@Override
public Long reduce(Long value1, Long value2) {
counter.update(counter.value() + 1);
return value1 + value2;
// add the second field of the input value
currentSum.f1 += input.f1;

// update the state
sum.update(currentSum);

// if the count reaches 2, emit the average and clear the state
if (currentSum.f0 >= 2) {
out.collect(new Tuple2<>(input.f0, currentSum.f1 / currentSum.f0));
sum.clear();
}
}

@Override
public void open(Configuration config) {
counter = getRuntimeContext().getKeyValueState("myCounter", Long.class, 0L);
ValueStateDescriptor<Tuple2<Long, Long>> descriptor =
new ValueStateDescriptor<>(
"average", // the state name
TypeInformation.of(new TypeHint<Tuple2<Long, Long>>() {}), // type information
Tuple2.of(0L, 0L)); // default value of the state, if nothing was set
sum = getRuntimeContext().getState(descriptor);
}
}

// this can be used in a streaming program like this (assuming we have a StreamExecutionEnvironment env)
env.fromElements(Tuple2.of(1L, 3L), Tuple2.of(1L, 5L), Tuple2.of(1L, 7L), Tuple2.of(1L, 4L), Tuple2.of(1L, 2L))
.keyBy(0)
.flatMap(new CountWindowAverage())
.print();

// the printed output will be (1,4) and (1,5)
{% endhighlight %}

State updated by this is usually kept locally inside the flink process (unless one configures explicitly
an external state backend). This means that lookups and updates are process local and this very fast.
This example implements a poor man's counting window. We key the tuples by the first field
(in the example all have the same key `1`). The function stores the count and a running sum in
a `ValueState`, once the count reaches 2 it will emit the average and clear the state so that
we start over from `0`. Note that this would keep a different state value for each different input
key if we had tuples with different values in the first field.

The important implication of having the keys set implicitly is that it forces programs to group the stream
by key (via the `keyBy()` function), making the key partitioning transparent to Flink. That allows the system
to efficiently restore and redistribute keys and state.
### State in the Scala DataStream API

The Scala API has shortcuts that for stateful `map()` or `flatMap()` functions on `KeyedStream`, which give the
state of the current key as an option directly into the function, and return the result with a state update:
In addition to the interface described above, the Scala API has shortcuts for stateful
`map()` or `flatMap()` functions with a single `ValueState` on `KeyedStream`. The user function
gets the current value of the `ValueState` in an `Option` and must return an updated value that
will be used to update the state.

{% highlight scala %}
val stream: DataStream[(String, Int)] = ...
Expand All @@ -143,14 +172,80 @@ val counts: DataStream[(String, Int)] = stream
})
{% endhighlight %}

## Checkpointing Instance Fields

Instance fields can be checkpointed by using the `Checkpointed` interface.

When the user-defined function implements the `Checkpointed` interface, the `snapshotState(…)` and `restoreState(…)`
methods will be executed to draw and restore function state.

In addition to that, user functions can also implement the `CheckpointNotifier` interface to receive notifications on
completed checkpoints via the `notifyCheckpointComplete(long checkpointId)` method.
Note that there is no guarantee for the user function to receive a notification if a failure happens between
checkpoint completion and notification. The notifications should hence be treated in a way that notifications from
later checkpoints can subsume missing notifications.

The above example for `ValueState` can be implemented using instance fields like this:

{% highlight java %}

public class CountWindowAverage
extends RichFlatMapFunction<Tuple2<Long, Long>, Tuple2<Long, Long>>
implements Checkpointed<Tuple2<Long, Long>> {

private Tuple2<Long, Long> sum = null;

@Override
public void flatMap(Tuple2<Long, Long> input, Collector<Tuple2<Long, Long>> out) throws Exception {

// update the count
sum.f0 += 1;

// add the second field of the input value
sum.f1 += input.f1;


// if the count reaches 2, emit the average and clear the state
if (sum.f0 >= 2) {
out.collect(new Tuple2<>(input.f0, sum.f1 / sum.f0));
sum = Tuple2.of(0L, 0L);
}
}

@Override
public void open(Configuration config) {
if (sum == null) {
// only recreate if null
// restoreState will be called before open()
// so this will already set the sum to the restored value
sum = Tuple2.of(0L, 0L);
}
}

// regularly persists state during normal operation
@Override
public Serializable snapshotState(long checkpointId, long checkpointTimestamp) {
return sum;
}

// restores state on recovery from failure
@Override
public void restoreState(Tuple2<Long, Long> state) {
sum = state;
}
}
{% endhighlight %}

## Stateful Source Functions

Stateful sources require a bit more care as opposed to other operators.
In order to make the updates to the state and output collection atomic (required for exactly-once semantics
on failure/recovery), the user is required to get a lock from the source's context.

{% highlight java %}
public static class CounterSource extends RichParallelSourceFunction<Long>, Checkpointed<Long> {
public static class CounterSource
extends RichParallelSourceFunction<Long>
implements Checkpointed<Long> {

/** current offset for exactly once semantics */
private long offset;
Expand Down

0 comments on commit e650d13

Please sign in to comment.