After using namespaced kubernetes client, we will not need to always set the namespace when creating kubernetes resources(e.g. deployment, pods, configmap, watch, etc.).

Address comments: Update the unit test to verify the configmap watch is created in appropriate namespace

This closes #14570.

This commits adds DeclarativeSlotPool.setResourceRequirements which sets the absolutely
required resources. Hence, this method can be used to overwrite the currently set
resource requirements.

This closes #14589.

This closes #14580

ExeucutorUtils.setBatchProperties(StreamExecutionEnvironment) does not set
InputDependencyContraint anymore as it's not used anywhere.

FLINK-20589

This work is part of FLINK-20589

This closes #14579

[FLINK-20600][runtime] Deleted unused SlotSharingManager and corresponding test class SlotSharingManagerTest.

This closes #14578

[FLINK-20596][runtime] Removed SlotProvider and Scheduler interfaces, corresponding classes, and test cases. Comments were adapted accordingly.

This closes #14598

[FLINK-20857][table-planner-blink] Introduce BatchPhysicalPythonGroupWindowAggregate, and make BatchExecPythonGroupWindowAggregate only extended from ExecNode

This closes #14574

[FLINK-20857][table-planner-blink] Introduce BatchPhysicalSortWindowAggregate & BatchPhysicalLocalSortWindowAggregate, and make BatchExecSortWindowAggregate only extended from ExecNode

This closes #14574

[FLINK-20857][table-planner-blink] Introduce BatchPhysicalHashWindowAggregate & BatchPhysicalLocalHashWindowAggregate, and make BatchExecHashWindowAggregate only extended from ExecNode

This closes #14574

[FLINK-20857][table-planner-blink] Rename BatchExecWindowAggregateBase to BatchPhysicalWindowAggregateBase and do some refactoring

This closes #14574

This closes #14539
Co-authored-by: NAlex Wang <alex.wang.dev@gmail.com>

This closes #14583

The YarnClusterDescriptor.zookeeperNamespace has been replaced by the configuration which
is deployed together with the Yarn cluster. Hence, it is no longer needed.

In order to easily connect to an HA enabled Yarn cluster, this commit sets the
high-availability.cluster-id to the application id if not configured. This is
symmetric to how we deploy HA enabled clusters and makes it unnecessary to
explicitly set the high-availability.cluster-id if one tries to reconnect
to the cluster.

This closes #14577.

This closes #14576

This closes #14560

[FLINK-15660][runtime] request slot should success if the slot is already allocated with the same allocation id

This closes #14378.

Completes the envisioned design of FLIP-32 and FLIP-65. It makes
implementing built-in functions as easy as implementing user-defined
functions while keeping function definition and body (aka runtime
implementation) separately.

Adding a built-in function requires changes in only 2 classes:
- BuiltInFunctionDefinitions for the definition and
- the class that contains the actual runtime logic.

Introduces the concept of SpecializedFunctions to hook into the planning phase
for generating runtime implementation lazily.

1. Make BytesMap decouple with BinaryRowData key type
2. Extract pages operations from AbstractRowDataSerializer into PagedTypeSerializer
3. Introduce WindowKey structure and WindowKeySerializer
4. Introduce BytesWindowHashMap and BytesWindowMultiMap to support window key
5. Refactor tests to also cover the new introduced BytesWindowHashMap and BytesWindowMultiMap

This closes #14572

This closes #14572

In regard to InputChannels, there are 3 cases when a checkpoint has to be declined:
1. Channel state is not fully consumed
2. Channel was not yet converted from Recovered to normal
3. Channel was not yet converted from Unknown to normal

In the 1st case, new checkpoint may skip some recovered buffers.
In the 2nd and 3rd cases, not handling checkpointStarted() call
by normal channels can prevent incoming buffers from being captured.
In all these cases new checkpoint would be inconsistent.

This closes #14508

[FLINK-20856][table-planner-blink] Introduce StreamPhysicalPythonGroupWindowAggregate, and make StreamExecPythonGroupWindowAggregate only extended from ExecNode

This closes #14569

[FLINK-20856][table-planner-blink] Introduce StreamPhysicalGroupWindowAggregate & StreamPhysicalGroupWindowTableAggregate, and make StreamExecGroupWindowAggregate only extended from ExecNode

This closes #14569

[FLINK-20836][runtime] Pass total resource and default slot resource in registering TaskManager to SlotManager

This closes #14561

[FLINK-20836][runtime] Add total resource profile and default slot resource profile to WorkerRegistration