org.apache.flink.contrib.streaming.state.snapshot

Class RocksIncrementalSnapshotStrategy<K>

java.lang.Object

org.apache.flink.contrib.streaming.state.snapshot.RocksDBSnapshotStrategyBase<K,RocksDBSnapshotStrategyBase.NativeRocksDBSnapshotResources>

org.apache.flink.contrib.streaming.state.snapshot.RocksIncrementalSnapshotStrategy<K>

Type Parameters:

K - type of the backend keys.

All Implemented Interfaces:

AutoCloseable, CheckpointListener, SnapshotStrategy<KeyedStateHandle,RocksDBSnapshotStrategyBase.NativeRocksDBSnapshotResources>

public class RocksIncrementalSnapshotStrategy<K>
extends RocksDBSnapshotStrategyBase<K,RocksDBSnapshotStrategyBase.NativeRocksDBSnapshotResources>

Snapshot strategy for RocksDBKeyedStateBackend that is based on RocksDB's native checkpoints and creates incremental snapshots.

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.flink.contrib.streaming.state.snapshot.RocksDBSnapshotStrategyBase

RocksDBSnapshotStrategyBase.NativeRocksDBSnapshotResources, RocksDBSnapshotStrategyBase.PreviousSnapshot, RocksDBSnapshotStrategyBase.RocksDBSnapshotOperation

Nested classes/interfaces inherited from interface org.apache.flink.runtime.state.SnapshotStrategy

SnapshotStrategy.SnapshotResultSupplier<S extends StateObject>

Field Summary

Fields inherited from class org.apache.flink.contrib.streaming.state.snapshot.RocksDBSnapshotStrategyBase

backendUID, db, EMPTY_PREVIOUS_SNAPSHOT, instanceBasePath, keyGroupPrefixBytes, keyGroupRange, keySerializer, kvStateInformation, localDirectoryName, localRecoveryConfig, rocksDBResourceGuard

Constructor Summary

Constructors

Constructor and Description

RocksIncrementalSnapshotStrategy(org.rocksdb.RocksDB db, ResourceGuard rocksDBResourceGuard, TypeSerializer<K> keySerializer, LinkedHashMap<String,RocksDBKeyedStateBackend.RocksDbKvStateInfo> kvStateInformation, KeyGroupRange keyGroupRange, int keyGroupPrefixBytes, LocalRecoveryConfig localRecoveryConfig, CloseableRegistry cancelStreamRegistry, File instanceBasePath, UUID backendUID, SortedMap<Long,Collection<IncrementalKeyedStateHandle.HandleAndLocalPath>> uploadedStateHandles, RocksDBStateUploader rocksDBStateUploader, long lastCompletedCheckpointId)

Method Summary

Modifier and Type	Method and Description
SnapshotStrategy.SnapshotResultSupplier<KeyedStateHandle>	asyncSnapshot(RocksDBSnapshotStrategyBase.NativeRocksDBSnapshotResources snapshotResources, long checkpointId, long timestamp, CheckpointStreamFactory checkpointStreamFactory, CheckpointOptions checkpointOptions) Operation that writes a snapshot into a stream that is provided by the given CheckpointStreamFactory and returns a @SupplierWithException that gives a state handle to the snapshot.
void	close()
void	notifyCheckpointAborted(long abortedCheckpointId) This method is called as a notification once a distributed checkpoint has been aborted.
void	notifyCheckpointComplete(long completedCheckpointId) Notifies the listener that the checkpoint with the given checkpointId completed and was committed.
protected RocksDBSnapshotStrategyBase.PreviousSnapshot	snapshotMetaData(long checkpointId, List<StateMetaInfoSnapshot> stateMetaInfoSnapshots)

Methods inherited from class org.apache.flink.contrib.streaming.state.snapshot.RocksDBSnapshotStrategyBase

cleanupIncompleteSnapshot, getDescription, materializeMetaData, prepareLocalSnapshotDirectory, syncPrepareResources

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

RocksIncrementalSnapshotStrategy

public RocksIncrementalSnapshotStrategy(@Nonnull
                                        org.rocksdb.RocksDB db,
                                        @Nonnull
                                        ResourceGuard rocksDBResourceGuard,
                                        @Nonnull
                                        TypeSerializer<K> keySerializer,
                                        @Nonnull
                                        LinkedHashMap<String,RocksDBKeyedStateBackend.RocksDbKvStateInfo> kvStateInformation,
                                        @Nonnull
                                        KeyGroupRange keyGroupRange,
                                        @Nonnegative
                                        int keyGroupPrefixBytes,
                                        @Nonnull
                                        LocalRecoveryConfig localRecoveryConfig,
                                        @Nonnull
                                        CloseableRegistry cancelStreamRegistry,
                                        @Nonnull
                                        File instanceBasePath,
                                        @Nonnull
                                        UUID backendUID,
                                        @Nonnull
                                        SortedMap<Long,Collection<IncrementalKeyedStateHandle.HandleAndLocalPath>> uploadedStateHandles,
                                        @Nonnull
                                        RocksDBStateUploader rocksDBStateUploader,
                                        long lastCompletedCheckpointId)

Method Detail

asyncSnapshot

public SnapshotStrategy.SnapshotResultSupplier<KeyedStateHandle> asyncSnapshot(RocksDBSnapshotStrategyBase.NativeRocksDBSnapshotResources snapshotResources,
                                                                               long checkpointId,
                                                                               long timestamp,
                                                                               @Nonnull
                                                                               CheckpointStreamFactory checkpointStreamFactory,
                                                                               @Nonnull
                                                                               CheckpointOptions checkpointOptions)

Description copied from interface: SnapshotStrategy

Operation that writes a snapshot into a stream that is provided by the given CheckpointStreamFactory and returns a @SupplierWithException that gives a state handle to the snapshot.

checkpointId - The ID of the checkpoint.

timestamp - The timestamp of the checkpoint.

checkpointStreamFactory - The factory that we can use for writing our state to streams.

checkpointOptions - Options for how to perform this checkpoint.

Returns:

A supplier that will yield a StateObject.

notifyCheckpointComplete

public void notifyCheckpointComplete(long completedCheckpointId)

Description copied from interface: CheckpointListener

Notifies the listener that the checkpoint with the given checkpointId completed and was committed.

These notifications are "best effort", meaning they can sometimes be skipped. To behave properly, implementers need to follow the "Checkpoint Subsuming Contract". Please see the class-level JavaDocs for details.

Please note that checkpoints may generally overlap, so you cannot assume that the notifyCheckpointComplete() call is always for the latest prior checkpoint (or snapshot) that was taken on the function/operator implementing this interface. It might be for a checkpoint that was triggered earlier. Implementing the "Checkpoint Subsuming Contract" (see above) properly handles this situation correctly as well.

Please note that throwing exceptions from this method will not cause the completed checkpoint to be revoked. Throwing exceptions will typically cause task/job failure and trigger recovery.

Parameters:

completedCheckpointId - The ID of the checkpoint that has been completed.

notifyCheckpointAborted

public void notifyCheckpointAborted(long abortedCheckpointId)

Description copied from interface: CheckpointListener

This method is called as a notification once a distributed checkpoint has been aborted.

Important: The fact that a checkpoint has been aborted does NOT mean that the data and artifacts produced between the previous checkpoint and the aborted checkpoint are to be discarded. The expected behavior is as if this checkpoint was never triggered in the first place, and the next successful checkpoint simply covers a longer time span. See the "Checkpoint Subsuming Contract" in the class-level JavaDocs for details.

These notifications are "best effort", meaning they can sometimes be skipped.

This method is very rarely necessary to implement. The "best effort" guarantee, together with the fact that this method should not result in discarding any data (per the "Checkpoint Subsuming Contract") means it is mainly useful for earlier cleanups of auxiliary resources. One example is to pro-actively clear a local per-checkpoint state cache upon checkpoint failure.

Parameters:

abortedCheckpointId - The ID of the checkpoint that has been aborted.

close

public void close()

Specified by:

close in interface AutoCloseable

Specified by:

close in class RocksDBSnapshotStrategyBase<K,RocksDBSnapshotStrategyBase.NativeRocksDBSnapshotResources>

snapshotMetaData

protected RocksDBSnapshotStrategyBase.PreviousSnapshot snapshotMetaData(long checkpointId,
                                                                        @Nonnull
                                                                        List<StateMetaInfoSnapshot> stateMetaInfoSnapshots)

Specified by:

snapshotMetaData in class RocksDBSnapshotStrategyBase<K,RocksDBSnapshotStrategyBase.NativeRocksDBSnapshotResources>