Package org.apache.flink.state.rocksdb

Class RocksDBKeyedStateBackend<K>

java.lang.Object
org.apache.flink.runtime.state.AbstractKeyedStateBackend<K>
org.apache.flink.state.rocksdb.RocksDBKeyedStateBackend<K>
All Implemented Interfaces:
Closeable, AutoCloseable, org.apache.flink.api.common.state.CheckpointListener, org.apache.flink.api.common.state.InternalCheckpointListener, org.apache.flink.runtime.state.CheckpointableKeyedStateBackend<K>, org.apache.flink.runtime.state.InternalKeyContext<K>, org.apache.flink.runtime.state.KeyedStateBackend<K>, org.apache.flink.runtime.state.KeyedStateFactory, org.apache.flink.runtime.state.PriorityQueueSetFactory, org.apache.flink.runtime.state.Snapshotable<org.apache.flink.runtime.state.SnapshotResult<org.apache.flink.runtime.state.KeyedStateHandle>>, org.apache.flink.runtime.state.TestableKeyedStateBackend<K>, org.apache.flink.util.Disposable
public class RocksDBKeyedStateBackend<K> extends org.apache.flink.runtime.state.AbstractKeyedStateBackend<K>
An AbstractKeyedStateBackend that stores its state in RocksDB and serializes state to streams provided by a CheckpointStreamFactory upon checkpointing. This state backend can store very large state that exceeds memory and spills to disk. Except for the snapshotting, this class should be accessed as if it is not threadsafe.

This class follows the rules for closing/releasing native RocksDB resources as described in + this document.

  • Nested Class Summary

    Nested Classes
    Modifier and Type
    Class
    Description
    static class 
    RocksDBKeyedStateBackend.RocksDbKvStateInfo
    Rocks DB specific information about the k/v states.

    Nested classes/interfaces inherited from class org.apache.flink.runtime.state.AbstractKeyedStateBackend

    org.apache.flink.runtime.state.AbstractKeyedStateBackend.PartitionStateFactory

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

    org.apache.flink.runtime.state.KeyedStateBackend.KeySelectionListener<K extends Object>

  • Field Summary

    Fields
    Modifier and Type
    Field
    Description
    protected final org.rocksdb.RocksDB
    db
    Our RocksDB database, this is used by the actual subclasses of AbstractRocksDBState to store state.
    static final String
    MERGE_OPERATOR_NAME
    The name of the merge operator in RocksDB.

    Fields inherited from class org.apache.flink.runtime.state.AbstractKeyedStateBackend

    cancelStreamRegistry, keyContext, keyGroupCompressionDecorator, keyGroupRange, keySerializer, kvStateRegistry, latencyTrackingStateConfig, numberOfKeyGroups, ttlTimeProvider, userCodeClassLoader

  • Constructor Summary

    Constructors
    Constructor
    Description
    RocksDBKeyedStateBackend(ClassLoader userCodeClassLoader, File instanceBasePath, RocksDBResourceContainer optionsContainer, Function<String,org.rocksdb.ColumnFamilyOptions> columnFamilyOptionsFactory, org.apache.flink.runtime.query.TaskKvStateRegistry kvStateRegistry, org.apache.flink.api.common.typeutils.TypeSerializer<K> keySerializer, org.apache.flink.api.common.ExecutionConfig executionConfig, org.apache.flink.runtime.state.ttl.TtlTimeProvider ttlTimeProvider, org.apache.flink.runtime.state.metrics.LatencyTrackingStateConfig latencyTrackingStateConfig, org.rocksdb.RocksDB db, LinkedHashMap<String,RocksDBKeyedStateBackend.RocksDbKvStateInfo> kvStateInformation, Map<String,org.apache.flink.runtime.state.heap.HeapPriorityQueueSnapshotRestoreWrapper<?>> registeredPQStates, int keyGroupPrefixBytes, org.apache.flink.core.fs.CloseableRegistry cancelStreamRegistry, org.apache.flink.runtime.state.StreamCompressionDecorator keyGroupCompressionDecorator, org.apache.flink.util.ResourceGuard rocksDBResourceGuard, RocksDBSnapshotStrategyBase<K,?> checkpointSnapshotStrategy, RocksDBWriteBatchWrapper writeBatchWrapper, org.rocksdb.ColumnFamilyHandle defaultColumnFamilyHandle, RocksDBNativeMetricMonitor nativeMetricMonitor, org.apache.flink.runtime.state.SerializedCompositeKeyBuilder<K> sharedRocksKeyBuilder, org.apache.flink.runtime.state.PriorityQueueSetFactory priorityQueueFactory, RocksDbTtlCompactFiltersManager ttlCompactFiltersManager, org.apache.flink.runtime.state.InternalKeyContext<K> keyContext, long writeBatchSize, CompletableFuture<Void> asyncCompactFuture, RocksDBManualCompactionManager rocksDBManualCompactionManager)
     

  • Method Summary

    Modifier and Type
    Method
    Description
    void
    compactState(org.apache.flink.api.common.state.StateDescriptor<?,?> stateDesc)
     
    <T extends org.apache.flink.runtime.state.heap.HeapPriorityQueueElement & org.apache.flink.runtime.state.PriorityComparable<? super T> & org.apache.flink.runtime.state.Keyed<?>>
    org.apache.flink.runtime.state.KeyGroupedInternalPriorityQueue<T>
    create(String stateName, org.apache.flink.api.common.typeutils.TypeSerializer<T> byteOrderedElementSerializer)
     
    <T extends org.apache.flink.runtime.state.heap.HeapPriorityQueueElement & org.apache.flink.runtime.state.PriorityComparable<? super T> & org.apache.flink.runtime.state.Keyed<?>>
    org.apache.flink.runtime.state.KeyGroupedInternalPriorityQueue<T>
    create(String stateName, org.apache.flink.api.common.typeutils.TypeSerializer<T> byteOrderedElementSerializer, boolean allowFutureMetadataUpdates)
     
    <N, SV, SEV, S extends org.apache.flink.api.common.state.State, IS extends S>
    IS
    createOrUpdateInternalState(org.apache.flink.api.common.typeutils.TypeSerializer<N> namespaceSerializer, org.apache.flink.api.common.state.StateDescriptor<S,SV> stateDesc, org.apache.flink.runtime.state.StateSnapshotTransformer.StateSnapshotTransformFactory<SEV> snapshotTransformFactory)
     
    <N, SV, SEV, S extends org.apache.flink.api.common.state.State, IS extends S>
    IS
    createOrUpdateInternalState(org.apache.flink.api.common.typeutils.TypeSerializer<N> namespaceSerializer, org.apache.flink.api.common.state.StateDescriptor<S,SV> stateDesc, org.apache.flink.runtime.state.StateSnapshotTransformer.StateSnapshotTransformFactory<SEV> snapshotTransformFactory, boolean allowFutureMetadataUpdates)
     
    void
    dispose()
    Should only be called by one thread, and only after all accesses to the DB happened.
    Optional<CompletableFuture<Void>>
    getAsyncCompactAfterRestoreFuture()
     
    int
    getKeyGroupPrefixBytes()
     
    <N> Stream<K>
    getKeys(String state, N namespace)
     
    <N> Stream<org.apache.flink.api.java.tuple.Tuple2<K,N>>
    getKeysAndNamespaces(String state)
     
    org.rocksdb.ReadOptions
    getReadOptions()
     
    org.rocksdb.WriteOptions
    getWriteOptions()
     
    boolean
    isSafeToReuseKVState()
     
    void
    notifyCheckpointAborted(long checkpointId)
     
    void
    notifyCheckpointComplete(long completedCheckpointId)
     
    int
    numKeyValueStateEntries()
     
    boolean
    requiresLegacySynchronousTimerSnapshots(org.apache.flink.runtime.checkpoint.SnapshotType checkpointType)
     
    org.apache.flink.runtime.state.SavepointResources<K>
    savepoint()
     
    void
    setCurrentKey(K newKey)
     
    void
    setCurrentKeyAndKeyGroup(K newKey, int newKeyGroupIndex)
     
    RunnableFuture<org.apache.flink.runtime.state.SnapshotResult<org.apache.flink.runtime.state.KeyedStateHandle>>
    snapshot(long checkpointId, long timestamp, org.apache.flink.runtime.state.CheckpointStreamFactory streamFactory, org.apache.flink.runtime.checkpoint.CheckpointOptions checkpointOptions)
    Triggers an asynchronous snapshot of the keyed state backend from RocksDB.

    Methods inherited from class org.apache.flink.runtime.state.AbstractKeyedStateBackend

    applyToAllKeys, applyToAllKeys, close, deregisterKeySelectionListener, getCurrentKey, getCurrentKeyGroupIndex, getKeyContext, getKeyGroupCompressionDecorator, getKeyGroupRange, getKeySerializer, getLatencyTrackingStateConfig, getNumberOfKeyGroups, getOrCreateKeyedState, getPartitionedState, notifyCheckpointSubsumed, numKeyValueStatesByName, publishQueryableStateIfEnabled, registerKeySelectionListener, setCurrentKeyGroupIndex

    Methods inherited from class java.lang.Object

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

    Methods inherited from interface org.apache.flink.runtime.state.KeyedStateFactory

    createOrUpdateInternalState

    Methods inherited from interface org.apache.flink.runtime.state.TestableKeyedStateBackend

    getDelegatedKeyedStateBackend

  • Field Details

    • MERGE_OPERATOR_NAME

      public static final String MERGE_OPERATOR_NAME
      The name of the merge operator in RocksDB. Do not change except you know exactly what you do.
      See Also:

    • db

      protected final org.rocksdb.RocksDB db
      Our RocksDB database, this is used by the actual subclasses of AbstractRocksDBState to store state. The different k/v states that we have don't each have their own RocksDB instance. They all write to this instance but to their own column family.

  • Constructor Details

    • RocksDBKeyedStateBackend

      public RocksDBKeyedStateBackend(ClassLoader userCodeClassLoader, File instanceBasePath, RocksDBResourceContainer optionsContainer, Function<String,org.rocksdb.ColumnFamilyOptions> columnFamilyOptionsFactory, org.apache.flink.runtime.query.TaskKvStateRegistry kvStateRegistry, org.apache.flink.api.common.typeutils.TypeSerializer<K> keySerializer, org.apache.flink.api.common.ExecutionConfig executionConfig, org.apache.flink.runtime.state.ttl.TtlTimeProvider ttlTimeProvider, org.apache.flink.runtime.state.metrics.LatencyTrackingStateConfig latencyTrackingStateConfig, org.rocksdb.RocksDB db, LinkedHashMap<String,RocksDBKeyedStateBackend.RocksDbKvStateInfo> kvStateInformation, Map<String,org.apache.flink.runtime.state.heap.HeapPriorityQueueSnapshotRestoreWrapper<?>> registeredPQStates, int keyGroupPrefixBytes, org.apache.flink.core.fs.CloseableRegistry cancelStreamRegistry, org.apache.flink.runtime.state.StreamCompressionDecorator keyGroupCompressionDecorator, org.apache.flink.util.ResourceGuard rocksDBResourceGuard, RocksDBSnapshotStrategyBase<K,?> checkpointSnapshotStrategy, RocksDBWriteBatchWrapper writeBatchWrapper, org.rocksdb.ColumnFamilyHandle defaultColumnFamilyHandle, RocksDBNativeMetricMonitor nativeMetricMonitor, org.apache.flink.runtime.state.SerializedCompositeKeyBuilder<K> sharedRocksKeyBuilder, org.apache.flink.runtime.state.PriorityQueueSetFactory priorityQueueFactory, RocksDbTtlCompactFiltersManager ttlCompactFiltersManager, org.apache.flink.runtime.state.InternalKeyContext<K> keyContext, @Nonnegative long writeBatchSize, @Nullable CompletableFuture<Void> asyncCompactFuture, RocksDBManualCompactionManager rocksDBManualCompactionManager)

  • Method Details

    • getKeys

      public <N> Stream<K> getKeys(String state, N namespace)

    • getKeysAndNamespaces

      public <N> Stream<org.apache.flink.api.java.tuple.Tuple2<K,N>> getKeysAndNamespaces(String state)

    • setCurrentKey

      public void setCurrentKey(K newKey)
      Specified by:
      setCurrentKey in interface org.apache.flink.runtime.state.InternalKeyContext<K>
      Specified by:
      setCurrentKey in interface org.apache.flink.runtime.state.KeyedStateBackend<K>
      Overrides:
      setCurrentKey in class org.apache.flink.runtime.state.AbstractKeyedStateBackend<K>

    • setCurrentKeyAndKeyGroup

      public void setCurrentKeyAndKeyGroup(K newKey, int newKeyGroupIndex)
      Specified by:
      setCurrentKeyAndKeyGroup in interface org.apache.flink.runtime.state.KeyedStateBackend<K>
      Overrides:
      setCurrentKeyAndKeyGroup in class org.apache.flink.runtime.state.AbstractKeyedStateBackend<K>

    • dispose

      public void dispose()
      Should only be called by one thread, and only after all accesses to the DB happened.
      Specified by:
      dispose in interface org.apache.flink.util.Disposable
      Specified by:
      dispose in interface org.apache.flink.runtime.state.KeyedStateBackend<K>
      Overrides:
      dispose in class org.apache.flink.runtime.state.AbstractKeyedStateBackend<K>

    • create

      @Nonnull public <T extends org.apache.flink.runtime.state.heap.HeapPriorityQueueElement & org.apache.flink.runtime.state.PriorityComparable<? super T> & org.apache.flink.runtime.state.Keyed<?>> org.apache.flink.runtime.state.KeyGroupedInternalPriorityQueue<T> create(@Nonnull String stateName, @Nonnull org.apache.flink.api.common.typeutils.TypeSerializer<T> byteOrderedElementSerializer)

    • create

      public <T extends org.apache.flink.runtime.state.heap.HeapPriorityQueueElement & org.apache.flink.runtime.state.PriorityComparable<? super T> & org.apache.flink.runtime.state.Keyed<?>> org.apache.flink.runtime.state.KeyGroupedInternalPriorityQueue<T> create(@Nonnull String stateName, @Nonnull org.apache.flink.api.common.typeutils.TypeSerializer<T> byteOrderedElementSerializer, boolean allowFutureMetadataUpdates)

    • getKeyGroupPrefixBytes

      public int getKeyGroupPrefixBytes()

    • getWriteOptions

      public org.rocksdb.WriteOptions getWriteOptions()

    • getReadOptions

      public org.rocksdb.ReadOptions getReadOptions()

    • snapshot

      @Nonnull public RunnableFuture<org.apache.flink.runtime.state.SnapshotResult<org.apache.flink.runtime.state.KeyedStateHandle>> snapshot(long checkpointId, long timestamp, @Nonnull org.apache.flink.runtime.state.CheckpointStreamFactory streamFactory, @Nonnull org.apache.flink.runtime.checkpoint.CheckpointOptions checkpointOptions) throws Exception
      Triggers an asynchronous snapshot of the keyed state backend from RocksDB. This snapshot can be canceled and is also stopped when the backend is closed through dispose(). For each backend, this method must always be called by the same thread.
      Parameters:
      checkpointId - The Id of the checkpoint.
      timestamp - The timestamp of the checkpoint.
      streamFactory - The factory that we can use for writing our state to streams.
      checkpointOptions - Options for how to perform this checkpoint.
      Returns:
      Future to the state handle of the snapshot data.
      Throws:
      Exception - indicating a problem in the synchronous part of the checkpoint.

    • savepoint

      @Nonnull public org.apache.flink.runtime.state.SavepointResources<K> savepoint() throws Exception
      Throws:
      Exception

    • notifyCheckpointComplete

      public void notifyCheckpointComplete(long completedCheckpointId) throws Exception
      Throws:
      Exception

    • notifyCheckpointAborted

      public void notifyCheckpointAborted(long checkpointId) throws Exception
      Throws:
      Exception

    • createOrUpdateInternalState

      @Nonnull public <N, SV, SEV, S extends org.apache.flink.api.common.state.State, IS extends S> IS createOrUpdateInternalState(@Nonnull org.apache.flink.api.common.typeutils.TypeSerializer<N> namespaceSerializer, @Nonnull org.apache.flink.api.common.state.StateDescriptor<S,SV> stateDesc, @Nonnull org.apache.flink.runtime.state.StateSnapshotTransformer.StateSnapshotTransformFactory<SEV> snapshotTransformFactory) throws Exception
      Throws:
      Exception

    • createOrUpdateInternalState

      @Nonnull public <N, SV, SEV, S extends org.apache.flink.api.common.state.State, IS extends S> IS createOrUpdateInternalState(@Nonnull org.apache.flink.api.common.typeutils.TypeSerializer<N> namespaceSerializer, @Nonnull org.apache.flink.api.common.state.StateDescriptor<S,SV> stateDesc, @Nonnull org.apache.flink.runtime.state.StateSnapshotTransformer.StateSnapshotTransformFactory<SEV> snapshotTransformFactory, boolean allowFutureMetadataUpdates) throws Exception
      Throws:
      Exception

    • numKeyValueStateEntries

      @VisibleForTesting public int numKeyValueStateEntries()

    • requiresLegacySynchronousTimerSnapshots

      public boolean requiresLegacySynchronousTimerSnapshots(org.apache.flink.runtime.checkpoint.SnapshotType checkpointType)
      Overrides:
      requiresLegacySynchronousTimerSnapshots in class org.apache.flink.runtime.state.AbstractKeyedStateBackend<K>

    • isSafeToReuseKVState

      public boolean isSafeToReuseKVState()

    • compactState

      @VisibleForTesting public void compactState(org.apache.flink.api.common.state.StateDescriptor<?,?> stateDesc) throws org.rocksdb.RocksDBException
      Throws:
      org.rocksdb.RocksDBException

    • getAsyncCompactAfterRestoreFuture

      @VisibleForTesting public Optional<CompletableFuture<Void>> getAsyncCompactAfterRestoreFuture()