kafka ProcessorStateManager 源码
kafka ProcessorStateManager 代码
文件路径：/streams/src/main/java/org/apache/kafka/streams/processor/internals/ProcessorStateManager.java
/*
* 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.
*/
package org.apache.kafka.streams.processor.internals;
import org.apache.kafka.clients.consumer.ConsumerRecord;
import org.apache.kafka.common.TopicPartition;
import org.apache.kafka.common.utils.FixedOrderMap;
import org.apache.kafka.common.utils.LogContext;
import org.apache.kafka.streams.errors.ProcessorStateException;
import org.apache.kafka.streams.errors.StreamsException;
import org.apache.kafka.streams.errors.TaskCorruptedException;
import org.apache.kafka.streams.errors.TaskMigratedException;
import org.apache.kafka.streams.processor.CommitCallback;
import org.apache.kafka.streams.processor.StateRestoreCallback;
import org.apache.kafka.streams.processor.StateRestoreListener;
import org.apache.kafka.streams.processor.StateStore;
import org.apache.kafka.streams.processor.StateStoreContext;
import org.apache.kafka.streams.processor.TaskId;
import org.apache.kafka.streams.processor.internals.Task.TaskType;
import org.apache.kafka.streams.state.internals.CachedStateStore;
import org.apache.kafka.streams.state.internals.OffsetCheckpoint;
import org.apache.kafka.streams.state.internals.RecordConverter;
import org.apache.kafka.streams.state.internals.TimeOrderedKeyValueBuffer;
import org.slf4j.Logger;
import java.io.File;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Collection;
import java.util.Collections;
import java.util.HashMap;
import java.util.LinkedList;
import java.util.List;
import java.util.Map;
import java.util.Set;
import java.util.stream.Collectors;
import static java.lang.String.format;
import static org.apache.kafka.streams.processor.internals.StateManagerUtil.CHECKPOINT_FILE_NAME;
import static org.apache.kafka.streams.processor.internals.StateManagerUtil.converterForStore;
import static org.apache.kafka.streams.processor.internals.StateRestoreCallbackAdapter.adapt;
import static org.apache.kafka.streams.state.internals.OffsetCheckpoint.OFFSET_UNKNOWN;
/**
* ProcessorStateManager is the source of truth for the current offset for each state store,
* which is either the read offset during restoring, or the written offset during normal processing.
*
* The offset is initialized as null when the state store is registered, and then it can be updated by
* loading checkpoint file, restore state stores, or passing from the record collector's written offsets.
*
* When checkpointing, if the offset is not null it would be written to the file.
*
* The manager is also responsible for restoring state stores via their registered restore callback,
* which is used for both updating standby tasks as well as restoring active tasks.
*/
public class ProcessorStateManager implements StateManager {
public static class StateStoreMetadata {
private final StateStore stateStore;
// corresponding changelog partition of the store, this and the following two fields
// will only be not-null if the state store is logged (i.e. changelog partition and restorer provided)
private final TopicPartition changelogPartition;
// could be used for both active restoration and standby
private final StateRestoreCallback restoreCallback;
private final CommitCallback commitCallback;
// record converters used for restoration and standby
private final RecordConverter recordConverter;
// indicating the current snapshot of the store as the offset of last changelog record that has been
// applied to the store used for both restoration (active and standby tasks restored offset) and
// normal processing that update stores (written offset); could be null (when initialized)
//
// the offset is updated in three ways:
// 1. when loading from the checkpoint file, when the corresponding task has acquired the state
// directory lock and have registered all the state store; it is only one-time
// 2. when updating with restore records (by both restoring active and standby),
// update to the last restore record's offset
// 3. when checkpointing with the given written offsets from record collector,
// update blindly with the given offset
private Long offset;
// corrupted state store should not be included in checkpointing
private boolean corrupted;
private StateStoreMetadata(final StateStore stateStore,
final CommitCallback commitCallback) {
this.stateStore = stateStore;
this.commitCallback = commitCallback;
this.restoreCallback = null;
this.recordConverter = null;
this.changelogPartition = null;
this.corrupted = false;
this.offset = null;
}
private StateStoreMetadata(final StateStore stateStore,
final TopicPartition changelogPartition,
final StateRestoreCallback restoreCallback,
final CommitCallback commitCallback,
final RecordConverter recordConverter) {
if (restoreCallback == null) {
throw new IllegalStateException("Log enabled store should always provide a restore callback upon registration");
}
this.stateStore = stateStore;
this.changelogPartition = changelogPartition;
this.restoreCallback = restoreCallback;
this.commitCallback = commitCallback;
this.recordConverter = recordConverter;
this.offset = null;
}
private void setOffset(final Long offset) {
this.offset = offset;
}
// the offset is exposed to the changelog reader to determine if restoration is completed
Long offset() {
return this.offset;
}
TopicPartition changelogPartition() {
return this.changelogPartition;
}
StateStore store() {
return this.stateStore;
}
@Override
public String toString() {
return "StateStoreMetadata (" + stateStore.name() + " : " + changelogPartition + " @ " + offset;
}
}
private static final String STATE_CHANGELOG_TOPIC_SUFFIX = "-changelog";
private Logger log;
private String logPrefix;
private final TaskId taskId;
private final boolean eosEnabled;
private final ChangelogRegister changelogReader;
private final Collection<TopicPartition> sourcePartitions;
private final Map<String, String> storeToChangelogTopic;
// must be maintained in topological order
private final FixedOrderMap<String, StateStoreMetadata> stores = new FixedOrderMap<>();
private final FixedOrderMap<String, StateStore> globalStores = new FixedOrderMap<>();
private final File baseDir;
private final OffsetCheckpoint checkpointFile;
private TaskType taskType;
private final boolean stateUpdaterEnabled;
public static String storeChangelogTopic(final String prefix, final String storeName, final String namedTopology) {
if (namedTopology == null) {
return prefix + "-" + storeName + STATE_CHANGELOG_TOPIC_SUFFIX;
} else {
return prefix + "-" + namedTopology + "-" + storeName + STATE_CHANGELOG_TOPIC_SUFFIX;
}
}
/**
* @throws ProcessorStateException if the task directory does not exist and could not be created
*/
public ProcessorStateManager(final TaskId taskId,
final TaskType taskType,
final boolean eosEnabled,
final LogContext logContext,
final StateDirectory stateDirectory,
final ChangelogRegister changelogReader,
final Map<String, String> storeToChangelogTopic,
final Collection<TopicPartition> sourcePartitions,
final boolean stateUpdaterEnabled) throws ProcessorStateException {
this.storeToChangelogTopic = storeToChangelogTopic;
this.log = logContext.logger(ProcessorStateManager.class);
this.logPrefix = logContext.logPrefix();
this.taskId = taskId;
this.taskType = taskType;
this.eosEnabled = eosEnabled;
this.changelogReader = changelogReader;
this.sourcePartitions = sourcePartitions;
this.stateUpdaterEnabled = stateUpdaterEnabled;
this.baseDir = stateDirectory.getOrCreateDirectoryForTask(taskId);
this.checkpointFile = new OffsetCheckpoint(stateDirectory.checkpointFileFor(taskId));
log.debug("Created state store manager for task {}", taskId);
}
void registerStateStores(final List<StateStore> allStores, final InternalProcessorContext processorContext) {
processorContext.uninitialize();
for (final StateStore store : allStores) {
if (stores.containsKey(store.name())) {
if (!stateUpdaterEnabled) {
maybeRegisterStoreWithChangelogReader(store.name());
}
} else {
store.init((StateStoreContext) processorContext, store);
}
log.trace("Registered state store {}", store.name());
}
}
void registerGlobalStateStores(final List<StateStore> stateStores) {
log.debug("Register global stores {}", stateStores);
for (final StateStore stateStore : stateStores) {
globalStores.put(stateStore.name(), stateStore);
}
}
@Override
public StateStore getGlobalStore(final String name) {
return globalStores.get(name);
}
// package-private for test only
void initializeStoreOffsetsFromCheckpoint(final boolean storeDirIsEmpty) {
try {
final Map<TopicPartition, Long> loadedCheckpoints = checkpointFile.read();
log.trace("Loaded offsets from the checkpoint file: {}", loadedCheckpoints);
for (final StateStoreMetadata store : stores.values()) {
if (store.corrupted) {
log.error("Tried to initialize store offsets for corrupted store {}", store);
throw new IllegalStateException("Should not initialize offsets for a corrupted task");
}
if (store.changelogPartition == null) {
log.info("State store {} is not logged and hence would not be restored", store.stateStore.name());
} else if (!store.stateStore.persistent()) {
log.info("Initializing to the starting offset for changelog {} of in-memory state store {}",
store.changelogPartition, store.stateStore.name());
} else if (store.offset() == null) {
if (loadedCheckpoints.containsKey(store.changelogPartition)) {
final Long offset = changelogOffsetFromCheckpointedOffset(loadedCheckpoints.remove(store.changelogPartition));
store.setOffset(offset);
log.debug("State store {} initialized from checkpoint with offset {} at changelog {}",
store.stateStore.name(), store.offset, store.changelogPartition);
} else {
// with EOS, if the previous run did not shutdown gracefully, we may lost the checkpoint file
// and hence we are uncertain that the current local state only contains committed data;
// in that case we need to treat it as a task-corrupted exception
if (eosEnabled && !storeDirIsEmpty) {
log.warn("State store {} did not find checkpoint offsets while stores are not empty, " +
"since under EOS it has the risk of getting uncommitted data in stores we have to " +
"treat it as a task corruption error and wipe out the local state of task {} " +
"before re-bootstrapping", store.stateStore.name(), taskId);
throw new TaskCorruptedException(Collections.singleton(taskId));
} else {
log.info("State store {} did not find checkpoint offset, hence would " +
"default to the starting offset at changelog {}",
store.stateStore.name(), store.changelogPartition);
}
}
} else {
loadedCheckpoints.remove(store.changelogPartition);
log.debug("Skipping re-initialization of offset from checkpoint for recycled store {}",
store.stateStore.name());
}
}
if (!loadedCheckpoints.isEmpty()) {
log.warn("Some loaded checkpoint offsets cannot find their corresponding state stores: {}", loadedCheckpoints);
}
if (eosEnabled) {
checkpointFile.delete();
}
} catch (final TaskCorruptedException e) {
throw e;
} catch (final IOException | RuntimeException e) {
// both IOException or runtime exception like number parsing can throw
throw new ProcessorStateException(format("%sError loading and deleting checkpoint file when creating the state manager",
logPrefix), e);
}
}
private void maybeRegisterStoreWithChangelogReader(final String storeName) {
if (isLoggingEnabled(storeName)) {
changelogReader.register(getStorePartition(storeName), this);
}
}
private List<TopicPartition> getAllChangelogTopicPartitions() {
final List<TopicPartition> allChangelogPartitions = new ArrayList<>();
for (final StateStoreMetadata storeMetadata : stores.values()) {
if (storeMetadata.changelogPartition != null) {
allChangelogPartitions.add(storeMetadata.changelogPartition);
}
}
return allChangelogPartitions;
}
@Override
public File baseDir() {
return baseDir;
}
@Override
public void registerStore(final StateStore store,
final StateRestoreCallback stateRestoreCallback,
final CommitCallback commitCallback) {
final String storeName = store.name();
// TODO (KAFKA-12887): we should not trigger user's exception handler for illegal-argument but always
// fail-crash; in this case we would not need to immediately close the state store before throwing
if (CHECKPOINT_FILE_NAME.equals(storeName)) {
store.close();
throw new IllegalArgumentException(format("%sIllegal store name: %s, which collides with the pre-defined " +
"checkpoint file name", logPrefix, storeName));
}
if (stores.containsKey(storeName)) {
store.close();
throw new IllegalArgumentException(format("%sStore %s has already been registered.", logPrefix, storeName));
}
if (stateRestoreCallback instanceof StateRestoreListener) {
log.warn("The registered state restore callback is also implementing the state restore listener interface, " +
"which is not expected and would be ignored");
}
final StateStoreMetadata storeMetadata = isLoggingEnabled(storeName) ?
new StateStoreMetadata(
store,
getStorePartition(storeName),
stateRestoreCallback,
commitCallback,
converterForStore(store)) :
new StateStoreMetadata(store, commitCallback);
// register the store first, so that if later an exception is thrown then eventually while we call `close`
// on the state manager this state store would be closed as well
stores.put(storeName, storeMetadata);
if (!stateUpdaterEnabled) {
maybeRegisterStoreWithChangelogReader(storeName);
}
log.debug("Registered state store {} to its state manager", storeName);
}
@Override
public StateStore getStore(final String name) {
if (stores.containsKey(name)) {
return stores.get(name).stateStore;
} else {
return null;
}
}
Set<TopicPartition> changelogPartitions() {
return Collections.unmodifiableSet(changelogOffsets().keySet());
}
void markChangelogAsCorrupted(final Collection<TopicPartition> partitions) {
final Collection<TopicPartition> partitionsToMarkAsCorrupted = new LinkedList<>(partitions);
for (final StateStoreMetadata storeMetadata : stores.values()) {
if (partitionsToMarkAsCorrupted.contains(storeMetadata.changelogPartition)) {
storeMetadata.corrupted = true;
partitionsToMarkAsCorrupted.remove(storeMetadata.changelogPartition);
}
}
if (!partitionsToMarkAsCorrupted.isEmpty()) {
throw new IllegalStateException("Some partitions " + partitionsToMarkAsCorrupted + " are not contained in " +
"the store list of task " + taskId + " marking as corrupted, this is not expected");
}
}
@Override
public Map<TopicPartition, Long> changelogOffsets() {
// return the current offsets for those logged stores
final Map<TopicPartition, Long> changelogOffsets = new HashMap<>();
for (final StateStoreMetadata storeMetadata : stores.values()) {
if (storeMetadata.changelogPartition != null) {
// for changelog whose offset is unknown, use 0L indicating earliest offset
// otherwise return the current offset + 1 as the next offset to fetch
changelogOffsets.put(
storeMetadata.changelogPartition,
storeMetadata.offset == null ? 0L : storeMetadata.offset + 1L);
}
}
return changelogOffsets;
}
TaskId taskId() {
return taskId;
}
// used by the changelog reader only
boolean changelogAsSource(final TopicPartition partition) {
return sourcePartitions.contains(partition);
}
@Override
public TaskType taskType() {
return taskType;
}
// used by the changelog reader only
StateStoreMetadata storeMetadata(final TopicPartition partition) {
for (final StateStoreMetadata storeMetadata : stores.values()) {
if (partition.equals(storeMetadata.changelogPartition)) {
return storeMetadata;
}
}
return null;
}
// used by the changelog reader only
void restore(final StateStoreMetadata storeMetadata, final List<ConsumerRecord<byte[], byte[]>> restoreRecords) {
if (!stores.containsValue(storeMetadata)) {
throw new IllegalStateException("Restoring " + storeMetadata + " which is not registered in this state manager, " +
"this should not happen.");
}
if (!restoreRecords.isEmpty()) {
// restore states from changelog records and update the snapshot offset as the batch end record's offset
final Long batchEndOffset = restoreRecords.get(restoreRecords.size() - 1).offset();
final RecordBatchingStateRestoreCallback restoreCallback = adapt(storeMetadata.restoreCallback);
final List<ConsumerRecord<byte[], byte[]>> convertedRecords = restoreRecords.stream()
.map(storeMetadata.recordConverter::convert)
.collect(Collectors.toList());
try {
restoreCallback.restoreBatch(convertedRecords);
} catch (final RuntimeException e) {
throw new ProcessorStateException(
format("%sException caught while trying to restore state from %s", logPrefix, storeMetadata.changelogPartition),
e
);
}
storeMetadata.setOffset(batchEndOffset);
}
}
/**
* @throws TaskMigratedException recoverable error sending changelog records that would cause the task to be removed
* @throws StreamsException fatal error when flushing the state store, for example sending changelog records failed
* or flushing state store get IO errors; such error should cause the thread to die
*/
@Override
public void flush() {
RuntimeException firstException = null;
// attempting to flush the stores
if (!stores.isEmpty()) {
log.debug("Flushing all stores registered in the state manager: {}", stores);
for (final StateStoreMetadata metadata : stores.values()) {
final StateStore store = metadata.stateStore;
log.trace("Flushing store {}", store.name());
try {
store.flush();
} catch (final RuntimeException exception) {
if (firstException == null) {
// do NOT wrap the error if it is actually caused by Streams itself
if (exception instanceof StreamsException)
firstException = exception;
else
firstException = new ProcessorStateException(
format("%sFailed to flush state store %s", logPrefix, store.name()), exception);
}
log.error("Failed to flush state store {}: ", store.name(), exception);
}
}
}
if (firstException != null) {
throw firstException;
}
}
public void flushCache() {
RuntimeException firstException = null;
// attempting to flush the stores
if (!stores.isEmpty()) {
log.debug("Flushing all store caches registered in the state manager: {}", stores);
for (final StateStoreMetadata metadata : stores.values()) {
final StateStore store = metadata.stateStore;
try {
// buffer should be flushed to send all records to changelog
if (store instanceof TimeOrderedKeyValueBuffer) {
store.flush();
} else if (store instanceof CachedStateStore) {
((CachedStateStore) store).flushCache();
}
log.trace("Flushed cache or buffer {}", store.name());
} catch (final RuntimeException exception) {
if (firstException == null) {
// do NOT wrap the error if it is actually caused by Streams itself
if (exception instanceof StreamsException) {
firstException = exception;
} else {
firstException = new ProcessorStateException(
format("%sFailed to flush cache of store %s", logPrefix, store.name()),
exception
);
}
}
log.error("Failed to flush cache of store {}: ", store.name(), exception);
}
}
}
if (firstException != null) {
throw firstException;
}
}
/**
* {@link StateStore#close() Close} all stores (even in case of failure).
* Log all exceptions and re-throw the first exception that occurred at the end.
*
* @throws ProcessorStateException if any error happens when closing the state stores
*/
@Override
public void close() throws ProcessorStateException {
log.debug("Closing its state manager and all the registered state stores: {}", stores);
if (!stateUpdaterEnabled) {
changelogReader.unregister(getAllChangelogTopicPartitions());
}
RuntimeException firstException = null;
// attempting to close the stores, just in case they
// are not closed by a ProcessorNode yet
if (!stores.isEmpty()) {
for (final Map.Entry<String, StateStoreMetadata> entry : stores.entrySet()) {
final StateStore store = entry.getValue().stateStore;
log.trace("Closing store {}", store.name());
try {
store.close();
} catch (final RuntimeException exception) {
if (firstException == null) {
// do NOT wrap the error if it is actually caused by Streams itself
if (exception instanceof StreamsException)
firstException = exception;
else
firstException = new ProcessorStateException(
format("%sFailed to close state store %s", logPrefix, store.name()), exception);
}
log.error("Failed to close state store {}: ", store.name(), exception);
}
}
stores.clear();
}
if (firstException != null) {
throw firstException;
}
}
/**
* Alternative to {@link #close()} that just resets the changelogs without closing any of the underlying state
* or unregistering the stores themselves
*/
void recycle() {
log.debug("Recycling state for {} task {}.", taskType, taskId);
if (!stateUpdaterEnabled) {
final List<TopicPartition> allChangelogs = getAllChangelogTopicPartitions();
changelogReader.unregister(allChangelogs);
}
}
void transitionTaskType(final TaskType newType) {
if (taskType.equals(newType)) {
throw new IllegalStateException("Tried to recycle state for task type conversion but new type was the same.");
}
taskType = newType;
}
@Override
public void updateChangelogOffsets(final Map<TopicPartition, Long> writtenOffsets) {
for (final Map.Entry<TopicPartition, Long> entry : writtenOffsets.entrySet()) {
final StateStoreMetadata store = findStore(entry.getKey());
if (store != null) {
store.setOffset(entry.getValue());
log.debug("State store {} updated to written offset {} at changelog {}",
store.stateStore.name(), store.offset, store.changelogPartition);
}
}
}
@Override
public void checkpoint() {
// checkpoint those stores that are only logged and persistent to the checkpoint file
final Map<TopicPartition, Long> checkpointingOffsets = new HashMap<>();
for (final StateStoreMetadata storeMetadata : stores.values()) {
if (storeMetadata.commitCallback != null && !storeMetadata.corrupted) {
try {
storeMetadata.commitCallback.onCommit();
} catch (final IOException e) {
throw new ProcessorStateException(
format("%sException caught while trying to checkpoint store, " +
"changelog partition %s", logPrefix, storeMetadata.changelogPartition),
e
);
}
}
// store is logged, persistent, not corrupted, and has a valid current offset
if (storeMetadata.changelogPartition != null &&
storeMetadata.stateStore.persistent() &&
!storeMetadata.corrupted) {
final long checkpointableOffset = checkpointableOffsetFromChangelogOffset(storeMetadata.offset);
checkpointingOffsets.put(storeMetadata.changelogPartition, checkpointableOffset);
}
}
log.debug("Writing checkpoint: {}", checkpointingOffsets);
try {
checkpointFile.write(checkpointingOffsets);
} catch (final IOException e) {
log.warn("Failed to write offset checkpoint file to [{}]." +
" This may occur if OS cleaned the state.dir in case when it located in ${java.io.tmpdir} directory." +
" This may also occur due to running multiple instances on the same machine using the same state dir." +
" Changing the location of state.dir may resolve the problem.",
checkpointFile, e);
}
}
private TopicPartition getStorePartition(final String storeName) {
// NOTE we assume the partition of the topic can always be inferred from the task id;
// if user ever use a custom partition grouper (deprecated in KIP-528) this would break and
// it is not a regression (it would always break anyways)
return new TopicPartition(changelogFor(storeName), taskId.partition());
}
private boolean isLoggingEnabled(final String storeName) {
// if the store name does not exist in the changelog map, it means the underlying store
// is not log enabled (including global stores)
return changelogFor(storeName) != null;
}
private StateStoreMetadata findStore(final TopicPartition changelogPartition) {
final List<StateStoreMetadata> found = stores.values().stream()
.filter(metadata -> changelogPartition.equals(metadata.changelogPartition))
.collect(Collectors.toList());
if (found.size() > 1) {
throw new IllegalStateException("Multiple state stores are found for changelog partition " + changelogPartition +
", this should never happen: " + found);
}
return found.isEmpty() ? null : found.get(0);
}
// Pass in a sentinel value to checkpoint when the changelog offset is not yet initialized/known
private long checkpointableOffsetFromChangelogOffset(final Long offset) {
return offset != null ? offset : OFFSET_UNKNOWN;
}
// Convert the written offsets in the checkpoint file back to the changelog offset
private Long changelogOffsetFromCheckpointedOffset(final long offset) {
return offset != OFFSET_UNKNOWN ? offset : null;
}
public TopicPartition registeredChangelogPartitionFor(final String storeName) {
final StateStoreMetadata storeMetadata = stores.get(storeName);
if (storeMetadata == null) {
throw new IllegalStateException("State store " + storeName
+ " for which the registered changelog partition should be"
+ " retrieved has not been registered"
);
}
if (storeMetadata.changelogPartition == null) {
throw new IllegalStateException("Registered state store " + storeName
+ " does not have a registered changelog partition."
+ " This may happen if logging is disabled for the state store."
);
}
return storeMetadata.changelogPartition;
}
public String changelogFor(final String storeName) {
return storeToChangelogTopic.get(storeName);
}
public void deleteCheckPointFileIfEOSEnabled() throws IOException {
if (eosEnabled) {
checkpointFile.delete();
}
}
}
相关信息
相关文章
kafka AbstractProcessorContext 源码
kafka AbstractReadOnlyDecorator 源码
kafka AbstractReadWriteDecorator 源码
0
赞
热门推荐
-
2、 - 优质文章
-
3、 gate.io
-
7、 golang
-
9、 openharmony
-
10、 Vue中input框自动聚焦