2 * Copyright (c) 2014 Cisco Systems, Inc. and others. All rights reserved.
4 * This program and the accompanying materials are made available under the
5 * terms of the Eclipse Public License v1.0 which accompanies this distribution,
6 * and is available at http://www.eclipse.org/legal/epl-v10.html
9 package org.opendaylight.controller.cluster.datastore;
11 import akka.actor.ActorRef;
12 import akka.actor.ActorSelection;
13 import akka.actor.Cancellable;
14 import akka.actor.PoisonPill;
15 import akka.actor.Props;
16 import akka.event.Logging;
17 import akka.event.LoggingAdapter;
18 import akka.japi.Creator;
19 import akka.persistence.RecoveryFailure;
20 import akka.serialization.Serialization;
21 import com.google.common.annotations.VisibleForTesting;
22 import com.google.common.base.Optional;
23 import com.google.common.base.Preconditions;
24 import com.google.common.collect.Lists;
25 import com.google.common.util.concurrent.FutureCallback;
26 import com.google.common.util.concurrent.Futures;
27 import com.google.common.util.concurrent.ListenableFuture;
28 import com.google.protobuf.ByteString;
29 import com.google.protobuf.InvalidProtocolBufferException;
30 import java.util.Collection;
31 import java.util.HashMap;
32 import java.util.List;
34 import java.util.concurrent.ExecutionException;
35 import java.util.concurrent.TimeUnit;
36 import javax.annotation.Nonnull;
37 import org.opendaylight.controller.cluster.DataPersistenceProvider;
38 import org.opendaylight.controller.cluster.common.actor.CommonConfig;
39 import org.opendaylight.controller.cluster.common.actor.MeteringBehavior;
40 import org.opendaylight.controller.cluster.datastore.ShardCommitCoordinator.CohortEntry;
41 import org.opendaylight.controller.cluster.datastore.compat.BackwardsCompatibleThreePhaseCommitCohort;
42 import org.opendaylight.controller.cluster.datastore.exceptions.NoShardLeaderException;
43 import org.opendaylight.controller.cluster.datastore.identifiers.ShardIdentifier;
44 import org.opendaylight.controller.cluster.datastore.identifiers.ShardTransactionIdentifier;
45 import org.opendaylight.controller.cluster.datastore.jmx.mbeans.shard.ShardMBeanFactory;
46 import org.opendaylight.controller.cluster.datastore.jmx.mbeans.shard.ShardStats;
47 import org.opendaylight.controller.cluster.datastore.messages.AbortTransaction;
48 import org.opendaylight.controller.cluster.datastore.messages.AbortTransactionReply;
49 import org.opendaylight.controller.cluster.datastore.messages.ActorInitialized;
50 import org.opendaylight.controller.cluster.datastore.messages.CanCommitTransaction;
51 import org.opendaylight.controller.cluster.datastore.messages.CloseTransactionChain;
52 import org.opendaylight.controller.cluster.datastore.messages.CommitTransaction;
53 import org.opendaylight.controller.cluster.datastore.messages.CommitTransactionReply;
54 import org.opendaylight.controller.cluster.datastore.messages.CreateTransaction;
55 import org.opendaylight.controller.cluster.datastore.messages.CreateTransactionReply;
56 import org.opendaylight.controller.cluster.datastore.messages.EnableNotification;
57 import org.opendaylight.controller.cluster.datastore.messages.ForwardedReadyTransaction;
58 import org.opendaylight.controller.cluster.datastore.messages.PeerAddressResolved;
59 import org.opendaylight.controller.cluster.datastore.messages.ReadData;
60 import org.opendaylight.controller.cluster.datastore.messages.ReadDataReply;
61 import org.opendaylight.controller.cluster.datastore.messages.ReadyTransactionReply;
62 import org.opendaylight.controller.cluster.datastore.messages.RegisterChangeListener;
63 import org.opendaylight.controller.cluster.datastore.messages.RegisterChangeListenerReply;
64 import org.opendaylight.controller.cluster.datastore.messages.UpdateSchemaContext;
65 import org.opendaylight.controller.cluster.datastore.modification.Modification;
66 import org.opendaylight.controller.cluster.datastore.modification.MutableCompositeModification;
67 import org.opendaylight.controller.cluster.datastore.node.NormalizedNodeToNodeCodec;
68 import org.opendaylight.controller.cluster.notifications.RoleChangeNotifier;
69 import org.opendaylight.controller.cluster.raft.RaftActor;
70 import org.opendaylight.controller.cluster.raft.ReplicatedLogEntry;
71 import org.opendaylight.controller.cluster.raft.base.messages.CaptureSnapshotReply;
72 import org.opendaylight.controller.cluster.raft.protobuff.client.messages.CompositeModificationByteStringPayload;
73 import org.opendaylight.controller.cluster.raft.protobuff.client.messages.CompositeModificationPayload;
74 import org.opendaylight.controller.cluster.raft.protobuff.client.messages.Payload;
75 import org.opendaylight.controller.md.sal.common.api.data.AsyncDataChangeListener;
76 import org.opendaylight.controller.md.sal.dom.store.impl.InMemoryDOMDataStore;
77 import org.opendaylight.controller.md.sal.dom.store.impl.InMemoryDOMDataStoreFactory;
78 import org.opendaylight.controller.protobuff.messages.common.NormalizedNodeMessages;
79 import org.opendaylight.controller.sal.core.spi.data.DOMStoreThreePhaseCommitCohort;
80 import org.opendaylight.controller.sal.core.spi.data.DOMStoreTransactionChain;
81 import org.opendaylight.controller.sal.core.spi.data.DOMStoreTransactionFactory;
82 import org.opendaylight.controller.sal.core.spi.data.DOMStoreWriteTransaction;
83 import org.opendaylight.yangtools.concepts.ListenerRegistration;
84 import org.opendaylight.yangtools.yang.data.api.YangInstanceIdentifier;
85 import org.opendaylight.yangtools.yang.data.api.schema.NormalizedNode;
86 import org.opendaylight.yangtools.yang.model.api.SchemaContext;
87 import scala.concurrent.duration.Duration;
88 import scala.concurrent.duration.FiniteDuration;
91 * A Shard represents a portion of the logical data tree <br/>
93 * Our Shard uses InMemoryDataStore as it's internal representation and delegates all requests it
96 public class Shard extends RaftActor {
98 private static final Object TX_COMMIT_TIMEOUT_CHECK_MESSAGE = "txCommitTimeoutCheck";
100 public static final String DEFAULT_NAME = "default";
102 // The state of this Shard
103 private final InMemoryDOMDataStore store;
105 private final LoggingAdapter LOG =
106 Logging.getLogger(getContext().system(), this);
108 /// The name of this shard
109 private final ShardIdentifier name;
111 private final ShardStats shardMBean;
113 private final List<ActorSelection> dataChangeListeners = Lists.newArrayList();
115 private final List<DelayedListenerRegistration> delayedListenerRegistrations =
116 Lists.newArrayList();
118 private final DatastoreContext datastoreContext;
120 private final DataPersistenceProvider dataPersistenceProvider;
122 private SchemaContext schemaContext;
124 private ActorRef createSnapshotTransaction;
126 private int createSnapshotTransactionCounter;
128 private final ShardCommitCoordinator commitCoordinator;
130 private final long transactionCommitTimeout;
132 private Cancellable txCommitTimeoutCheckSchedule;
134 private final Optional<ActorRef> roleChangeNotifier;
137 * Coordinates persistence recovery on startup.
139 private ShardRecoveryCoordinator recoveryCoordinator;
140 private List<Object> currentLogRecoveryBatch;
142 private final Map<String, DOMStoreTransactionChain> transactionChains = new HashMap<>();
144 protected Shard(final ShardIdentifier name, final Map<ShardIdentifier, String> peerAddresses,
145 final DatastoreContext datastoreContext, final SchemaContext schemaContext) {
146 super(name.toString(), mapPeerAddresses(peerAddresses),
147 Optional.of(datastoreContext.getShardRaftConfig()));
150 this.datastoreContext = datastoreContext;
151 this.schemaContext = schemaContext;
152 this.dataPersistenceProvider = (datastoreContext.isPersistent()) ? new PersistentDataProvider() : new NonPersistentRaftDataProvider();
154 LOG.info("Shard created : {} persistent : {}", name, datastoreContext.isPersistent());
156 store = InMemoryDOMDataStoreFactory.create(name.toString(), null,
157 datastoreContext.getDataStoreProperties());
159 if(schemaContext != null) {
160 store.onGlobalContextUpdated(schemaContext);
163 shardMBean = ShardMBeanFactory.getShardStatsMBean(name.toString(),
164 datastoreContext.getDataStoreMXBeanType());
165 shardMBean.setNotificationManager(store.getDataChangeListenerNotificationManager());
167 if (isMetricsCaptureEnabled()) {
168 getContext().become(new MeteringBehavior(this));
171 commitCoordinator = new ShardCommitCoordinator(TimeUnit.SECONDS.convert(1, TimeUnit.MINUTES),
172 datastoreContext.getShardTransactionCommitQueueCapacity());
174 transactionCommitTimeout = TimeUnit.MILLISECONDS.convert(
175 datastoreContext.getShardTransactionCommitTimeoutInSeconds(), TimeUnit.SECONDS);
177 // create a notifier actor for each cluster member
178 roleChangeNotifier = createRoleChangeNotifier(name.toString());
181 private static Map<String, String> mapPeerAddresses(
182 final Map<ShardIdentifier, String> peerAddresses) {
183 Map<String, String> map = new HashMap<>();
185 for (Map.Entry<ShardIdentifier, String> entry : peerAddresses
187 map.put(entry.getKey().toString(), entry.getValue());
193 public static Props props(final ShardIdentifier name,
194 final Map<ShardIdentifier, String> peerAddresses,
195 final DatastoreContext datastoreContext, final SchemaContext schemaContext) {
196 Preconditions.checkNotNull(name, "name should not be null");
197 Preconditions.checkNotNull(peerAddresses, "peerAddresses should not be null");
198 Preconditions.checkNotNull(datastoreContext, "dataStoreContext should not be null");
199 Preconditions.checkNotNull(schemaContext, "schemaContext should not be null");
201 return Props.create(new ShardCreator(name, peerAddresses, datastoreContext, schemaContext));
204 private Optional<ActorRef> createRoleChangeNotifier(String shardId) {
205 ActorRef shardRoleChangeNotifier = this.getContext().actorOf(
206 RoleChangeNotifier.getProps(shardId), shardId + "-notifier");
207 return Optional.<ActorRef>of(shardRoleChangeNotifier);
211 public void postStop() {
214 if(txCommitTimeoutCheckSchedule != null) {
215 txCommitTimeoutCheckSchedule.cancel();
220 public void onReceiveRecover(final Object message) throws Exception {
221 if(LOG.isDebugEnabled()) {
222 LOG.debug("onReceiveRecover: Received message {} from {}",
223 message.getClass().toString(),
227 if (message instanceof RecoveryFailure){
228 LOG.error(((RecoveryFailure) message).cause(), "Recovery failed because of this cause");
230 // Even though recovery failed, we still need to finish our recovery, eg send the
231 // ActorInitialized message and start the txCommitTimeoutCheckSchedule.
232 onRecoveryComplete();
234 super.onReceiveRecover(message);
239 public void onReceiveCommand(final Object message) throws Exception {
240 if(LOG.isDebugEnabled()) {
241 LOG.debug("onReceiveCommand: Received message {} from {}", message, getSender());
244 if(message.getClass().equals(ReadDataReply.SERIALIZABLE_CLASS)) {
245 handleReadDataReply(message);
246 } else if (message.getClass().equals(CreateTransaction.SERIALIZABLE_CLASS)) {
247 handleCreateTransaction(message);
248 } else if(message instanceof ForwardedReadyTransaction) {
249 handleForwardedReadyTransaction((ForwardedReadyTransaction)message);
250 } else if(message.getClass().equals(CanCommitTransaction.SERIALIZABLE_CLASS)) {
251 handleCanCommitTransaction(CanCommitTransaction.fromSerializable(message));
252 } else if(message.getClass().equals(CommitTransaction.SERIALIZABLE_CLASS)) {
253 handleCommitTransaction(CommitTransaction.fromSerializable(message));
254 } else if(message.getClass().equals(AbortTransaction.SERIALIZABLE_CLASS)) {
255 handleAbortTransaction(AbortTransaction.fromSerializable(message));
256 } else if (message.getClass().equals(CloseTransactionChain.SERIALIZABLE_CLASS)){
257 closeTransactionChain(CloseTransactionChain.fromSerializable(message));
258 } else if (message instanceof RegisterChangeListener) {
259 registerChangeListener((RegisterChangeListener) message);
260 } else if (message instanceof UpdateSchemaContext) {
261 updateSchemaContext((UpdateSchemaContext) message);
262 } else if (message instanceof PeerAddressResolved) {
263 PeerAddressResolved resolved = (PeerAddressResolved) message;
264 setPeerAddress(resolved.getPeerId().toString(),
265 resolved.getPeerAddress());
266 } else if(message.equals(TX_COMMIT_TIMEOUT_CHECK_MESSAGE)) {
267 handleTransactionCommitTimeoutCheck();
269 super.onReceiveCommand(message);
274 protected Optional<ActorRef> getRoleChangeNotifier() {
275 return roleChangeNotifier;
278 private void handleTransactionCommitTimeoutCheck() {
279 CohortEntry cohortEntry = commitCoordinator.getCurrentCohortEntry();
280 if(cohortEntry != null) {
281 long elapsed = System.currentTimeMillis() - cohortEntry.getLastAccessTime();
282 if(elapsed > transactionCommitTimeout) {
283 LOG.warning("Current transaction {} has timed out after {} ms - aborting",
284 cohortEntry.getTransactionID(), transactionCommitTimeout);
286 doAbortTransaction(cohortEntry.getTransactionID(), null);
291 private void handleCommitTransaction(final CommitTransaction commit) {
292 final String transactionID = commit.getTransactionID();
294 LOG.debug("Committing transaction {}", transactionID);
296 // Get the current in-progress cohort entry in the commitCoordinator if it corresponds to
298 final CohortEntry cohortEntry = commitCoordinator.getCohortEntryIfCurrent(transactionID);
299 if(cohortEntry == null) {
300 // We're not the current Tx - the Tx was likely expired b/c it took too long in
301 // between the canCommit and commit messages.
302 IllegalStateException ex = new IllegalStateException(
303 String.format("Cannot commit transaction %s - it is not the current transaction",
305 LOG.error(ex.getMessage());
306 shardMBean.incrementFailedTransactionsCount();
307 getSender().tell(new akka.actor.Status.Failure(ex), getSelf());
311 // We perform the preCommit phase here atomically with the commit phase. This is an
312 // optimization to eliminate the overhead of an extra preCommit message. We lose front-end
313 // coordination of preCommit across shards in case of failure but preCommit should not
314 // normally fail since we ensure only one concurrent 3-phase commit.
317 // We block on the future here so we don't have to worry about possibly accessing our
318 // state on a different thread outside of our dispatcher. Also, the data store
319 // currently uses a same thread executor anyway.
320 cohortEntry.getCohort().preCommit().get();
322 // If we do not have any followers and we are not using persistence we can
323 // apply modification to the state immediately
324 if(!hasFollowers() && !persistence().isRecoveryApplicable()){
325 applyModificationToState(getSender(), transactionID, cohortEntry.getModification());
327 Shard.this.persistData(getSender(), transactionID,
328 new CompositeModificationByteStringPayload(cohortEntry.getModification().toSerializable()));
330 } catch (InterruptedException | ExecutionException e) {
331 LOG.error(e, "An exception occurred while preCommitting transaction {}",
332 cohortEntry.getTransactionID());
333 shardMBean.incrementFailedTransactionsCount();
334 getSender().tell(new akka.actor.Status.Failure(e), getSelf());
337 cohortEntry.updateLastAccessTime();
340 private void finishCommit(@Nonnull final ActorRef sender, final @Nonnull String transactionID) {
341 // With persistence enabled, this method is called via applyState by the leader strategy
342 // after the commit has been replicated to a majority of the followers.
344 CohortEntry cohortEntry = commitCoordinator.getCohortEntryIfCurrent(transactionID);
345 if(cohortEntry == null) {
346 // The transaction is no longer the current commit. This can happen if the transaction
347 // was aborted prior, most likely due to timeout in the front-end. We need to finish
348 // committing the transaction though since it was successfully persisted and replicated
349 // however we can't use the original cohort b/c it was already preCommitted and may
350 // conflict with the current commit or may have been aborted so we commit with a new
352 cohortEntry = commitCoordinator.getAndRemoveCohortEntry(transactionID);
353 if(cohortEntry != null) {
354 commitWithNewTransaction(cohortEntry.getModification());
355 sender.tell(CommitTransactionReply.INSTANCE.toSerializable(), getSelf());
357 // This really shouldn't happen - it likely means that persistence or replication
358 // took so long to complete such that the cohort entry was expired from the cache.
359 IllegalStateException ex = new IllegalStateException(
360 String.format("Could not finish committing transaction %s - no CohortEntry found",
362 LOG.error(ex.getMessage());
363 sender.tell(new akka.actor.Status.Failure(ex), getSelf());
369 LOG.debug("Finishing commit for transaction {}", cohortEntry.getTransactionID());
372 // We block on the future here so we don't have to worry about possibly accessing our
373 // state on a different thread outside of our dispatcher. Also, the data store
374 // currently uses a same thread executor anyway.
375 cohortEntry.getCohort().commit().get();
377 sender.tell(CommitTransactionReply.INSTANCE.toSerializable(), getSelf());
379 shardMBean.incrementCommittedTransactionCount();
380 shardMBean.setLastCommittedTransactionTime(System.currentTimeMillis());
382 } catch (InterruptedException | ExecutionException e) {
383 sender.tell(new akka.actor.Status.Failure(e), getSelf());
385 LOG.error(e, "An exception occurred while committing transaction {}", transactionID);
386 shardMBean.incrementFailedTransactionsCount();
389 commitCoordinator.currentTransactionComplete(transactionID, true);
392 private void handleCanCommitTransaction(final CanCommitTransaction canCommit) {
393 LOG.debug("Can committing transaction {}", canCommit.getTransactionID());
394 commitCoordinator.handleCanCommit(canCommit, getSender(), self());
397 private void handleForwardedReadyTransaction(ForwardedReadyTransaction ready) {
398 LOG.debug("Readying transaction {}, client version {}", ready.getTransactionID(),
399 ready.getTxnClientVersion());
401 // This message is forwarded by the ShardTransaction on ready. We cache the cohort in the
402 // commitCoordinator in preparation for the subsequent three phase commit initiated by
404 commitCoordinator.transactionReady(ready.getTransactionID(), ready.getCohort(),
405 ready.getModification());
407 // Return our actor path as we'll handle the three phase commit, except if the Tx client
408 // version < 1 (Helium-1 version). This means the Tx was initiated by a base Helium version
409 // node. In that case, the subsequent 3-phase commit messages won't contain the
410 // transactionId so to maintain backwards compatibility, we create a separate cohort actor
411 // to provide the compatible behavior.
412 ActorRef replyActorPath = self();
413 if(ready.getTxnClientVersion() < DataStoreVersions.HELIUM_1_VERSION) {
414 LOG.debug("Creating BackwardsCompatibleThreePhaseCommitCohort");
415 replyActorPath = getContext().actorOf(BackwardsCompatibleThreePhaseCommitCohort.props(
416 ready.getTransactionID()));
419 ReadyTransactionReply readyTransactionReply = new ReadyTransactionReply(
420 Serialization.serializedActorPath(replyActorPath));
421 getSender().tell(ready.isReturnSerialized() ? readyTransactionReply.toSerializable() :
422 readyTransactionReply, getSelf());
425 private void handleAbortTransaction(final AbortTransaction abort) {
426 doAbortTransaction(abort.getTransactionID(), getSender());
429 void doAbortTransaction(final String transactionID, final ActorRef sender) {
430 final CohortEntry cohortEntry = commitCoordinator.getCohortEntryIfCurrent(transactionID);
431 if(cohortEntry != null) {
432 LOG.debug("Aborting transaction {}", transactionID);
434 // We don't remove the cached cohort entry here (ie pass false) in case the Tx was
435 // aborted during replication in which case we may still commit locally if replication
437 commitCoordinator.currentTransactionComplete(transactionID, false);
439 final ListenableFuture<Void> future = cohortEntry.getCohort().abort();
440 final ActorRef self = getSelf();
442 Futures.addCallback(future, new FutureCallback<Void>() {
444 public void onSuccess(final Void v) {
445 shardMBean.incrementAbortTransactionsCount();
448 sender.tell(AbortTransactionReply.INSTANCE.toSerializable(), self);
453 public void onFailure(final Throwable t) {
454 LOG.error(t, "An exception happened during abort");
457 sender.tell(new akka.actor.Status.Failure(t), self);
464 private void handleCreateTransaction(final Object message) {
466 createTransaction(CreateTransaction.fromSerializable(message));
467 } else if (getLeader() != null) {
468 getLeader().forward(message, getContext());
470 getSender().tell(new akka.actor.Status.Failure(new NoShardLeaderException(
471 "Could not find shard leader so transaction cannot be created. This typically happens" +
472 " when the system is coming up or recovering and a leader is being elected. Try again" +
473 " later.")), getSelf());
477 private void handleReadDataReply(final Object message) {
478 // This must be for install snapshot. Don't want to open this up and trigger
481 self().tell(new CaptureSnapshotReply(ReadDataReply.fromSerializableAsByteString(message)),
484 createSnapshotTransaction = null;
486 // Send a PoisonPill instead of sending close transaction because we do not really need
488 getSender().tell(PoisonPill.getInstance(), self());
491 private void closeTransactionChain(final CloseTransactionChain closeTransactionChain) {
492 DOMStoreTransactionChain chain =
493 transactionChains.remove(closeTransactionChain.getTransactionChainId());
500 private ActorRef createTypedTransactionActor(int transactionType,
501 ShardTransactionIdentifier transactionId, String transactionChainId,
502 short clientVersion ) {
504 DOMStoreTransactionFactory factory = store;
506 if(!transactionChainId.isEmpty()) {
507 factory = transactionChains.get(transactionChainId);
509 DOMStoreTransactionChain transactionChain = store.createTransactionChain();
510 transactionChains.put(transactionChainId, transactionChain);
511 factory = transactionChain;
515 if(this.schemaContext == null) {
516 throw new IllegalStateException("SchemaContext is not set");
519 if (transactionType == TransactionProxy.TransactionType.READ_ONLY.ordinal()) {
521 shardMBean.incrementReadOnlyTransactionCount();
523 return getContext().actorOf(
524 ShardTransaction.props(factory.newReadOnlyTransaction(), getSelf(),
525 schemaContext,datastoreContext, shardMBean,
526 transactionId.getRemoteTransactionId(), clientVersion),
527 transactionId.toString());
529 } else if (transactionType == TransactionProxy.TransactionType.READ_WRITE.ordinal()) {
531 shardMBean.incrementReadWriteTransactionCount();
533 return getContext().actorOf(
534 ShardTransaction.props(factory.newReadWriteTransaction(), getSelf(),
535 schemaContext, datastoreContext, shardMBean,
536 transactionId.getRemoteTransactionId(), clientVersion),
537 transactionId.toString());
540 } else if (transactionType == TransactionProxy.TransactionType.WRITE_ONLY.ordinal()) {
542 shardMBean.incrementWriteOnlyTransactionCount();
544 return getContext().actorOf(
545 ShardTransaction.props(factory.newWriteOnlyTransaction(), getSelf(),
546 schemaContext, datastoreContext, shardMBean,
547 transactionId.getRemoteTransactionId(), clientVersion),
548 transactionId.toString());
550 throw new IllegalArgumentException(
551 "Shard="+name + ":CreateTransaction message has unidentified transaction type="
556 private void createTransaction(CreateTransaction createTransaction) {
558 ActorRef transactionActor = createTransaction(createTransaction.getTransactionType(),
559 createTransaction.getTransactionId(), createTransaction.getTransactionChainId(),
560 createTransaction.getVersion());
562 getSender().tell(new CreateTransactionReply(Serialization.serializedActorPath(transactionActor),
563 createTransaction.getTransactionId()).toSerializable(), getSelf());
564 } catch (Exception e) {
565 getSender().tell(new akka.actor.Status.Failure(e), getSelf());
569 private ActorRef createTransaction(int transactionType, String remoteTransactionId,
570 String transactionChainId, short clientVersion) {
572 ShardTransactionIdentifier transactionId =
573 ShardTransactionIdentifier.builder()
574 .remoteTransactionId(remoteTransactionId)
577 if(LOG.isDebugEnabled()) {
578 LOG.debug("Creating transaction : {} ", transactionId);
581 ActorRef transactionActor = createTypedTransactionActor(transactionType, transactionId,
582 transactionChainId, clientVersion);
584 return transactionActor;
587 private void syncCommitTransaction(final DOMStoreWriteTransaction transaction)
588 throws ExecutionException, InterruptedException {
589 DOMStoreThreePhaseCommitCohort commitCohort = transaction.ready();
590 commitCohort.preCommit().get();
591 commitCohort.commit().get();
594 private void commitWithNewTransaction(final Modification modification) {
595 DOMStoreWriteTransaction tx = store.newWriteOnlyTransaction();
596 modification.apply(tx);
598 syncCommitTransaction(tx);
599 shardMBean.incrementCommittedTransactionCount();
600 shardMBean.setLastCommittedTransactionTime(System.currentTimeMillis());
601 } catch (InterruptedException | ExecutionException e) {
602 shardMBean.incrementFailedTransactionsCount();
603 LOG.error(e, "Failed to commit");
607 private void updateSchemaContext(final UpdateSchemaContext message) {
608 this.schemaContext = message.getSchemaContext();
609 updateSchemaContext(message.getSchemaContext());
610 store.onGlobalContextUpdated(message.getSchemaContext());
614 void updateSchemaContext(final SchemaContext schemaContext) {
615 store.onGlobalContextUpdated(schemaContext);
618 private void registerChangeListener(final RegisterChangeListener registerChangeListener) {
620 LOG.debug("registerDataChangeListener for {}", registerChangeListener.getPath());
622 ListenerRegistration<AsyncDataChangeListener<YangInstanceIdentifier,
623 NormalizedNode<?, ?>>> registration;
625 registration = doChangeListenerRegistration(registerChangeListener);
627 LOG.debug("Shard is not the leader - delaying registration");
629 DelayedListenerRegistration delayedReg =
630 new DelayedListenerRegistration(registerChangeListener);
631 delayedListenerRegistrations.add(delayedReg);
632 registration = delayedReg;
635 ActorRef listenerRegistration = getContext().actorOf(
636 DataChangeListenerRegistration.props(registration));
638 LOG.debug("registerDataChangeListener sending reply, listenerRegistrationPath = {} ",
639 listenerRegistration.path());
641 getSender().tell(new RegisterChangeListenerReply(listenerRegistration.path()), getSelf());
644 private ListenerRegistration<AsyncDataChangeListener<YangInstanceIdentifier,
645 NormalizedNode<?, ?>>> doChangeListenerRegistration(
646 final RegisterChangeListener registerChangeListener) {
648 ActorSelection dataChangeListenerPath = getContext().system().actorSelection(
649 registerChangeListener.getDataChangeListenerPath());
651 // Notify the listener if notifications should be enabled or not
652 // If this shard is the leader then it will enable notifications else
654 dataChangeListenerPath.tell(new EnableNotification(true), getSelf());
656 // Now store a reference to the data change listener so it can be notified
657 // at a later point if notifications should be enabled or disabled
658 dataChangeListeners.add(dataChangeListenerPath);
660 AsyncDataChangeListener<YangInstanceIdentifier, NormalizedNode<?, ?>> listener =
661 new DataChangeListenerProxy(dataChangeListenerPath);
663 LOG.debug("Registering for path {}", registerChangeListener.getPath());
665 return store.registerChangeListener(registerChangeListener.getPath(), listener,
666 registerChangeListener.getScope());
669 private boolean isMetricsCaptureEnabled(){
670 CommonConfig config = new CommonConfig(getContext().system().settings().config());
671 return config.isMetricCaptureEnabled();
676 void startLogRecoveryBatch(final int maxBatchSize) {
677 currentLogRecoveryBatch = Lists.newArrayListWithCapacity(maxBatchSize);
679 if(LOG.isDebugEnabled()) {
680 LOG.debug("{} : starting log recovery batch with max size {}", persistenceId(), maxBatchSize);
685 protected void appendRecoveredLogEntry(final Payload data) {
686 if (data instanceof CompositeModificationPayload) {
687 currentLogRecoveryBatch.add(((CompositeModificationPayload) data).getModification());
688 } else if (data instanceof CompositeModificationByteStringPayload) {
689 currentLogRecoveryBatch.add(((CompositeModificationByteStringPayload) data).getModification());
691 LOG.error("Unknown state received {} during recovery", data);
696 protected void applyRecoverySnapshot(final ByteString snapshot) {
697 if(recoveryCoordinator == null) {
698 recoveryCoordinator = new ShardRecoveryCoordinator(persistenceId(), schemaContext);
701 recoveryCoordinator.submit(snapshot, store.newWriteOnlyTransaction());
703 if(LOG.isDebugEnabled()) {
704 LOG.debug("{} : submitted recovery sbapshot", persistenceId());
709 protected void applyCurrentLogRecoveryBatch() {
710 if(recoveryCoordinator == null) {
711 recoveryCoordinator = new ShardRecoveryCoordinator(persistenceId(), schemaContext);
714 recoveryCoordinator.submit(currentLogRecoveryBatch, store.newWriteOnlyTransaction());
716 if(LOG.isDebugEnabled()) {
717 LOG.debug("{} : submitted log recovery batch with size {}", persistenceId(),
718 currentLogRecoveryBatch.size());
723 protected void onRecoveryComplete() {
724 if(recoveryCoordinator != null) {
725 Collection<DOMStoreWriteTransaction> txList = recoveryCoordinator.getTransactions();
727 if(LOG.isDebugEnabled()) {
728 LOG.debug("{} : recovery complete - committing {} Tx's", persistenceId(), txList.size());
731 for(DOMStoreWriteTransaction tx: txList) {
733 syncCommitTransaction(tx);
734 shardMBean.incrementCommittedTransactionCount();
735 } catch (InterruptedException | ExecutionException e) {
736 shardMBean.incrementFailedTransactionsCount();
737 LOG.error(e, "Failed to commit");
742 recoveryCoordinator = null;
743 currentLogRecoveryBatch = null;
744 updateJournalStats();
746 //notify shard manager
747 getContext().parent().tell(new ActorInitialized(), getSelf());
749 // Being paranoid here - this method should only be called once but just in case...
750 if(txCommitTimeoutCheckSchedule == null) {
751 // Schedule a message to be periodically sent to check if the current in-progress
752 // transaction should be expired and aborted.
753 FiniteDuration period = Duration.create(transactionCommitTimeout / 3, TimeUnit.MILLISECONDS);
754 txCommitTimeoutCheckSchedule = getContext().system().scheduler().schedule(
755 period, period, getSelf(),
756 TX_COMMIT_TIMEOUT_CHECK_MESSAGE, getContext().dispatcher(), ActorRef.noSender());
761 protected void applyState(final ActorRef clientActor, final String identifier, final Object data) {
763 if (data instanceof CompositeModificationPayload) {
764 Object modification = ((CompositeModificationPayload) data).getModification();
766 applyModificationToState(clientActor, identifier, modification);
767 } else if(data instanceof CompositeModificationByteStringPayload ){
768 Object modification = ((CompositeModificationByteStringPayload) data).getModification();
770 applyModificationToState(clientActor, identifier, modification);
773 LOG.error("Unknown state received {} Class loader = {} CompositeNodeMod.ClassLoader = {}",
774 data, data.getClass().getClassLoader(),
775 CompositeModificationPayload.class.getClassLoader());
778 updateJournalStats();
782 private void applyModificationToState(ActorRef clientActor, String identifier, Object modification) {
783 if(modification == null) {
785 "modification is null - this is very unexpected, clientActor = {}, identifier = {}",
786 identifier, clientActor != null ? clientActor.path().toString() : null);
787 } else if(clientActor == null) {
788 // There's no clientActor to which to send a commit reply so we must be applying
789 // replicated state from the leader.
790 commitWithNewTransaction(MutableCompositeModification.fromSerializable(
791 modification, schemaContext));
793 // This must be the OK to commit after replication consensus.
794 finishCommit(clientActor, identifier);
798 private void updateJournalStats() {
799 ReplicatedLogEntry lastLogEntry = getLastLogEntry();
801 if (lastLogEntry != null) {
802 shardMBean.setLastLogIndex(lastLogEntry.getIndex());
803 shardMBean.setLastLogTerm(lastLogEntry.getTerm());
806 shardMBean.setCommitIndex(getCommitIndex());
807 shardMBean.setLastApplied(getLastApplied());
808 shardMBean.setInMemoryJournalDataSize(getRaftActorContext().getReplicatedLog().dataSize());
812 protected void createSnapshot() {
813 if (createSnapshotTransaction == null) {
815 // Create a transaction. We are really going to treat the transaction as a worker
816 // so that this actor does not get block building the snapshot
817 createSnapshotTransaction = createTransaction(
818 TransactionProxy.TransactionType.READ_ONLY.ordinal(),
819 "createSnapshot" + ++createSnapshotTransactionCounter, "",
820 DataStoreVersions.CURRENT_VERSION);
822 createSnapshotTransaction.tell(
823 new ReadData(YangInstanceIdentifier.builder().build()).toSerializable(), self());
830 protected void applySnapshot(final ByteString snapshot) {
831 // Since this will be done only on Recovery or when this actor is a Follower
832 // we can safely commit everything in here. We not need to worry about event notifications
833 // as they would have already been disabled on the follower
835 LOG.info("Applying snapshot");
837 DOMStoreWriteTransaction transaction = store.newWriteOnlyTransaction();
838 NormalizedNodeMessages.Node serializedNode = NormalizedNodeMessages.Node.parseFrom(snapshot);
839 NormalizedNode<?, ?> node = new NormalizedNodeToNodeCodec(schemaContext)
840 .decode(serializedNode);
842 // delete everything first
843 transaction.delete(YangInstanceIdentifier.builder().build());
845 // Add everything from the remote node back
846 transaction.write(YangInstanceIdentifier.builder().build(), node);
847 syncCommitTransaction(transaction);
848 } catch (InvalidProtocolBufferException | InterruptedException | ExecutionException e) {
849 LOG.error(e, "An exception occurred when applying snapshot");
851 LOG.info("Done applying snapshot");
856 protected void onStateChanged() {
857 boolean isLeader = isLeader();
858 for (ActorSelection dataChangeListener : dataChangeListeners) {
859 dataChangeListener.tell(new EnableNotification(isLeader), getSelf());
863 for(DelayedListenerRegistration reg: delayedListenerRegistrations) {
864 if(!reg.isClosed()) {
865 reg.setDelegate(doChangeListenerRegistration(reg.getRegisterChangeListener()));
869 delayedListenerRegistrations.clear();
872 shardMBean.setRaftState(getRaftState().name());
873 shardMBean.setCurrentTerm(getCurrentTerm());
875 // If this actor is no longer the leader close all the transaction chains
877 for(Map.Entry<String, DOMStoreTransactionChain> entry : transactionChains.entrySet()){
878 if(LOG.isDebugEnabled()) {
880 "onStateChanged: Closing transaction chain {} because shard {} is no longer the leader",
881 entry.getKey(), getId());
883 entry.getValue().close();
886 transactionChains.clear();
891 protected DataPersistenceProvider persistence() {
892 return dataPersistenceProvider;
895 @Override protected void onLeaderChanged(final String oldLeader, final String newLeader) {
896 shardMBean.setLeader(newLeader);
899 @Override public String persistenceId() {
900 return this.name.toString();
904 DataPersistenceProvider getDataPersistenceProvider() {
905 return dataPersistenceProvider;
908 private static class ShardCreator implements Creator<Shard> {
910 private static final long serialVersionUID = 1L;
912 final ShardIdentifier name;
913 final Map<ShardIdentifier, String> peerAddresses;
914 final DatastoreContext datastoreContext;
915 final SchemaContext schemaContext;
917 ShardCreator(final ShardIdentifier name, final Map<ShardIdentifier, String> peerAddresses,
918 final DatastoreContext datastoreContext, final SchemaContext schemaContext) {
920 this.peerAddresses = peerAddresses;
921 this.datastoreContext = datastoreContext;
922 this.schemaContext = schemaContext;
926 public Shard create() throws Exception {
927 return new Shard(name, peerAddresses, datastoreContext, schemaContext);
932 InMemoryDOMDataStore getDataStore() {
937 ShardStats getShardMBean() {
941 private static class DelayedListenerRegistration implements
942 ListenerRegistration<AsyncDataChangeListener<YangInstanceIdentifier, NormalizedNode<?, ?>>> {
944 private volatile boolean closed;
946 private final RegisterChangeListener registerChangeListener;
948 private volatile ListenerRegistration<AsyncDataChangeListener<YangInstanceIdentifier,
949 NormalizedNode<?, ?>>> delegate;
951 DelayedListenerRegistration(final RegisterChangeListener registerChangeListener) {
952 this.registerChangeListener = registerChangeListener;
955 void setDelegate( final ListenerRegistration<AsyncDataChangeListener<YangInstanceIdentifier,
956 NormalizedNode<?, ?>>> registration) {
957 this.delegate = registration;
964 RegisterChangeListener getRegisterChangeListener() {
965 return registerChangeListener;
969 public AsyncDataChangeListener<YangInstanceIdentifier, NormalizedNode<?, ?>> getInstance() {
970 return delegate != null ? delegate.getInstance() : null;
974 public void close() {
976 if(delegate != null) {