opendaylight/md-sal/sal-distributed-datastore/src/main/java/org/opendaylight/controller/cluster/datastore/Shard.java

   1 /*
   2  * Copyright (c) 2014 Cisco Systems, Inc. and others.  All rights reserved.
   3  *
   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
   7  */
   8
   9 package org.opendaylight.controller.cluster.datastore;
  10
  11 import akka.actor.ActorRef;
  12 import akka.actor.ActorSelection;
  13 import akka.actor.Cancellable;
  14 import akka.actor.Props;
  15 import akka.japi.Creator;
  16 import akka.persistence.RecoveryFailure;
  17 import akka.serialization.Serialization;
  18 import com.google.common.annotations.VisibleForTesting;
  19 import com.google.common.base.Optional;
  20 import com.google.common.base.Preconditions;
  21 import com.google.common.util.concurrent.FutureCallback;
  22 import com.google.common.util.concurrent.Futures;
  23 import com.google.common.util.concurrent.ListenableFuture;
  24 import java.io.IOException;
  25 import java.util.HashMap;
  26 import java.util.Map;
  27 import java.util.concurrent.ExecutionException;
  28 import java.util.concurrent.TimeUnit;
  29 import javax.annotation.Nonnull;
  30 import org.opendaylight.controller.cluster.common.actor.CommonConfig;
  31 import org.opendaylight.controller.cluster.common.actor.MeteringBehavior;
  32 import org.opendaylight.controller.cluster.datastore.ShardCommitCoordinator.CohortEntry;
  33 import org.opendaylight.controller.cluster.datastore.exceptions.NoShardLeaderException;
  34 import org.opendaylight.controller.cluster.datastore.identifiers.ShardIdentifier;
  35 import org.opendaylight.controller.cluster.datastore.identifiers.ShardTransactionIdentifier;
  36 import org.opendaylight.controller.cluster.datastore.jmx.mbeans.shard.ShardMBeanFactory;
  37 import org.opendaylight.controller.cluster.datastore.jmx.mbeans.shard.ShardStats;
  38 import org.opendaylight.controller.cluster.datastore.messages.AbortTransaction;
  39 import org.opendaylight.controller.cluster.datastore.messages.AbortTransactionReply;
  40 import org.opendaylight.controller.cluster.datastore.messages.ActorInitialized;
  41 import org.opendaylight.controller.cluster.datastore.messages.BatchedModifications;
  42 import org.opendaylight.controller.cluster.datastore.messages.CanCommitTransaction;
  43 import org.opendaylight.controller.cluster.datastore.messages.CloseTransactionChain;
  44 import org.opendaylight.controller.cluster.datastore.messages.CommitTransaction;
  45 import org.opendaylight.controller.cluster.datastore.messages.CommitTransactionReply;
  46 import org.opendaylight.controller.cluster.datastore.messages.CreateTransaction;
  47 import org.opendaylight.controller.cluster.datastore.messages.CreateTransactionReply;
  48 import org.opendaylight.controller.cluster.datastore.messages.ForwardedReadyTransaction;
  49 import org.opendaylight.controller.cluster.datastore.messages.PeerAddressResolved;
  50 import org.opendaylight.controller.cluster.datastore.messages.RegisterChangeListener;
  51 import org.opendaylight.controller.cluster.datastore.messages.RegisterDataTreeChangeListener;
  52 import org.opendaylight.controller.cluster.datastore.messages.UpdateSchemaContext;
  53 import org.opendaylight.controller.cluster.datastore.modification.Modification;
  54 import org.opendaylight.controller.cluster.datastore.modification.ModificationPayload;
  55 import org.opendaylight.controller.cluster.datastore.modification.MutableCompositeModification;
  56 import org.opendaylight.controller.cluster.datastore.utils.Dispatchers;
  57 import org.opendaylight.controller.cluster.datastore.utils.MessageTracker;
  58 import org.opendaylight.controller.cluster.notifications.RegisterRoleChangeListener;
  59 import org.opendaylight.controller.cluster.notifications.RoleChangeNotifier;
  60 import org.opendaylight.controller.cluster.raft.RaftActor;
  61 import org.opendaylight.controller.cluster.raft.RaftActorRecoveryCohort;
  62 import org.opendaylight.controller.cluster.raft.RaftActorSnapshotCohort;
  63 import org.opendaylight.controller.cluster.raft.base.messages.FollowerInitialSyncUpStatus;
  64 import org.opendaylight.controller.cluster.raft.messages.AppendEntriesReply;
  65 import org.opendaylight.controller.cluster.raft.protobuff.client.messages.CompositeModificationByteStringPayload;
  66 import org.opendaylight.controller.cluster.raft.protobuff.client.messages.CompositeModificationPayload;
  67 import org.opendaylight.controller.md.sal.dom.store.impl.InMemoryDOMDataStore;
  68 import org.opendaylight.controller.md.sal.dom.store.impl.InMemoryDOMDataStoreFactory;
  69 import org.opendaylight.controller.sal.core.spi.data.DOMStoreWriteTransaction;
  70 import org.opendaylight.yangtools.yang.model.api.SchemaContext;
  71 import scala.concurrent.duration.Duration;
  72 import scala.concurrent.duration.FiniteDuration;
  73
  74 /**
  75  * A Shard represents a portion of the logical data tree <br/>
  76  * <p>
  77  * Our Shard uses InMemoryDataStore as it's internal representation and delegates all requests it
  78  * </p>
  79  */
  80 public class Shard extends RaftActor {
  81
  82     private static final Object TX_COMMIT_TIMEOUT_CHECK_MESSAGE = "txCommitTimeoutCheck";
  83
  84     @VisibleForTesting
  85     static final String DEFAULT_NAME = "default";
  86
  87     // The state of this Shard
  88     private final InMemoryDOMDataStore store;
  89
  90     /// The name of this shard
  91     private final String name;
  92
  93     private final ShardStats shardMBean;
  94
  95     private DatastoreContext datastoreContext;
  96
  97     private final ShardCommitCoordinator commitCoordinator;
  98
  99     private long transactionCommitTimeout;
 100
 101     private Cancellable txCommitTimeoutCheckSchedule;
 102
 103     private final Optional<ActorRef> roleChangeNotifier;
 104
 105     private final MessageTracker appendEntriesReplyTracker;
 106
 107     private final DOMTransactionFactory domTransactionFactory;
 108
 109     private final ShardTransactionActorFactory transactionActorFactory;
 110
 111     private final ShardSnapshotCohort snapshotCohort;
 112
 113     private final DataTreeChangeListenerSupport treeChangeSupport = new DataTreeChangeListenerSupport(this);
 114     private final DataChangeListenerSupport changeSupport = new DataChangeListenerSupport(this);
 115
 116     protected Shard(final ShardIdentifier name, final Map<String, String> peerAddresses,
 117             final DatastoreContext datastoreContext, final SchemaContext schemaContext) {
 118         super(name.toString(), new HashMap<>(peerAddresses), Optional.of(datastoreContext.getShardRaftConfig()));
 119
 120         this.name = name.toString();
 121         this.datastoreContext = datastoreContext;
 122
 123         setPersistence(datastoreContext.isPersistent());
 124
 125         LOG.info("Shard created : {}, persistent : {}", name, datastoreContext.isPersistent());
 126
 127         store = InMemoryDOMDataStoreFactory.create(name.toString(), null,
 128                 datastoreContext.getDataStoreProperties());
 129
 130         if (schemaContext != null) {
 131             store.onGlobalContextUpdated(schemaContext);
 132         }
 133
 134         shardMBean = ShardMBeanFactory.getShardStatsMBean(name.toString(),
 135                 datastoreContext.getDataStoreMXBeanType());
 136         shardMBean.setNotificationManager(store.getDataChangeListenerNotificationManager());
 137         shardMBean.setShardActor(getSelf());
 138
 139         if (isMetricsCaptureEnabled()) {
 140             getContext().become(new MeteringBehavior(this));
 141         }
 142
 143         domTransactionFactory = new DOMTransactionFactory(store, shardMBean, LOG, this.name);
 144
 145         commitCoordinator = new ShardCommitCoordinator(domTransactionFactory,
 146                 TimeUnit.SECONDS.convert(5, TimeUnit.MINUTES),
 147                 datastoreContext.getShardTransactionCommitQueueCapacity(), self(), LOG, this.name);
 148
 149         setTransactionCommitTimeout();
 150
 151         // create a notifier actor for each cluster member
 152         roleChangeNotifier = createRoleChangeNotifier(name.toString());
 153
 154         appendEntriesReplyTracker = new MessageTracker(AppendEntriesReply.class,
 155                 getRaftActorContext().getConfigParams().getIsolatedCheckIntervalInMillis());
 156
 157         transactionActorFactory = new ShardTransactionActorFactory(domTransactionFactory, datastoreContext,
 158                 new Dispatchers(context().system().dispatchers()).getDispatcherPath(
 159                         Dispatchers.DispatcherType.Transaction), self(), getContext(), shardMBean);
 160
 161         snapshotCohort = new ShardSnapshotCohort(transactionActorFactory, store, LOG, this.name);
 162     }
 163
 164     private void setTransactionCommitTimeout() {
 165         transactionCommitTimeout = TimeUnit.MILLISECONDS.convert(
 166                 datastoreContext.getShardTransactionCommitTimeoutInSeconds(), TimeUnit.SECONDS);
 167     }
 168
 169     public static Props props(final ShardIdentifier name,
 170         final Map<String, String> peerAddresses,
 171         final DatastoreContext datastoreContext, final SchemaContext schemaContext) {
 172         Preconditions.checkNotNull(name, "name should not be null");
 173         Preconditions.checkNotNull(peerAddresses, "peerAddresses should not be null");
 174         Preconditions.checkNotNull(datastoreContext, "dataStoreContext should not be null");
 175         Preconditions.checkNotNull(schemaContext, "schemaContext should not be null");
 176
 177         return Props.create(new ShardCreator(name, peerAddresses, datastoreContext, schemaContext));
 178     }
 179
 180     private Optional<ActorRef> createRoleChangeNotifier(String shardId) {
 181         ActorRef shardRoleChangeNotifier = this.getContext().actorOf(
 182             RoleChangeNotifier.getProps(shardId), shardId + "-notifier");
 183         return Optional.of(shardRoleChangeNotifier);
 184     }
 185
 186     @Override
 187     public void postStop() {
 188         LOG.info("Stopping Shard {}", persistenceId());
 189
 190         super.postStop();
 191
 192         if(txCommitTimeoutCheckSchedule != null) {
 193             txCommitTimeoutCheckSchedule.cancel();
 194         }
 195
 196         shardMBean.unregisterMBean();
 197     }
 198
 199     @Override
 200     public void onReceiveRecover(final Object message) throws Exception {
 201         if(LOG.isDebugEnabled()) {
 202             LOG.debug("{}: onReceiveRecover: Received message {} from {}", persistenceId(),
 203                 message.getClass().toString(), getSender());
 204         }
 205
 206         if (message instanceof RecoveryFailure){
 207             LOG.error("{}: Recovery failed because of this cause",
 208                     persistenceId(), ((RecoveryFailure) message).cause());
 209
 210             // Even though recovery failed, we still need to finish our recovery, eg send the
 211             // ActorInitialized message and start the txCommitTimeoutCheckSchedule.
 212             onRecoveryComplete();
 213         } else {
 214             super.onReceiveRecover(message);
 215             if(LOG.isTraceEnabled()) {
 216                 appendEntriesReplyTracker.begin();
 217             }
 218         }
 219     }
 220
 221     @Override
 222     public void onReceiveCommand(final Object message) throws Exception {
 223
 224         MessageTracker.Context context = appendEntriesReplyTracker.received(message);
 225
 226         if(context.error().isPresent()){
 227             LOG.trace("{} : AppendEntriesReply failed to arrive at the expected interval {}", persistenceId(),
 228                     context.error());
 229         }
 230
 231         try {
 232             if (CreateTransaction.SERIALIZABLE_CLASS.isInstance(message)) {
 233                 handleCreateTransaction(message);
 234             } else if (BatchedModifications.class.isInstance(message)) {
 235                 handleBatchedModifications((BatchedModifications)message);
 236             } else if (message instanceof ForwardedReadyTransaction) {
 237                 commitCoordinator.handleForwardedReadyTransaction((ForwardedReadyTransaction) message,
 238                         getSender(), this);
 239             } else if (CanCommitTransaction.SERIALIZABLE_CLASS.isInstance(message)) {
 240                 handleCanCommitTransaction(CanCommitTransaction.fromSerializable(message));
 241             } else if (CommitTransaction.SERIALIZABLE_CLASS.isInstance(message)) {
 242                 handleCommitTransaction(CommitTransaction.fromSerializable(message));
 243             } else if (AbortTransaction.SERIALIZABLE_CLASS.isInstance(message)) {
 244                 handleAbortTransaction(AbortTransaction.fromSerializable(message));
 245             } else if (CloseTransactionChain.SERIALIZABLE_CLASS.isInstance(message)) {
 246                 closeTransactionChain(CloseTransactionChain.fromSerializable(message));
 247             } else if (message instanceof RegisterChangeListener) {
 248                 changeSupport.onMessage((RegisterChangeListener) message, isLeader());
 249             } else if (message instanceof RegisterDataTreeChangeListener) {
 250                 treeChangeSupport.onMessage((RegisterDataTreeChangeListener) message, isLeader());
 251             } else if (message instanceof UpdateSchemaContext) {
 252                 updateSchemaContext((UpdateSchemaContext) message);
 253             } else if (message instanceof PeerAddressResolved) {
 254                 PeerAddressResolved resolved = (PeerAddressResolved) message;
 255                 setPeerAddress(resolved.getPeerId().toString(),
 256                         resolved.getPeerAddress());
 257             } else if (message.equals(TX_COMMIT_TIMEOUT_CHECK_MESSAGE)) {
 258                 handleTransactionCommitTimeoutCheck();
 259             } else if(message instanceof DatastoreContext) {
 260                 onDatastoreContext((DatastoreContext)message);
 261             } else if(message instanceof RegisterRoleChangeListener){
 262                 roleChangeNotifier.get().forward(message, context());
 263             } else if (message instanceof FollowerInitialSyncUpStatus){
 264                 shardMBean.setFollowerInitialSyncStatus(((FollowerInitialSyncUpStatus) message).isInitialSyncDone());
 265                 context().parent().tell(message, self());
 266             } else {
 267                 super.onReceiveCommand(message);
 268             }
 269         } finally {
 270             context.done();
 271         }
 272     }
 273
 274     @Override
 275     protected Optional<ActorRef> getRoleChangeNotifier() {
 276         return roleChangeNotifier;
 277     }
 278
 279     private void onDatastoreContext(DatastoreContext context) {
 280         datastoreContext = context;
 281
 282         commitCoordinator.setQueueCapacity(datastoreContext.getShardTransactionCommitQueueCapacity());
 283
 284         setTransactionCommitTimeout();
 285
 286         if(datastoreContext.isPersistent() && !persistence().isRecoveryApplicable()) {
 287             setPersistence(true);
 288         } else if(!datastoreContext.isPersistent() && persistence().isRecoveryApplicable()) {
 289             setPersistence(false);
 290         }
 291
 292         updateConfigParams(datastoreContext.getShardRaftConfig());
 293     }
 294
 295     private void handleTransactionCommitTimeoutCheck() {
 296         CohortEntry cohortEntry = commitCoordinator.getCurrentCohortEntry();
 297         if(cohortEntry != null) {
 298             long elapsed = System.currentTimeMillis() - cohortEntry.getLastAccessTime();
 299             if(elapsed > transactionCommitTimeout) {
 300                 LOG.warn("{}: Current transaction {} has timed out after {} ms - aborting",
 301                         persistenceId(), cohortEntry.getTransactionID(), transactionCommitTimeout);
 302
 303                 doAbortTransaction(cohortEntry.getTransactionID(), null);
 304             }
 305         }
 306     }
 307
 308     void continueCommit(final CohortEntry cohortEntry) throws Exception {
 309         // If we do not have any followers and we are not using persistence
 310         // or if cohortEntry has no modifications
 311         // we can apply modification to the state immediately
 312         if((!hasFollowers() && !persistence().isRecoveryApplicable()) || (!cohortEntry.hasModifications())){
 313             applyModificationToState(getSender(), cohortEntry.getTransactionID(), cohortEntry.getModification());
 314         } else {
 315             Shard.this.persistData(getSender(), cohortEntry.getTransactionID(),
 316                     new ModificationPayload(cohortEntry.getModification()));
 317         }
 318     }
 319
 320     private void handleCommitTransaction(final CommitTransaction commit) {
 321         if(!commitCoordinator.handleCommit(commit.getTransactionID(), getSender(), this)) {
 322             shardMBean.incrementFailedTransactionsCount();
 323         }
 324     }
 325
 326     private void finishCommit(@Nonnull final ActorRef sender, final @Nonnull String transactionID) {
 327         // With persistence enabled, this method is called via applyState by the leader strategy
 328         // after the commit has been replicated to a majority of the followers.
 329
 330         CohortEntry cohortEntry = commitCoordinator.getCohortEntryIfCurrent(transactionID);
 331         if(cohortEntry == null) {
 332             // The transaction is no longer the current commit. This can happen if the transaction
 333             // was aborted prior, most likely due to timeout in the front-end. We need to finish
 334             // committing the transaction though since it was successfully persisted and replicated
 335             // however we can't use the original cohort b/c it was already preCommitted and may
 336             // conflict with the current commit or may have been aborted so we commit with a new
 337             // transaction.
 338             cohortEntry = commitCoordinator.getAndRemoveCohortEntry(transactionID);
 339             if(cohortEntry != null) {
 340                 commitWithNewTransaction(cohortEntry.getModification());
 341                 sender.tell(CommitTransactionReply.INSTANCE.toSerializable(), getSelf());
 342             } else {
 343                 // This really shouldn't happen - it likely means that persistence or replication
 344                 // took so long to complete such that the cohort entry was expired from the cache.
 345                 IllegalStateException ex = new IllegalStateException(
 346                         String.format("%s: Could not finish committing transaction %s - no CohortEntry found",
 347                                 persistenceId(), transactionID));
 348                 LOG.error(ex.getMessage());
 349                 sender.tell(new akka.actor.Status.Failure(ex), getSelf());
 350             }
 351
 352             return;
 353         }
 354
 355         LOG.debug("{}: Finishing commit for transaction {}", persistenceId(), cohortEntry.getTransactionID());
 356
 357         try {
 358             // We block on the future here so we don't have to worry about possibly accessing our
 359             // state on a different thread outside of our dispatcher. Also, the data store
 360             // currently uses a same thread executor anyway.
 361             cohortEntry.getCohort().commit().get();
 362
 363             sender.tell(CommitTransactionReply.INSTANCE.toSerializable(), getSelf());
 364
 365             shardMBean.incrementCommittedTransactionCount();
 366             shardMBean.setLastCommittedTransactionTime(System.currentTimeMillis());
 367
 368         } catch (Exception e) {
 369             sender.tell(new akka.actor.Status.Failure(e), getSelf());
 370
 371             LOG.error("{}, An exception occurred while committing transaction {}", persistenceId(),
 372                     transactionID, e);
 373             shardMBean.incrementFailedTransactionsCount();
 374         } finally {
 375             commitCoordinator.currentTransactionComplete(transactionID, true);
 376         }
 377     }
 378
 379     private void handleCanCommitTransaction(final CanCommitTransaction canCommit) {
 380         LOG.debug("{}: Can committing transaction {}", persistenceId(), canCommit.getTransactionID());
 381         commitCoordinator.handleCanCommit(canCommit.getTransactionID(), getSender(), this);
 382     }
 383
 384     private void handleBatchedModifications(BatchedModifications batched) {
 385         // This message is sent to prepare the modificationsa transaction directly on the Shard as an
 386         // optimization to avoid the extra overhead of a separate ShardTransaction actor. On the last
 387         // BatchedModifications message, the caller sets the ready flag in the message indicating
 388         // modifications are complete. The reply contains the cohort actor path (this actor) for the caller
 389         // to initiate the 3-phase commit. This also avoids the overhead of sending an additional
 390         // ReadyTransaction message.
 391
 392         // If we're not the leader then forward to the leader. This is a safety measure - we shouldn't
 393         // normally get here if we're not the leader as the front-end (TransactionProxy) should determine
 394         // the primary/leader shard. However with timing and caching on the front-end, there's a small
 395         // window where it could have a stale leader during leadership transitions.
 396         //
 397         if(isLeader()) {
 398             try {
 399                 commitCoordinator.handleBatchedModifications(batched, getSender(), this);
 400             } catch (Exception e) {
 401                 LOG.error("{}: Error handling BatchedModifications for Tx {}", persistenceId(),
 402                         batched.getTransactionID(), e);
 403                 getSender().tell(new akka.actor.Status.Failure(e), getSelf());
 404             }
 405         } else {
 406             ActorSelection leader = getLeader();
 407             if(leader != null) {
 408                 // TODO: what if this is not the first batch and leadership changed in between batched messages?
 409                 // We could check if the commitCoordinator already has a cached entry and forward all the previous
 410                 // batched modifications.
 411                 LOG.debug("{}: Forwarding BatchedModifications to leader {}", persistenceId(), leader);
 412                 leader.forward(batched, getContext());
 413             } else {
 414                 // TODO: rather than throwing an immediate exception, we could schedule a timer to try again to make
 415                 // it more resilient in case we're in the process of electing a new leader.
 416                 getSender().tell(new akka.actor.Status.Failure(new NoShardLeaderException(String.format(
 417                     "Could not find the leader for shard %s. This typically happens" +
 418                     " when the system is coming up or recovering and a leader is being elected. Try again" +
 419                     " later.", persistenceId()))), getSelf());
 420             }
 421         }
 422     }
 423
 424     private void handleAbortTransaction(final AbortTransaction abort) {
 425         doAbortTransaction(abort.getTransactionID(), getSender());
 426     }
 427
 428     void doAbortTransaction(final String transactionID, final ActorRef sender) {
 429         final CohortEntry cohortEntry = commitCoordinator.getCohortEntryIfCurrent(transactionID);
 430         if(cohortEntry != null) {
 431             LOG.debug("{}: Aborting transaction {}", persistenceId(), transactionID);
 432
 433             // We don't remove the cached cohort entry here (ie pass false) in case the Tx was
 434             // aborted during replication in which case we may still commit locally if replication
 435             // succeeds.
 436             commitCoordinator.currentTransactionComplete(transactionID, false);
 437
 438             final ListenableFuture<Void> future = cohortEntry.getCohort().abort();
 439             final ActorRef self = getSelf();
 440
 441             Futures.addCallback(future, new FutureCallback<Void>() {
 442                 @Override
 443                 public void onSuccess(final Void v) {
 444                     shardMBean.incrementAbortTransactionsCount();
 445
 446                     if(sender != null) {
 447                         sender.tell(AbortTransactionReply.INSTANCE.toSerializable(), self);
 448                     }
 449                 }
 450
 451                 @Override
 452                 public void onFailure(final Throwable t) {
 453                     LOG.error("{}: An exception happened during abort", persistenceId(), t);
 454
 455                     if(sender != null) {
 456                         sender.tell(new akka.actor.Status.Failure(t), self);
 457                     }
 458                 }
 459             });
 460         }
 461     }
 462
 463     private void handleCreateTransaction(final Object message) {
 464         if (isLeader()) {
 465             createTransaction(CreateTransaction.fromSerializable(message));
 466         } else if (getLeader() != null) {
 467             getLeader().forward(message, getContext());
 468         } else {
 469             getSender().tell(new akka.actor.Status.Failure(new NoShardLeaderException(String.format(
 470                 "Could not find leader for shard %s so transaction cannot be created. This typically happens" +
 471                 " when the system is coming up or recovering and a leader is being elected. Try again" +
 472                 " later.", persistenceId()))), getSelf());
 473         }
 474     }
 475
 476     private void closeTransactionChain(final CloseTransactionChain closeTransactionChain) {
 477         domTransactionFactory.closeTransactionChain(closeTransactionChain.getTransactionChainId());
 478     }
 479
 480     private ActorRef createTypedTransactionActor(int transactionType,
 481             ShardTransactionIdentifier transactionId, String transactionChainId,
 482             short clientVersion ) {
 483
 484         return transactionActorFactory.newShardTransaction(TransactionProxy.TransactionType.fromInt(transactionType),
 485                 transactionId, transactionChainId, clientVersion);
 486     }
 487
 488     private void createTransaction(CreateTransaction createTransaction) {
 489         try {
 490             ActorRef transactionActor = createTransaction(createTransaction.getTransactionType(),
 491                 createTransaction.getTransactionId(), createTransaction.getTransactionChainId(),
 492                 createTransaction.getVersion());
 493
 494             getSender().tell(new CreateTransactionReply(Serialization.serializedActorPath(transactionActor),
 495                     createTransaction.getTransactionId()).toSerializable(), getSelf());
 496         } catch (Exception e) {
 497             getSender().tell(new akka.actor.Status.Failure(e), getSelf());
 498         }
 499     }
 500
 501     private ActorRef createTransaction(int transactionType, String remoteTransactionId,
 502             String transactionChainId, short clientVersion) {
 503
 504
 505         ShardTransactionIdentifier transactionId = new ShardTransactionIdentifier(remoteTransactionId);
 506
 507         if(LOG.isDebugEnabled()) {
 508             LOG.debug("{}: Creating transaction : {} ", persistenceId(), transactionId);
 509         }
 510
 511         ActorRef transactionActor = createTypedTransactionActor(transactionType, transactionId,
 512                 transactionChainId, clientVersion);
 513
 514         return transactionActor;
 515     }
 516
 517     private void commitWithNewTransaction(final Modification modification) {
 518         DOMStoreWriteTransaction tx = store.newWriteOnlyTransaction();
 519         modification.apply(tx);
 520         try {
 521             snapshotCohort.syncCommitTransaction(tx);
 522             shardMBean.incrementCommittedTransactionCount();
 523             shardMBean.setLastCommittedTransactionTime(System.currentTimeMillis());
 524         } catch (InterruptedException | ExecutionException e) {
 525             shardMBean.incrementFailedTransactionsCount();
 526             LOG.error("{}: Failed to commit", persistenceId(), e);
 527         }
 528     }
 529
 530     private void updateSchemaContext(final UpdateSchemaContext message) {
 531         updateSchemaContext(message.getSchemaContext());
 532     }
 533
 534     @VisibleForTesting
 535     void updateSchemaContext(final SchemaContext schemaContext) {
 536         store.onGlobalContextUpdated(schemaContext);
 537     }
 538
 539     private boolean isMetricsCaptureEnabled() {
 540         CommonConfig config = new CommonConfig(getContext().system().settings().config());
 541         return config.isMetricCaptureEnabled();
 542     }
 543
 544     @Override
 545     protected RaftActorSnapshotCohort getRaftActorSnapshotCohort() {
 546         return snapshotCohort;
 547     }
 548
 549     @Override
 550     @Nonnull
 551     protected RaftActorRecoveryCohort getRaftActorRecoveryCohort() {
 552         return new ShardRecoveryCoordinator(store, persistenceId(), LOG);
 553     }
 554
 555     @Override
 556     protected void onRecoveryComplete() {
 557         //notify shard manager
 558         getContext().parent().tell(new ActorInitialized(), getSelf());
 559
 560         // Being paranoid here - this method should only be called once but just in case...
 561         if(txCommitTimeoutCheckSchedule == null) {
 562             // Schedule a message to be periodically sent to check if the current in-progress
 563             // transaction should be expired and aborted.
 564             FiniteDuration period = Duration.create(transactionCommitTimeout / 3, TimeUnit.MILLISECONDS);
 565             txCommitTimeoutCheckSchedule = getContext().system().scheduler().schedule(
 566                     period, period, getSelf(),
 567                     TX_COMMIT_TIMEOUT_CHECK_MESSAGE, getContext().dispatcher(), ActorRef.noSender());
 568         }
 569     }
 570
 571     @Override
 572     protected void applyState(final ActorRef clientActor, final String identifier, final Object data) {
 573
 574         if(data instanceof ModificationPayload) {
 575             try {
 576                 applyModificationToState(clientActor, identifier, ((ModificationPayload) data).getModification());
 577             } catch (ClassNotFoundException | IOException e) {
 578                 LOG.error("{}: Error extracting ModificationPayload", persistenceId(), e);
 579             }
 580         }
 581         else if (data instanceof CompositeModificationPayload) {
 582             Object modification = ((CompositeModificationPayload) data).getModification();
 583
 584             applyModificationToState(clientActor, identifier, modification);
 585         } else if(data instanceof CompositeModificationByteStringPayload ){
 586             Object modification = ((CompositeModificationByteStringPayload) data).getModification();
 587
 588             applyModificationToState(clientActor, identifier, modification);
 589         } else {
 590             LOG.error("{}: Unknown state received {} Class loader = {} CompositeNodeMod.ClassLoader = {}",
 591                     persistenceId(), data, data.getClass().getClassLoader(),
 592                     CompositeModificationPayload.class.getClassLoader());
 593         }
 594     }
 595
 596     private void applyModificationToState(ActorRef clientActor, String identifier, Object modification) {
 597         if(modification == null) {
 598             LOG.error(
 599                     "{}: modification is null - this is very unexpected, clientActor = {}, identifier = {}",
 600                     persistenceId(), identifier, clientActor != null ? clientActor.path().toString() : null);
 601         } else if(clientActor == null) {
 602             // There's no clientActor to which to send a commit reply so we must be applying
 603             // replicated state from the leader.
 604             commitWithNewTransaction(MutableCompositeModification.fromSerializable(modification));
 605         } else {
 606             // This must be the OK to commit after replication consensus.
 607             finishCommit(clientActor, identifier);
 608         }
 609     }
 610
 611     @Override
 612     protected void onStateChanged() {
 613         boolean isLeader = isLeader();
 614         changeSupport.onLeadershipChange(isLeader);
 615         treeChangeSupport.onLeadershipChange(isLeader);
 616
 617         // If this actor is no longer the leader close all the transaction chains
 618         if (!isLeader) {
 619             if(LOG.isDebugEnabled()) {
 620                 LOG.debug(
 621                     "{}: onStateChanged: Closing all transaction chains because shard {} is no longer the leader",
 622                     persistenceId(), getId());
 623             }
 624
 625             domTransactionFactory.closeAllTransactionChains();
 626         }
 627     }
 628
 629     @Override
 630     protected void onLeaderChanged(String oldLeader, String newLeader) {
 631         shardMBean.incrementLeadershipChangeCount();
 632     }
 633
 634     @Override
 635     public String persistenceId() {
 636         return this.name;
 637     }
 638
 639     @VisibleForTesting
 640     ShardCommitCoordinator getCommitCoordinator() {
 641         return commitCoordinator;
 642     }
 643
 644
 645     private static class ShardCreator implements Creator<Shard> {
 646
 647         private static final long serialVersionUID = 1L;
 648
 649         final ShardIdentifier name;
 650         final Map<String, String> peerAddresses;
 651         final DatastoreContext datastoreContext;
 652         final SchemaContext schemaContext;
 653
 654         ShardCreator(final ShardIdentifier name, final Map<String, String> peerAddresses,
 655                 final DatastoreContext datastoreContext, final SchemaContext schemaContext) {
 656             this.name = name;
 657             this.peerAddresses = peerAddresses;
 658             this.datastoreContext = datastoreContext;
 659             this.schemaContext = schemaContext;
 660         }
 661
 662         @Override
 663         public Shard create() throws Exception {
 664             return new Shard(name, peerAddresses, datastoreContext, schemaContext);
 665         }
 666     }
 667
 668     @VisibleForTesting
 669     public InMemoryDOMDataStore getDataStore() {
 670         return store;
 671     }
 672
 673     @VisibleForTesting
 674     ShardStats getShardMBean() {
 675         return shardMBean;
 676     }
 677 }