opendaylight/md-sal/sal-akka-raft/src/main/java/org/opendaylight/controller/cluster/raft/RaftActor.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.raft;
  10
  11 import akka.actor.ActorRef;
  12 import akka.actor.ActorSelection;
  13 import akka.event.Logging;
  14 import akka.event.LoggingAdapter;
  15 import akka.japi.Procedure;
  16 import akka.persistence.RecoveryCompleted;
  17 import akka.persistence.SaveSnapshotFailure;
  18 import akka.persistence.SaveSnapshotSuccess;
  19 import akka.persistence.SnapshotOffer;
  20 import akka.persistence.SnapshotSelectionCriteria;
  21 import com.google.common.annotations.VisibleForTesting;
  22 import com.google.common.base.Optional;
  23 import com.google.common.base.Stopwatch;
  24 import com.google.protobuf.ByteString;
  25 import org.opendaylight.controller.cluster.DataPersistenceProvider;
  26 import org.opendaylight.controller.cluster.common.actor.AbstractUntypedPersistentActor;
  27 import org.opendaylight.controller.cluster.raft.base.messages.ApplyLogEntries;
  28 import org.opendaylight.controller.cluster.raft.base.messages.ApplySnapshot;
  29 import org.opendaylight.controller.cluster.raft.base.messages.ApplyState;
  30 import org.opendaylight.controller.cluster.raft.base.messages.CaptureSnapshot;
  31 import org.opendaylight.controller.cluster.raft.base.messages.CaptureSnapshotReply;
  32 import org.opendaylight.controller.cluster.raft.base.messages.Replicate;
  33 import org.opendaylight.controller.cluster.raft.base.messages.SendHeartBeat;
  34 import org.opendaylight.controller.cluster.raft.base.messages.SendInstallSnapshot;
  35 import org.opendaylight.controller.cluster.raft.behaviors.AbstractRaftActorBehavior;
  36 import org.opendaylight.controller.cluster.raft.behaviors.Follower;
  37 import org.opendaylight.controller.cluster.raft.behaviors.RaftActorBehavior;
  38 import org.opendaylight.controller.cluster.raft.client.messages.FindLeader;
  39 import org.opendaylight.controller.cluster.raft.client.messages.FindLeaderReply;
  40 import org.opendaylight.controller.cluster.raft.messages.AppendEntriesReply;
  41 import org.opendaylight.controller.cluster.raft.protobuff.client.messages.Payload;
  42 import org.opendaylight.controller.protobuff.messages.cluster.raft.AppendEntriesMessages;
  43
  44 import java.io.Serializable;
  45 import java.util.Map;
  46
  47 /**
  48  * RaftActor encapsulates a state machine that needs to be kept synchronized
  49  * in a cluster. It implements the RAFT algorithm as described in the paper
  50  * <a href='https://ramcloud.stanford.edu/wiki/download/attachments/11370504/raft.pdf'>
  51  * In Search of an Understandable Consensus Algorithm</a>
  52  * <p/>
  53  * RaftActor has 3 states and each state has a certain behavior associated
  54  * with it. A Raft actor can behave as,
  55  * <ul>
  56  * <li> A Leader </li>
  57  * <li> A Follower (or) </li>
  58  * <li> A Candidate </li>
  59  * </ul>
  60  * <p/>
  61  * <p/>
  62  * A RaftActor MUST be a Leader in order to accept requests from clients to
  63  * change the state of it's encapsulated state machine. Once a RaftActor becomes
  64  * a Leader it is also responsible for ensuring that all followers ultimately
  65  * have the same log and therefore the same state machine as itself.
  66  * <p/>
  67  * <p/>
  68  * The current behavior of a RaftActor determines how election for leadership
  69  * is initiated and how peer RaftActors react to request for votes.
  70  * <p/>
  71  * <p/>
  72  * Each RaftActor also needs to know the current election term. It uses this
  73  * information for a couple of things. One is to simply figure out who it
  74  * voted for in the last election. Another is to figure out if the message
  75  * it received to update it's state is stale.
  76  * <p/>
  77  * <p/>
  78  * The RaftActor uses akka-persistence to store it's replicated log.
  79  * Furthermore through it's behaviors a Raft Actor determines
  80  * <p/>
  81  * <ul>
  82  * <li> when a log entry should be persisted </li>
  83  * <li> when a log entry should be applied to the state machine (and) </li>
  84  * <li> when a snapshot should be saved </li>
  85  * </ul>
  86  */
  87 public abstract class RaftActor extends AbstractUntypedPersistentActor {
  88     protected final LoggingAdapter LOG =
  89         Logging.getLogger(getContext().system(), this);
  90
  91     /**
  92      * The current state determines the current behavior of a RaftActor
  93      * A Raft Actor always starts off in the Follower State
  94      */
  95     private RaftActorBehavior currentBehavior;
  96
  97     /**
  98      * This context should NOT be passed directly to any other actor it is
  99      * only to be consumed by the RaftActorBehaviors
 100      */
 101     private final RaftActorContext context;
 102
 103     /**
 104      * The in-memory journal
 105      */
 106     private ReplicatedLogImpl replicatedLog = new ReplicatedLogImpl();
 107
 108     private CaptureSnapshot captureSnapshot = null;
 109
 110     private volatile boolean hasSnapshotCaptureInitiated = false;
 111
 112     private Stopwatch recoveryTimer;
 113
 114     private int currentRecoveryBatchCount;
 115
 116     public RaftActor(String id, Map<String, String> peerAddresses) {
 117         this(id, peerAddresses, Optional.<ConfigParams>absent());
 118     }
 119
 120     public RaftActor(String id, Map<String, String> peerAddresses,
 121          Optional<ConfigParams> configParams) {
 122
 123         context = new RaftActorContextImpl(this.getSelf(),
 124             this.getContext(), id, new ElectionTermImpl(),
 125             -1, -1, replicatedLog, peerAddresses,
 126             (configParams.isPresent() ? configParams.get(): new DefaultConfigParamsImpl()),
 127             LOG);
 128     }
 129
 130     private void initRecoveryTimer() {
 131         if(recoveryTimer == null) {
 132             recoveryTimer = new Stopwatch();
 133             recoveryTimer.start();
 134         }
 135     }
 136
 137     @Override
 138     public void preStart() throws Exception {
 139         LOG.info("Starting recovery for {} with journal batch size {}", persistenceId(),
 140                 context.getConfigParams().getJournalRecoveryLogBatchSize());
 141
 142         super.preStart();
 143     }
 144
 145     @Override
 146     public void handleRecover(Object message) {
 147         if(persistence().isRecoveryApplicable()) {
 148             if (message instanceof SnapshotOffer) {
 149                 onRecoveredSnapshot((SnapshotOffer) message);
 150             } else if (message instanceof ReplicatedLogEntry) {
 151                 onRecoveredJournalLogEntry((ReplicatedLogEntry) message);
 152             } else if (message instanceof ApplyLogEntries) {
 153                 onRecoveredApplyLogEntries((ApplyLogEntries) message);
 154             } else if (message instanceof DeleteEntries) {
 155                 replicatedLog.removeFrom(((DeleteEntries) message).getFromIndex());
 156             } else if (message instanceof UpdateElectionTerm) {
 157                 context.getTermInformation().update(((UpdateElectionTerm) message).getCurrentTerm(),
 158                         ((UpdateElectionTerm) message).getVotedFor());
 159             } else if (message instanceof RecoveryCompleted) {
 160                 onRecoveryCompletedMessage();
 161             }
 162         } else {
 163             if (message instanceof RecoveryCompleted) {
 164                 // Delete all the messages from the akka journal so that we do not end up with consistency issues
 165                 // Note I am not using the dataPersistenceProvider and directly using the akka api here
 166                 deleteMessages(lastSequenceNr());
 167
 168                 // Delete all the akka snapshots as they will not be needed
 169                 deleteSnapshots(new SnapshotSelectionCriteria(scala.Long.MaxValue(), scala.Long.MaxValue()));
 170
 171                 onRecoveryComplete();
 172                 currentBehavior = new Follower(context);
 173                 onStateChanged();
 174             }
 175         }
 176     }
 177
 178     private void onRecoveredSnapshot(SnapshotOffer offer) {
 179         if(LOG.isDebugEnabled()) {
 180             LOG.debug("SnapshotOffer called..");
 181         }
 182
 183         initRecoveryTimer();
 184
 185         Snapshot snapshot = (Snapshot) offer.snapshot();
 186
 187         // Create a replicated log with the snapshot information
 188         // The replicated log can be used later on to retrieve this snapshot
 189         // when we need to install it on a peer
 190         replicatedLog = new ReplicatedLogImpl(snapshot);
 191
 192         context.setReplicatedLog(replicatedLog);
 193         context.setLastApplied(snapshot.getLastAppliedIndex());
 194         context.setCommitIndex(snapshot.getLastAppliedIndex());
 195
 196         Stopwatch timer = new Stopwatch();
 197         timer.start();
 198
 199         // Apply the snapshot to the actors state
 200         applyRecoverySnapshot(ByteString.copyFrom(snapshot.getState()));
 201
 202         timer.stop();
 203         LOG.info("Recovery snapshot applied for {} in {}: snapshotIndex={}, snapshotTerm={}, journal-size=" +
 204                 replicatedLog.size(), persistenceId(), timer.toString(),
 205                 replicatedLog.snapshotIndex, replicatedLog.snapshotTerm);
 206     }
 207
 208     private void onRecoveredJournalLogEntry(ReplicatedLogEntry logEntry) {
 209         if(LOG.isDebugEnabled()) {
 210             LOG.debug("Received ReplicatedLogEntry for recovery: {}", logEntry.getIndex());
 211         }
 212
 213         replicatedLog.append(logEntry);
 214     }
 215
 216     private void onRecoveredApplyLogEntries(ApplyLogEntries ale) {
 217         if(LOG.isDebugEnabled()) {
 218             LOG.debug("Received ApplyLogEntries for recovery, applying to state: {} to {}",
 219                     context.getLastApplied() + 1, ale.getToIndex());
 220         }
 221
 222         for (long i = context.getLastApplied() + 1; i <= ale.getToIndex(); i++) {
 223             batchRecoveredLogEntry(replicatedLog.get(i));
 224         }
 225
 226         context.setLastApplied(ale.getToIndex());
 227         context.setCommitIndex(ale.getToIndex());
 228     }
 229
 230     private void batchRecoveredLogEntry(ReplicatedLogEntry logEntry) {
 231         initRecoveryTimer();
 232
 233         int batchSize = context.getConfigParams().getJournalRecoveryLogBatchSize();
 234         if(currentRecoveryBatchCount == 0) {
 235             startLogRecoveryBatch(batchSize);
 236         }
 237
 238         appendRecoveredLogEntry(logEntry.getData());
 239
 240         if(++currentRecoveryBatchCount >= batchSize) {
 241             endCurrentLogRecoveryBatch();
 242         }
 243     }
 244
 245     private void endCurrentLogRecoveryBatch() {
 246         applyCurrentLogRecoveryBatch();
 247         currentRecoveryBatchCount = 0;
 248     }
 249
 250     private void onRecoveryCompletedMessage() {
 251         if(currentRecoveryBatchCount > 0) {
 252             endCurrentLogRecoveryBatch();
 253         }
 254
 255         onRecoveryComplete();
 256
 257         String recoveryTime = "";
 258         if(recoveryTimer != null) {
 259             recoveryTimer.stop();
 260             recoveryTime = " in " + recoveryTimer.toString();
 261             recoveryTimer = null;
 262         }
 263
 264         LOG.info(
 265             "Recovery completed" + recoveryTime + " - Switching actor to Follower - " +
 266                 "Persistence Id =  " + persistenceId() +
 267                 " Last index in log={}, snapshotIndex={}, snapshotTerm={}, " +
 268                 "journal-size={}",
 269             replicatedLog.lastIndex(), replicatedLog.snapshotIndex,
 270             replicatedLog.snapshotTerm, replicatedLog.size());
 271
 272         currentBehavior = new Follower(context);
 273         onStateChanged();
 274     }
 275
 276     @Override public void handleCommand(Object message) {
 277         if (message instanceof ApplyState){
 278             ApplyState applyState = (ApplyState) message;
 279
 280             if(LOG.isDebugEnabled()) {
 281                 LOG.debug("Applying state for log index {} data {}",
 282                     applyState.getReplicatedLogEntry().getIndex(),
 283                     applyState.getReplicatedLogEntry().getData());
 284             }
 285
 286             applyState(applyState.getClientActor(), applyState.getIdentifier(),
 287                 applyState.getReplicatedLogEntry().getData());
 288
 289         } else if (message instanceof ApplyLogEntries){
 290             ApplyLogEntries ale = (ApplyLogEntries) message;
 291             if(LOG.isDebugEnabled()) {
 292                 LOG.debug("Persisting ApplyLogEntries with index={}", ale.getToIndex());
 293             }
 294             persistence().persist(new ApplyLogEntries(ale.getToIndex()), new Procedure<ApplyLogEntries>() {
 295                 @Override
 296                 public void apply(ApplyLogEntries param) throws Exception {
 297                 }
 298             });
 299
 300         } else if(message instanceof ApplySnapshot ) {
 301             Snapshot snapshot = ((ApplySnapshot) message).getSnapshot();
 302
 303             if(LOG.isDebugEnabled()) {
 304                 LOG.debug("ApplySnapshot called on Follower Actor " +
 305                         "snapshotIndex:{}, snapshotTerm:{}", snapshot.getLastAppliedIndex(),
 306                     snapshot.getLastAppliedTerm()
 307                 );
 308             }
 309             applySnapshot(ByteString.copyFrom(snapshot.getState()));
 310
 311             //clears the followers log, sets the snapshot index to ensure adjusted-index works
 312             replicatedLog = new ReplicatedLogImpl(snapshot);
 313             context.setReplicatedLog(replicatedLog);
 314             context.setLastApplied(snapshot.getLastAppliedIndex());
 315
 316         } else if (message instanceof FindLeader) {
 317             getSender().tell(
 318                 new FindLeaderReply(getLeaderAddress()),
 319                 getSelf()
 320             );
 321
 322         } else if (message instanceof SaveSnapshotSuccess) {
 323             SaveSnapshotSuccess success = (SaveSnapshotSuccess) message;
 324             LOG.info("SaveSnapshotSuccess received for snapshot");
 325
 326             long sequenceNumber = success.metadata().sequenceNr();
 327
 328             commitSnapshot(sequenceNumber);
 329
 330         } else if (message instanceof SaveSnapshotFailure) {
 331             SaveSnapshotFailure saveSnapshotFailure = (SaveSnapshotFailure) message;
 332
 333             LOG.info("saveSnapshotFailure.metadata():{}", saveSnapshotFailure.metadata().toString());
 334             LOG.error(saveSnapshotFailure.cause(), "SaveSnapshotFailure received for snapshot Cause:");
 335
 336             context.getReplicatedLog().snapshotRollback();
 337
 338             LOG.info("Replicated Log rollbacked. Snapshot will be attempted in the next cycle." +
 339                 "snapshotIndex:{}, snapshotTerm:{}, log-size:{}",
 340                 context.getReplicatedLog().getSnapshotIndex(),
 341                 context.getReplicatedLog().getSnapshotTerm(),
 342                 context.getReplicatedLog().size());
 343
 344         } else if (message instanceof CaptureSnapshot) {
 345             LOG.info("CaptureSnapshot received by actor");
 346             CaptureSnapshot cs = (CaptureSnapshot)message;
 347             captureSnapshot = cs;
 348             createSnapshot();
 349
 350         } else if (message instanceof CaptureSnapshotReply){
 351             LOG.info("CaptureSnapshotReply received by actor");
 352             CaptureSnapshotReply csr = (CaptureSnapshotReply) message;
 353
 354             ByteString stateInBytes = csr.getSnapshot();
 355             LOG.info("CaptureSnapshotReply stateInBytes size:{}", stateInBytes.size());
 356             handleCaptureSnapshotReply(stateInBytes);
 357
 358         } else {
 359             if (!(message instanceof AppendEntriesMessages.AppendEntries)
 360                 && !(message instanceof AppendEntriesReply) && !(message instanceof SendHeartBeat)) {
 361                 if(LOG.isDebugEnabled()) {
 362                     LOG.debug("onReceiveCommand: message: {}", message.getClass());
 363                 }
 364             }
 365
 366             RaftActorBehavior oldBehavior = currentBehavior;
 367             currentBehavior = currentBehavior.handleMessage(getSender(), message);
 368
 369             if(oldBehavior != currentBehavior){
 370                 onStateChanged();
 371             }
 372
 373             onLeaderChanged(oldBehavior.getLeaderId(), currentBehavior.getLeaderId());
 374         }
 375     }
 376
 377     public java.util.Set<String> getPeers() {
 378
 379         return context.getPeerAddresses().keySet();
 380     }
 381
 382     protected String getReplicatedLogState() {
 383         return "snapshotIndex=" + context.getReplicatedLog().getSnapshotIndex()
 384             + ", snapshotTerm=" + context.getReplicatedLog().getSnapshotTerm()
 385             + ", im-mem journal size=" + context.getReplicatedLog().size();
 386     }
 387
 388
 389     /**
 390      * When a derived RaftActor needs to persist something it must call
 391      * persistData.
 392      *
 393      * @param clientActor
 394      * @param identifier
 395      * @param data
 396      */
 397     protected void persistData(ActorRef clientActor, String identifier,
 398         Payload data) {
 399
 400         ReplicatedLogEntry replicatedLogEntry = new ReplicatedLogImplEntry(
 401             context.getReplicatedLog().lastIndex() + 1,
 402             context.getTermInformation().getCurrentTerm(), data);
 403
 404         if(LOG.isDebugEnabled()) {
 405             LOG.debug("Persist data {}", replicatedLogEntry);
 406         }
 407
 408         replicatedLog
 409             .appendAndPersist(clientActor, identifier, replicatedLogEntry);
 410     }
 411
 412     protected String getId() {
 413         return context.getId();
 414     }
 415
 416     /**
 417      * Derived actors can call the isLeader method to check if the current
 418      * RaftActor is the Leader or not
 419      *
 420      * @return true it this RaftActor is a Leader false otherwise
 421      */
 422     protected boolean isLeader() {
 423         return context.getId().equals(currentBehavior.getLeaderId());
 424     }
 425
 426     /**
 427      * Derived actor can call getLeader if they need a reference to the Leader.
 428      * This would be useful for example in forwarding a request to an actor
 429      * which is the leader
 430      *
 431      * @return A reference to the leader if known, null otherwise
 432      */
 433     protected ActorSelection getLeader(){
 434         String leaderAddress = getLeaderAddress();
 435
 436         if(leaderAddress == null){
 437             return null;
 438         }
 439
 440         return context.actorSelection(leaderAddress);
 441     }
 442
 443     /**
 444      *
 445      * @return the current leader's id
 446      */
 447     protected String getLeaderId(){
 448         return currentBehavior.getLeaderId();
 449     }
 450
 451     protected RaftState getRaftState() {
 452         return currentBehavior.state();
 453     }
 454
 455     protected ReplicatedLogEntry getLastLogEntry() {
 456         return replicatedLog.last();
 457     }
 458
 459     protected Long getCurrentTerm(){
 460         return context.getTermInformation().getCurrentTerm();
 461     }
 462
 463     protected Long getCommitIndex(){
 464         return context.getCommitIndex();
 465     }
 466
 467     protected Long getLastApplied(){
 468         return context.getLastApplied();
 469     }
 470
 471     protected RaftActorContext getRaftActorContext() {
 472         return context;
 473     }
 474
 475     /**
 476      * setPeerAddress sets the address of a known peer at a later time.
 477      * <p>
 478      * This is to account for situations where a we know that a peer
 479      * exists but we do not know an address up-front. This may also be used in
 480      * situations where a known peer starts off in a different location and we
 481      * need to change it's address
 482      * <p>
 483      * Note that if the peerId does not match the list of peers passed to
 484      * this actor during construction an IllegalStateException will be thrown.
 485      *
 486      * @param peerId
 487      * @param peerAddress
 488      */
 489     protected void setPeerAddress(String peerId, String peerAddress){
 490         context.setPeerAddress(peerId, peerAddress);
 491     }
 492
 493     protected void commitSnapshot(long sequenceNumber) {
 494         context.getReplicatedLog().snapshotCommit();
 495
 496         // TODO: Not sure if we want to be this aggressive with trimming stuff
 497         trimPersistentData(sequenceNumber);
 498     }
 499
 500     /**
 501      * The applyState method will be called by the RaftActor when some data
 502      * needs to be applied to the actor's state
 503      *
 504      * @param clientActor A reference to the client who sent this message. This
 505      *                    is the same reference that was passed to persistData
 506      *                    by the derived actor. clientActor may be null when
 507      *                    the RaftActor is behaving as a follower or during
 508      *                    recovery.
 509      * @param identifier  The identifier of the persisted data. This is also
 510      *                    the same identifier that was passed to persistData by
 511      *                    the derived actor. identifier may be null when
 512      *                    the RaftActor is behaving as a follower or during
 513      *                    recovery
 514      * @param data        A piece of data that was persisted by the persistData call.
 515      *                    This should NEVER be null.
 516      */
 517     protected abstract void applyState(ActorRef clientActor, String identifier,
 518         Object data);
 519
 520     /**
 521      * This method is called during recovery at the start of a batch of state entries. Derived
 522      * classes should perform any initialization needed to start a batch.
 523      */
 524     protected abstract void startLogRecoveryBatch(int maxBatchSize);
 525
 526     /**
 527      * This method is called during recovery to append state data to the current batch. This method
 528      * is called 1 or more times after {@link #startLogRecoveryBatch}.
 529      *
 530      * @param data the state data
 531      */
 532     protected abstract void appendRecoveredLogEntry(Payload data);
 533
 534     /**
 535      * This method is called during recovery to reconstruct the state of the actor.
 536      *
 537      * @param snapshot A snapshot of the state of the actor
 538      */
 539     protected abstract void applyRecoverySnapshot(ByteString snapshot);
 540
 541     /**
 542      * This method is called during recovery at the end of a batch to apply the current batched
 543      * log entries. This method is called after {@link #appendRecoveredLogEntry}.
 544      */
 545     protected abstract void applyCurrentLogRecoveryBatch();
 546
 547     /**
 548      * This method is called when recovery is complete.
 549      */
 550     protected abstract void onRecoveryComplete();
 551
 552     /**
 553      * This method will be called by the RaftActor when a snapshot needs to be
 554      * created. The derived actor should respond with its current state.
 555      * <p/>
 556      * During recovery the state that is returned by the derived actor will
 557      * be passed back to it by calling the applySnapshot  method
 558      *
 559      * @return The current state of the actor
 560      */
 561     protected abstract void createSnapshot();
 562
 563     /**
 564      * This method can be called at any other point during normal
 565      * operations when the derived actor is out of sync with it's peers
 566      * and the only way to bring it in sync is by applying a snapshot
 567      *
 568      * @param snapshot A snapshot of the state of the actor
 569      */
 570     protected abstract void applySnapshot(ByteString snapshot);
 571
 572     /**
 573      * This method will be called by the RaftActor when the state of the
 574      * RaftActor changes. The derived actor can then use methods like
 575      * isLeader or getLeader to do something useful
 576      */
 577     protected abstract void onStateChanged();
 578
 579     protected abstract DataPersistenceProvider persistence();
 580
 581     protected void onLeaderChanged(String oldLeader, String newLeader){};
 582
 583     private void trimPersistentData(long sequenceNumber) {
 584         // Trim akka snapshots
 585         // FIXME : Not sure how exactly the SnapshotSelectionCriteria is applied
 586         // For now guessing that it is ANDed.
 587         persistence().deleteSnapshots(new SnapshotSelectionCriteria(
 588             sequenceNumber - context.getConfigParams().getSnapshotBatchCount(), 43200000));
 589
 590         // Trim akka journal
 591         persistence().deleteMessages(sequenceNumber);
 592     }
 593
 594     private String getLeaderAddress(){
 595         if(isLeader()){
 596             return getSelf().path().toString();
 597         }
 598         String leaderId = currentBehavior.getLeaderId();
 599         if (leaderId == null) {
 600             return null;
 601         }
 602         String peerAddress = context.getPeerAddress(leaderId);
 603         if(LOG.isDebugEnabled()) {
 604             LOG.debug("getLeaderAddress leaderId = {} peerAddress = {}",
 605                     leaderId, peerAddress);
 606         }
 607
 608         return peerAddress;
 609     }
 610
 611     private void handleCaptureSnapshotReply(ByteString stateInBytes) {
 612         // create a snapshot object from the state provided and save it
 613         // when snapshot is saved async, SaveSnapshotSuccess is raised.
 614
 615         Snapshot sn = Snapshot.create(stateInBytes.toByteArray(),
 616             context.getReplicatedLog().getFrom(captureSnapshot.getLastAppliedIndex() + 1),
 617             captureSnapshot.getLastIndex(), captureSnapshot.getLastTerm(),
 618             captureSnapshot.getLastAppliedIndex(), captureSnapshot.getLastAppliedTerm());
 619
 620         persistence().saveSnapshot(sn);
 621
 622         LOG.info("Persisting of snapshot done:{}", sn.getLogMessage());
 623
 624         //be greedy and remove entries from in-mem journal which are in the snapshot
 625         // and update snapshotIndex and snapshotTerm without waiting for the success,
 626
 627         context.getReplicatedLog().snapshotPreCommit(
 628             captureSnapshot.getLastAppliedIndex(),
 629             captureSnapshot.getLastAppliedTerm());
 630
 631         LOG.info("Removed in-memory snapshotted entries, adjusted snaphsotIndex:{} " +
 632             "and term:{}", captureSnapshot.getLastAppliedIndex(),
 633             captureSnapshot.getLastAppliedTerm());
 634
 635         if (isLeader() && captureSnapshot.isInstallSnapshotInitiated()) {
 636             // this would be call straight to the leader and won't initiate in serialization
 637             currentBehavior.handleMessage(getSelf(), new SendInstallSnapshot(stateInBytes));
 638         }
 639
 640         captureSnapshot = null;
 641         hasSnapshotCaptureInitiated = false;
 642     }
 643
 644     private class ReplicatedLogImpl extends AbstractReplicatedLogImpl {
 645
 646         public ReplicatedLogImpl(Snapshot snapshot) {
 647             super(snapshot.getLastAppliedIndex(), snapshot.getLastAppliedTerm(),
 648                 snapshot.getUnAppliedEntries());
 649         }
 650
 651         public ReplicatedLogImpl() {
 652             super();
 653         }
 654
 655         @Override public void removeFromAndPersist(long logEntryIndex) {
 656             int adjustedIndex = adjustedIndex(logEntryIndex);
 657
 658             if (adjustedIndex < 0) {
 659                 return;
 660             }
 661
 662             // FIXME: Maybe this should be done after the command is saved
 663             journal.subList(adjustedIndex , journal.size()).clear();
 664
 665             persistence().persist(new DeleteEntries(adjustedIndex), new Procedure<DeleteEntries>(){
 666
 667                 @Override public void apply(DeleteEntries param)
 668                     throws Exception {
 669                     //FIXME : Doing nothing for now
 670                 }
 671             });
 672         }
 673
 674         @Override public void appendAndPersist(
 675             final ReplicatedLogEntry replicatedLogEntry) {
 676             appendAndPersist(null, null, replicatedLogEntry);
 677         }
 678
 679         public void appendAndPersist(final ActorRef clientActor,
 680             final String identifier,
 681             final ReplicatedLogEntry replicatedLogEntry) {
 682
 683             if(LOG.isDebugEnabled()) {
 684                 LOG.debug("Append log entry and persist {} ", replicatedLogEntry);
 685             }
 686
 687             // FIXME : By adding the replicated log entry to the in-memory journal we are not truly ensuring durability of the logs
 688             journal.add(replicatedLogEntry);
 689
 690             // When persisting events with persist it is guaranteed that the
 691             // persistent actor will not receive further commands between the
 692             // persist call and the execution(s) of the associated event
 693             // handler. This also holds for multiple persist calls in context
 694             // of a single command.
 695             persistence().persist(replicatedLogEntry,
 696                 new Procedure<ReplicatedLogEntry>() {
 697                     @Override
 698                     public void apply(ReplicatedLogEntry evt) throws Exception {
 699                         // when a snaphsot is being taken, captureSnapshot != null
 700                         if (hasSnapshotCaptureInitiated == false &&
 701                             journal.size() % context.getConfigParams().getSnapshotBatchCount() == 0) {
 702
 703                             LOG.info("Initiating Snapshot Capture..");
 704                             long lastAppliedIndex = -1;
 705                             long lastAppliedTerm = -1;
 706
 707                             ReplicatedLogEntry lastAppliedEntry = get(context.getLastApplied());
 708                             if (lastAppliedEntry != null) {
 709                                 lastAppliedIndex = lastAppliedEntry.getIndex();
 710                                 lastAppliedTerm = lastAppliedEntry.getTerm();
 711                             }
 712
 713                             if(LOG.isDebugEnabled()) {
 714                                 LOG.debug("Snapshot Capture logSize: {}", journal.size());
 715                                 LOG.debug("Snapshot Capture lastApplied:{} ",
 716                                     context.getLastApplied());
 717                                 LOG.debug("Snapshot Capture lastAppliedIndex:{}", lastAppliedIndex);
 718                                 LOG.debug("Snapshot Capture lastAppliedTerm:{}", lastAppliedTerm);
 719                             }
 720
 721                             // send a CaptureSnapshot to self to make the expensive operation async.
 722                             getSelf().tell(new CaptureSnapshot(
 723                                 lastIndex(), lastTerm(), lastAppliedIndex, lastAppliedTerm),
 724                                 null);
 725                             hasSnapshotCaptureInitiated = true;
 726                         }
 727                         // Send message for replication
 728                         if (clientActor != null) {
 729                             currentBehavior.handleMessage(getSelf(),
 730                                 new Replicate(clientActor, identifier,
 731                                     replicatedLogEntry)
 732                             );
 733                         }
 734                     }
 735                 }
 736             );
 737         }
 738
 739     }
 740
 741     static class DeleteEntries implements Serializable {
 742         private static final long serialVersionUID = 1L;
 743         private final int fromIndex;
 744
 745         public DeleteEntries(int fromIndex) {
 746             this.fromIndex = fromIndex;
 747         }
 748
 749         public int getFromIndex() {
 750             return fromIndex;
 751         }
 752     }
 753
 754
 755     private class ElectionTermImpl implements ElectionTerm {
 756         /**
 757          * Identifier of the actor whose election term information this is
 758          */
 759         private long currentTerm = 0;
 760         private String votedFor = null;
 761
 762         @Override
 763         public long getCurrentTerm() {
 764             return currentTerm;
 765         }
 766
 767         @Override
 768         public String getVotedFor() {
 769             return votedFor;
 770         }
 771
 772         @Override public void update(long currentTerm, String votedFor) {
 773             if(LOG.isDebugEnabled()) {
 774                 LOG.debug("Set currentTerm={}, votedFor={}", currentTerm, votedFor);
 775             }
 776             this.currentTerm = currentTerm;
 777             this.votedFor = votedFor;
 778         }
 779
 780         @Override
 781         public void updateAndPersist(long currentTerm, String votedFor){
 782             update(currentTerm, votedFor);
 783             // FIXME : Maybe first persist then update the state
 784             persistence().persist(new UpdateElectionTerm(this.currentTerm, this.votedFor), new Procedure<UpdateElectionTerm>(){
 785
 786                 @Override public void apply(UpdateElectionTerm param)
 787                     throws Exception {
 788
 789                 }
 790             });
 791         }
 792     }
 793
 794     static class UpdateElectionTerm implements Serializable {
 795         private static final long serialVersionUID = 1L;
 796         private final long currentTerm;
 797         private final String votedFor;
 798
 799         public UpdateElectionTerm(long currentTerm, String votedFor) {
 800             this.currentTerm = currentTerm;
 801             this.votedFor = votedFor;
 802         }
 803
 804         public long getCurrentTerm() {
 805             return currentTerm;
 806         }
 807
 808         public String getVotedFor() {
 809             return votedFor;
 810         }
 811     }
 812
 813     protected class NonPersistentRaftDataProvider extends NonPersistentDataProvider {
 814
 815         public NonPersistentRaftDataProvider(){
 816
 817         }
 818
 819         /**
 820          * The way snapshotting works is,
 821          * <ol>
 822          * <li> RaftActor calls createSnapshot on the Shard
 823          * <li> Shard sends a CaptureSnapshotReply and RaftActor then calls saveSnapshot
 824          * <li> When saveSnapshot is invoked on the akka-persistence API it uses the SnapshotStore to save the snapshot.
 825          * The SnapshotStore sends SaveSnapshotSuccess or SaveSnapshotFailure. When the RaftActor gets SaveSnapshot
 826          * success it commits the snapshot to the in-memory journal. This commitSnapshot is mimicking what is done
 827          * in SaveSnapshotSuccess.
 828          * </ol>
 829          * @param o
 830          */
 831         @Override
 832         public void saveSnapshot(Object o) {
 833             // Make saving Snapshot successful
 834             commitSnapshot(-1L);
 835         }
 836     }
 837
 838     @VisibleForTesting
 839     void setCurrentBehavior(AbstractRaftActorBehavior behavior) {
 840         currentBehavior = behavior;
 841     }
 842
 843     protected RaftActorBehavior getCurrentBehavior() {
 844         return currentBehavior;
 845     }
 846 }