opendaylight/md-sal/sal-akka-raft/src/main/java/org/opendaylight/controller/cluster/raft/behaviors/AbstractLeader.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.behaviors;
  10
  11 import akka.actor.ActorRef;
  12 import akka.actor.ActorSelection;
  13 import akka.actor.Cancellable;
  14 import com.google.common.annotations.VisibleForTesting;
  15 import com.google.common.base.Optional;
  16 import com.google.common.base.Preconditions;
  17 import com.google.common.collect.ImmutableMap;
  18 import com.google.common.collect.ImmutableMap.Builder;
  19 import com.google.protobuf.ByteString;
  20 import java.io.IOException;
  21 import java.util.Collection;
  22 import java.util.Collections;
  23 import java.util.HashMap;
  24 import java.util.Iterator;
  25 import java.util.LinkedList;
  26 import java.util.List;
  27 import java.util.Map;
  28 import java.util.Map.Entry;
  29 import org.opendaylight.controller.cluster.raft.ClientRequestTracker;
  30 import org.opendaylight.controller.cluster.raft.ClientRequestTrackerImpl;
  31 import org.opendaylight.controller.cluster.raft.FollowerLogInformation;
  32 import org.opendaylight.controller.cluster.raft.FollowerLogInformationImpl;
  33 import org.opendaylight.controller.cluster.raft.RaftActorContext;
  34 import org.opendaylight.controller.cluster.raft.RaftState;
  35 import org.opendaylight.controller.cluster.raft.ReplicatedLogEntry;
  36 import org.opendaylight.controller.cluster.raft.base.messages.CaptureSnapshot;
  37 import org.opendaylight.controller.cluster.raft.base.messages.Replicate;
  38 import org.opendaylight.controller.cluster.raft.base.messages.SendHeartBeat;
  39 import org.opendaylight.controller.cluster.raft.base.messages.SendInstallSnapshot;
  40 import org.opendaylight.controller.cluster.raft.messages.AppendEntries;
  41 import org.opendaylight.controller.cluster.raft.messages.AppendEntriesReply;
  42 import org.opendaylight.controller.cluster.raft.messages.InstallSnapshot;
  43 import org.opendaylight.controller.cluster.raft.messages.InstallSnapshotReply;
  44 import org.opendaylight.controller.cluster.raft.messages.RaftRPC;
  45 import org.opendaylight.controller.cluster.raft.messages.RequestVoteReply;
  46 import scala.concurrent.duration.FiniteDuration;
  47
  48 /**
  49  * The behavior of a RaftActor when it is in the Leader state
  50  * <p/>
  51  * Leaders:
  52  * <ul>
  53  * <li> Upon election: send initial empty AppendEntries RPCs
  54  * (heartbeat) to each server; repeat during idle periods to
  55  * prevent election timeouts (§5.2)
  56  * <li> If command received from client: append entry to local log,
  57  * respond after entry applied to state machine (§5.3)
  58  * <li> If last log index ≥ nextIndex for a follower: send
  59  * AppendEntries RPC with log entries starting at nextIndex
  60  * <ul>
  61  * <li> If successful: update nextIndex and matchIndex for
  62  * follower (§5.3)
  63  * <li> If AppendEntries fails because of log inconsistency:
  64  * decrement nextIndex and retry (§5.3)
  65  * </ul>
  66  * <li> If there exists an N such that N > commitIndex, a majority
  67  * of matchIndex[i] ≥ N, and log[N].term == currentTerm:
  68  * set commitIndex = N (§5.3, §5.4).
  69  */
  70 public abstract class AbstractLeader extends AbstractRaftActorBehavior {
  71
  72     // The index of the first chunk that is sent when installing a snapshot
  73     public static final int FIRST_CHUNK_INDEX = 1;
  74
  75     // The index that the follower should respond with if it needs the install snapshot to be reset
  76     public static final int INVALID_CHUNK_INDEX = -1;
  77
  78     // This would be passed as the hash code of the last chunk when sending the first chunk
  79     public static final int INITIAL_LAST_CHUNK_HASH_CODE = -1;
  80
  81     private final Map<String, FollowerLogInformation> followerToLog;
  82     private final Map<String, FollowerToSnapshot> mapFollowerToSnapshot = new HashMap<>();
  83
  84     private Cancellable heartbeatSchedule = null;
  85
  86     private final Collection<ClientRequestTracker> trackerList = new LinkedList<>();
  87
  88     protected final int minReplicationCount;
  89
  90     protected final int minIsolatedLeaderPeerCount;
  91
  92     private Optional<ByteString> snapshot;
  93
  94     public AbstractLeader(RaftActorContext context) {
  95         super(context, RaftState.Leader);
  96
  97         final Builder<String, FollowerLogInformation> ftlBuilder = ImmutableMap.builder();
  98         for (String followerId : context.getPeerAddresses().keySet()) {
  99             FollowerLogInformation followerLogInformation =
 100                 new FollowerLogInformationImpl(followerId,
 101                     context.getCommitIndex(), -1,
 102                     context.getConfigParams().getElectionTimeOutInterval());
 103
 104             ftlBuilder.put(followerId, followerLogInformation);
 105         }
 106         followerToLog = ftlBuilder.build();
 107
 108         leaderId = context.getId();
 109
 110         LOG.debug("{}: Election: Leader has following peers: {}", logName(), getFollowerIds());
 111
 112         minReplicationCount = getMajorityVoteCount(getFollowerIds().size());
 113
 114         // the isolated Leader peer count will be 1 less than the majority vote count.
 115         // this is because the vote count has the self vote counted in it
 116         // for e.g
 117         // 0 peers = 1 votesRequired , minIsolatedLeaderPeerCount = 0
 118         // 2 peers = 2 votesRequired , minIsolatedLeaderPeerCount = 1
 119         // 4 peers = 3 votesRequired, minIsolatedLeaderPeerCount = 2
 120         minIsolatedLeaderPeerCount = minReplicationCount > 0 ? (minReplicationCount - 1) : 0;
 121
 122         snapshot = Optional.absent();
 123
 124         // Immediately schedule a heartbeat
 125         // Upon election: send initial empty AppendEntries RPCs
 126         // (heartbeat) to each server; repeat during idle periods to
 127         // prevent election timeouts (§5.2)
 128         sendAppendEntries(0, false);
 129     }
 130
 131     /**
 132      * Return an immutable collection of follower identifiers.
 133      *
 134      * @return Collection of follower IDs
 135      */
 136     protected final Collection<String> getFollowerIds() {
 137         return followerToLog.keySet();
 138     }
 139
 140     @VisibleForTesting
 141     void setSnapshot(Optional<ByteString> snapshot) {
 142         this.snapshot = snapshot;
 143     }
 144
 145     @Override
 146     protected RaftActorBehavior handleAppendEntries(ActorRef sender,
 147         AppendEntries appendEntries) {
 148
 149         LOG.debug("{}: handleAppendEntries: {}", logName(), appendEntries);
 150
 151         return this;
 152     }
 153
 154     @Override
 155     protected RaftActorBehavior handleAppendEntriesReply(ActorRef sender,
 156         AppendEntriesReply appendEntriesReply) {
 157
 158         if(LOG.isTraceEnabled()) {
 159             LOG.trace("{}: handleAppendEntriesReply: {}", logName(), appendEntriesReply);
 160         } else if(LOG.isDebugEnabled() && !appendEntriesReply.isSuccess()) {
 161             LOG.debug("{}: handleAppendEntriesReply: {}", logName(), appendEntriesReply);
 162         }
 163
 164         // Update the FollowerLogInformation
 165         String followerId = appendEntriesReply.getFollowerId();
 166         FollowerLogInformation followerLogInformation =
 167             followerToLog.get(followerId);
 168
 169         if(followerLogInformation == null){
 170             LOG.error("{}: handleAppendEntriesReply - unknown follower {}", logName(), followerId);
 171             return this;
 172         }
 173
 174         followerLogInformation.markFollowerActive();
 175
 176         if (appendEntriesReply.isSuccess()) {
 177             followerLogInformation
 178                 .setMatchIndex(appendEntriesReply.getLogLastIndex());
 179             followerLogInformation
 180                 .setNextIndex(appendEntriesReply.getLogLastIndex() + 1);
 181         } else {
 182
 183             // TODO: When we find that the follower is out of sync with the
 184             // Leader we simply decrement that followers next index by 1.
 185             // Would it be possible to do better than this? The RAFT spec
 186             // does not explicitly deal with it but may be something for us to
 187             // think about
 188
 189             followerLogInformation.decrNextIndex();
 190         }
 191
 192         // Now figure out if this reply warrants a change in the commitIndex
 193         // If there exists an N such that N > commitIndex, a majority
 194         // of matchIndex[i] ≥ N, and log[N].term == currentTerm:
 195         // set commitIndex = N (§5.3, §5.4).
 196         for (long N = context.getCommitIndex() + 1; ; N++) {
 197             int replicatedCount = 1;
 198
 199             for (FollowerLogInformation info : followerToLog.values()) {
 200                 if (info.getMatchIndex() >= N) {
 201                     replicatedCount++;
 202                 }
 203             }
 204
 205             if (replicatedCount >= minReplicationCount) {
 206                 ReplicatedLogEntry replicatedLogEntry = context.getReplicatedLog().get(N);
 207                 if (replicatedLogEntry != null &&
 208                     replicatedLogEntry.getTerm() == currentTerm()) {
 209                     context.setCommitIndex(N);
 210                 }
 211             } else {
 212                 break;
 213             }
 214         }
 215
 216         // Apply the change to the state machine
 217         if (context.getCommitIndex() > context.getLastApplied()) {
 218             LOG.debug("{}: handleAppendEntriesReply: applying to log - commitIndex: {}, lastAppliedIndex: {}",
 219                     logName(), context.getCommitIndex(), context.getLastApplied());
 220
 221             applyLogToStateMachine(context.getCommitIndex());
 222         }
 223
 224         if (!context.isSnapshotCaptureInitiated()) {
 225             purgeInMemoryLog();
 226         }
 227
 228         //Send the next log entry immediately, if possible, no need to wait for heartbeat to trigger that event
 229         sendUpdatesToFollower(followerId, followerLogInformation, false, false);
 230         return this;
 231     }
 232
 233     private void purgeInMemoryLog() {
 234         //find the lowest index across followers which has been replicated to all.
 235         // lastApplied if there are no followers, so that we keep clearing the log for single-node
 236         // we would delete the in-mem log from that index on, in-order to minimize mem usage
 237         // we would also share this info thru AE with the followers so that they can delete their log entries as well.
 238         long minReplicatedToAllIndex = followerToLog.isEmpty() ? context.getLastApplied() : Long.MAX_VALUE;
 239         for (FollowerLogInformation info : followerToLog.values()) {
 240             minReplicatedToAllIndex = Math.min(minReplicatedToAllIndex, info.getMatchIndex());
 241         }
 242
 243         super.performSnapshotWithoutCapture(minReplicatedToAllIndex);
 244     }
 245
 246     @Override
 247     protected ClientRequestTracker removeClientRequestTracker(long logIndex) {
 248         final Iterator<ClientRequestTracker> it = trackerList.iterator();
 249         while (it.hasNext()) {
 250             final ClientRequestTracker t = it.next();
 251             if (t.getIndex() == logIndex) {
 252                 it.remove();
 253                 return t;
 254             }
 255         }
 256
 257         return null;
 258     }
 259
 260     @Override
 261     protected ClientRequestTracker findClientRequestTracker(long logIndex) {
 262         for (ClientRequestTracker tracker : trackerList) {
 263             if (tracker.getIndex() == logIndex) {
 264                 return tracker;
 265             }
 266         }
 267         return null;
 268     }
 269
 270     @Override
 271     protected RaftActorBehavior handleRequestVoteReply(ActorRef sender,
 272         RequestVoteReply requestVoteReply) {
 273         return this;
 274     }
 275
 276     @Override
 277     public RaftActorBehavior handleMessage(ActorRef sender, Object originalMessage) {
 278         Preconditions.checkNotNull(sender, "sender should not be null");
 279
 280         Object message = fromSerializableMessage(originalMessage);
 281
 282         if (message instanceof RaftRPC) {
 283             RaftRPC rpc = (RaftRPC) message;
 284             // If RPC request or response contains term T > currentTerm:
 285             // set currentTerm = T, convert to follower (§5.1)
 286             // This applies to all RPC messages and responses
 287             if (rpc.getTerm() > context.getTermInformation().getCurrentTerm()) {
 288                 LOG.debug("{}: Term {} in \"{}\" message is greater than leader's term {} - switching to Follower",
 289                         logName(), rpc.getTerm(), rpc, context.getTermInformation().getCurrentTerm());
 290
 291                 context.getTermInformation().updateAndPersist(rpc.getTerm(), null);
 292
 293                 return switchBehavior(new Follower(context));
 294             }
 295         }
 296
 297         try {
 298             if (message instanceof SendHeartBeat) {
 299                 sendHeartBeat();
 300                 return this;
 301
 302             } else if(message instanceof SendInstallSnapshot) {
 303                 // received from RaftActor
 304                 setSnapshot(Optional.of(((SendInstallSnapshot) message).getSnapshot()));
 305                 sendInstallSnapshot();
 306
 307             } else if (message instanceof Replicate) {
 308                 replicate((Replicate) message);
 309
 310             } else if (message instanceof InstallSnapshotReply){
 311                 handleInstallSnapshotReply((InstallSnapshotReply) message);
 312
 313             }
 314         } finally {
 315             scheduleHeartBeat(context.getConfigParams().getHeartBeatInterval());
 316         }
 317
 318         return super.handleMessage(sender, message);
 319     }
 320
 321     private void handleInstallSnapshotReply(InstallSnapshotReply reply) {
 322         LOG.debug("{}: handleInstallSnapshotReply: {}", logName(), reply);
 323
 324         String followerId = reply.getFollowerId();
 325         FollowerToSnapshot followerToSnapshot = mapFollowerToSnapshot.get(followerId);
 326
 327         if (followerToSnapshot == null) {
 328             LOG.error("{}: FollowerId {} in InstallSnapshotReply not known to Leader",
 329                     logName(), followerId);
 330             return;
 331         }
 332
 333         FollowerLogInformation followerLogInformation = followerToLog.get(followerId);
 334         followerLogInformation.markFollowerActive();
 335
 336         if (followerToSnapshot.getChunkIndex() == reply.getChunkIndex()) {
 337             boolean wasLastChunk = false;
 338             if (reply.isSuccess()) {
 339                 if(followerToSnapshot.isLastChunk(reply.getChunkIndex())) {
 340                     //this was the last chunk reply
 341                     if(LOG.isDebugEnabled()) {
 342                         LOG.debug("{}: InstallSnapshotReply received, " +
 343                                 "last chunk received, Chunk: {}. Follower: {} Setting nextIndex: {}",
 344                                 logName(), reply.getChunkIndex(), followerId,
 345                             context.getReplicatedLog().getSnapshotIndex() + 1
 346                         );
 347                     }
 348
 349                     followerLogInformation.setMatchIndex(
 350                         context.getReplicatedLog().getSnapshotIndex());
 351                     followerLogInformation.setNextIndex(
 352                         context.getReplicatedLog().getSnapshotIndex() + 1);
 353                     mapFollowerToSnapshot.remove(followerId);
 354
 355                     LOG.debug("{}: follower: {}, matchIndex set to {}, nextIndex set to {}",
 356                                 logName(), followerId, followerLogInformation.getMatchIndex(),
 357                                 followerLogInformation.getNextIndex());
 358
 359                     if (mapFollowerToSnapshot.isEmpty()) {
 360                         // once there are no pending followers receiving snapshots
 361                         // we can remove snapshot from the memory
 362                         setSnapshot(Optional.<ByteString>absent());
 363                     }
 364                     wasLastChunk = true;
 365
 366                 } else {
 367                     followerToSnapshot.markSendStatus(true);
 368                 }
 369             } else {
 370                 LOG.info("{}: InstallSnapshotReply received sending snapshot chunk failed, Will retry, Chunk: {}",
 371                         logName(), reply.getChunkIndex());
 372
 373                 followerToSnapshot.markSendStatus(false);
 374             }
 375
 376             if (!wasLastChunk && followerToSnapshot.canSendNextChunk()) {
 377                 ActorSelection followerActor = context.getPeerActorSelection(followerId);
 378                 if(followerActor != null) {
 379                     sendSnapshotChunk(followerActor, followerId);
 380                 }
 381             }
 382
 383         } else {
 384             LOG.error("{}: Chunk index {} in InstallSnapshotReply from follower {} does not match expected index {}",
 385                     logName(), reply.getChunkIndex(), followerId,
 386                     followerToSnapshot.getChunkIndex());
 387
 388             if(reply.getChunkIndex() == INVALID_CHUNK_INDEX){
 389                 // Since the Follower did not find this index to be valid we should reset the follower snapshot
 390                 // so that Installing the snapshot can resume from the beginning
 391                 followerToSnapshot.reset();
 392             }
 393         }
 394     }
 395
 396     private void replicate(Replicate replicate) {
 397         long logIndex = replicate.getReplicatedLogEntry().getIndex();
 398
 399         LOG.debug("{}: Replicate message: identifier: {}, logIndex: {}", logName(),
 400                 replicate.getIdentifier(), logIndex);
 401
 402         // Create a tracker entry we will use this later to notify the
 403         // client actor
 404         trackerList.add(
 405             new ClientRequestTrackerImpl(replicate.getClientActor(),
 406                 replicate.getIdentifier(),
 407                 logIndex)
 408         );
 409
 410         if (followerToLog.isEmpty()) {
 411             context.setCommitIndex(logIndex);
 412             applyLogToStateMachine(logIndex);
 413         } else {
 414             sendAppendEntries(0, false);
 415         }
 416     }
 417
 418     private void sendAppendEntries(long timeSinceLastActivityInterval, boolean isHeartbeat) {
 419         // Send an AppendEntries to all followers
 420         for (Entry<String, FollowerLogInformation> e : followerToLog.entrySet()) {
 421             final String followerId = e.getKey();
 422             final FollowerLogInformation followerLogInformation = e.getValue();
 423             // This checks helps not to send a repeat message to the follower
 424             if(!followerLogInformation.isFollowerActive() ||
 425                     followerLogInformation.timeSinceLastActivity() >= timeSinceLastActivityInterval) {
 426                 sendUpdatesToFollower(followerId, followerLogInformation, true, isHeartbeat);
 427             }
 428         }
 429     }
 430
 431     /**
 432      *
 433      * This method checks if any update needs to be sent to the given follower. This includes append log entries,
 434      * sending next snapshot chunk, and initiating a snapshot.
 435      * @return true if any update is sent, false otherwise
 436      */
 437
 438     private void sendUpdatesToFollower(String followerId, FollowerLogInformation followerLogInformation,
 439                                        boolean sendHeartbeat, boolean isHeartbeat) {
 440
 441         ActorSelection followerActor = context.getPeerActorSelection(followerId);
 442         if (followerActor != null) {
 443             long followerNextIndex = followerLogInformation.getNextIndex();
 444             boolean isFollowerActive = followerLogInformation.isFollowerActive();
 445
 446             if (mapFollowerToSnapshot.get(followerId) != null) {
 447                 // if install snapshot is in process , then sent next chunk if possible
 448                 if (isFollowerActive && mapFollowerToSnapshot.get(followerId).canSendNextChunk()) {
 449                     sendSnapshotChunk(followerActor, followerId);
 450                 } else if(sendHeartbeat) {
 451                     // we send a heartbeat even if we have not received a reply for the last chunk
 452                     sendAppendEntriesToFollower(followerActor, followerLogInformation.getNextIndex(),
 453                         Collections.<ReplicatedLogEntry>emptyList(), followerId);
 454                 }
 455             } else {
 456                 long leaderLastIndex = context.getReplicatedLog().lastIndex();
 457                 long leaderSnapShotIndex = context.getReplicatedLog().getSnapshotIndex();
 458
 459                 if(!isHeartbeat || LOG.isTraceEnabled()) {
 460                     LOG.debug("{}: Checking sendAppendEntries for follower {}, leaderLastIndex: {}, leaderSnapShotIndex: {}",
 461                             logName(), followerId, leaderLastIndex, leaderSnapShotIndex);
 462                 }
 463
 464                 if (isFollowerActive && context.getReplicatedLog().isPresent(followerNextIndex)) {
 465
 466                     LOG.debug("{}: sendAppendEntries: {} is present for follower {}", logName(),
 467                             followerNextIndex, followerId);
 468
 469                     // FIXME : Sending one entry at a time
 470                     final List<ReplicatedLogEntry> entries = context.getReplicatedLog().getFrom(followerNextIndex, 1);
 471
 472                     sendAppendEntriesToFollower(followerActor, followerNextIndex, entries, followerId);
 473
 474                 } else if (isFollowerActive && followerNextIndex >= 0 &&
 475                     leaderLastIndex > followerNextIndex && !context.isSnapshotCaptureInitiated()) {
 476                     // if the followers next index is not present in the leaders log, and
 477                     // if the follower is just not starting and if leader's index is more than followers index
 478                     // then snapshot should be sent
 479
 480                     if (LOG.isDebugEnabled()) {
 481                         LOG.debug(String.format("%s: InitiateInstallSnapshot to follower: %s," +
 482                                     "follower-nextIndex: %d, leader-snapshot-index: %d,  " +
 483                                     "leader-last-index: %d", logName(), followerId,
 484                                     followerNextIndex, leaderSnapShotIndex, leaderLastIndex));
 485                     }
 486
 487                     // Send heartbeat to follower whenever install snapshot is initiated.
 488                     sendAppendEntriesToFollower(followerActor, followerLogInformation.getNextIndex(),
 489                             Collections.<ReplicatedLogEntry>emptyList(), followerId);
 490
 491                     initiateCaptureSnapshot(followerId, followerNextIndex);
 492
 493                 } else if(sendHeartbeat) {
 494                     //we send an AppendEntries, even if the follower is inactive
 495                     // in-order to update the followers timestamp, in case it becomes active again
 496                     sendAppendEntriesToFollower(followerActor, followerLogInformation.getNextIndex(),
 497                         Collections.<ReplicatedLogEntry>emptyList(), followerId);
 498                 }
 499
 500             }
 501         }
 502     }
 503
 504     private void sendAppendEntriesToFollower(ActorSelection followerActor, long followerNextIndex,
 505         List<ReplicatedLogEntry> entries, String followerId) {
 506         AppendEntries appendEntries = new AppendEntries(currentTerm(), context.getId(),
 507             prevLogIndex(followerNextIndex),
 508             prevLogTerm(followerNextIndex), entries,
 509             context.getCommitIndex(), super.getReplicatedToAllIndex());
 510
 511         if(!entries.isEmpty() || LOG.isTraceEnabled()) {
 512             LOG.debug("{}: Sending AppendEntries to follower {}: {}", logName(), followerId,
 513                     appendEntries);
 514         }
 515
 516         followerActor.tell(appendEntries.toSerializable(), actor());
 517     }
 518
 519     /**
 520      * Install Snapshot works as follows
 521      * 1. Leader initiates the capture snapshot by sending a CaptureSnapshot message to actor
 522      * 2. RaftActor on receipt of the CaptureSnapshotReply (from Shard), stores the received snapshot in the replicated log
 523      * and makes a call to Leader's handleMessage , with SendInstallSnapshot message.
 524      * 3. Leader , picks the snapshot from im-mem ReplicatedLog and sends it in chunks to the Follower
 525      * 4. On complete, Follower sends back a InstallSnapshotReply.
 526      * 5. On receipt of the InstallSnapshotReply for the last chunk, Leader marks the install complete for that follower
 527      * and replenishes the memory by deleting the snapshot in Replicated log.
 528      * 6. If another follower requires a snapshot and a snapshot has been collected (via CaptureSnapshotReply)
 529      * then send the existing snapshot in chunks to the follower.
 530      * @param followerId
 531      * @param followerNextIndex
 532      */
 533     private void initiateCaptureSnapshot(String followerId, long followerNextIndex) {
 534         if (!context.getReplicatedLog().isPresent(followerNextIndex) &&
 535                 context.getReplicatedLog().isInSnapshot(followerNextIndex)) {
 536
 537             if (snapshot.isPresent()) {
 538                 // if a snapshot is present in the memory, most likely another install is in progress
 539                 // no need to capture snapshot.
 540                 // This could happen if another follower needs an install when one is going on.
 541                 final ActorSelection followerActor = context.getPeerActorSelection(followerId);
 542                 sendSnapshotChunk(followerActor, followerId);
 543
 544             } else if (!context.isSnapshotCaptureInitiated()) {
 545
 546                 LOG.info("{}: Initiating Snapshot Capture to Install Snapshot, Leader:{}", logName(), getLeaderId());
 547                 ReplicatedLogEntry lastAppliedEntry = context.getReplicatedLog().get(context.getLastApplied());
 548                 long lastAppliedIndex = -1;
 549                 long lastAppliedTerm = -1;
 550
 551                 if (lastAppliedEntry != null) {
 552                     lastAppliedIndex = lastAppliedEntry.getIndex();
 553                     lastAppliedTerm = lastAppliedEntry.getTerm();
 554                 } else if (context.getReplicatedLog().getSnapshotIndex() > -1) {
 555                     lastAppliedIndex = context.getReplicatedLog().getSnapshotIndex();
 556                     lastAppliedTerm = context.getReplicatedLog().getSnapshotTerm();
 557                 }
 558
 559                 boolean isInstallSnapshotInitiated = true;
 560                 long replicatedToAllIndex = super.getReplicatedToAllIndex();
 561                 ReplicatedLogEntry replicatedToAllEntry = context.getReplicatedLog().get(replicatedToAllIndex);
 562                 actor().tell(new CaptureSnapshot(lastIndex(), lastTerm(), lastAppliedIndex, lastAppliedTerm,
 563                     (replicatedToAllEntry != null ? replicatedToAllEntry.getIndex() : -1),
 564                     (replicatedToAllEntry != null ? replicatedToAllEntry.getTerm() : -1),
 565                     isInstallSnapshotInitiated), actor());
 566                 context.setSnapshotCaptureInitiated(true);
 567             }
 568         }
 569     }
 570
 571
 572     private void sendInstallSnapshot() {
 573         LOG.debug("{}: sendInstallSnapshot", logName());
 574         for (Entry<String, FollowerLogInformation> e : followerToLog.entrySet()) {
 575             ActorSelection followerActor = context.getPeerActorSelection(e.getKey());
 576
 577             if (followerActor != null) {
 578                 long nextIndex = e.getValue().getNextIndex();
 579
 580                 if (!context.getReplicatedLog().isPresent(nextIndex) &&
 581                     context.getReplicatedLog().isInSnapshot(nextIndex)) {
 582                     sendSnapshotChunk(followerActor, e.getKey());
 583                 }
 584             }
 585         }
 586     }
 587
 588     /**
 589      *  Sends a snapshot chunk to a given follower
 590      *  InstallSnapshot should qualify as a heartbeat too.
 591      */
 592     private void sendSnapshotChunk(ActorSelection followerActor, String followerId) {
 593         try {
 594             if (snapshot.isPresent()) {
 595                 ByteString nextSnapshotChunk = getNextSnapshotChunk(followerId,snapshot.get());
 596
 597                 // Note: the previous call to getNextSnapshotChunk has the side-effect of adding
 598                 // followerId to the followerToSnapshot map.
 599                 FollowerToSnapshot followerToSnapshot = mapFollowerToSnapshot.get(followerId);
 600
 601                 followerActor.tell(
 602                     new InstallSnapshot(currentTerm(), context.getId(),
 603                         context.getReplicatedLog().getSnapshotIndex(),
 604                         context.getReplicatedLog().getSnapshotTerm(),
 605                         nextSnapshotChunk,
 606                         followerToSnapshot.incrementChunkIndex(),
 607                         followerToSnapshot.getTotalChunks(),
 608                         Optional.of(followerToSnapshot.getLastChunkHashCode())
 609                     ).toSerializable(),
 610                     actor()
 611                 );
 612                 LOG.info("{}: InstallSnapshot sent to follower {}, Chunk: {}/{}",
 613                         logName(), followerActor.path(),
 614                         followerToSnapshot.getChunkIndex(),
 615                         followerToSnapshot.getTotalChunks());
 616             }
 617         } catch (IOException e) {
 618             LOG.error("{}: InstallSnapshot failed for Leader.", logName(), e);
 619         }
 620     }
 621
 622     /**
 623      * Acccepts snaphot as ByteString, enters into map for future chunks
 624      * creates and return a ByteString chunk
 625      */
 626     private ByteString getNextSnapshotChunk(String followerId, ByteString snapshotBytes) throws IOException {
 627         FollowerToSnapshot followerToSnapshot = mapFollowerToSnapshot.get(followerId);
 628         if (followerToSnapshot == null) {
 629             followerToSnapshot = new FollowerToSnapshot(snapshotBytes);
 630             mapFollowerToSnapshot.put(followerId, followerToSnapshot);
 631         }
 632         ByteString nextChunk = followerToSnapshot.getNextChunk();
 633
 634         LOG.debug("{}: next snapshot chunk size for follower {}: {}", logName(), followerId, nextChunk.size());
 635
 636         return nextChunk;
 637     }
 638
 639     private void sendHeartBeat() {
 640         if (!followerToLog.isEmpty()) {
 641             LOG.trace("{}: Sending heartbeat", logName());
 642             sendAppendEntries(context.getConfigParams().getHeartBeatInterval().toMillis(), true);
 643         }
 644     }
 645
 646     private void stopHeartBeat() {
 647         if (heartbeatSchedule != null && !heartbeatSchedule.isCancelled()) {
 648             heartbeatSchedule.cancel();
 649         }
 650     }
 651
 652     private void scheduleHeartBeat(FiniteDuration interval) {
 653         if (followerToLog.isEmpty()) {
 654             // Optimization - do not bother scheduling a heartbeat as there are
 655             // no followers
 656             return;
 657         }
 658
 659         stopHeartBeat();
 660
 661         // Schedule a heartbeat. When the scheduler triggers a SendHeartbeat
 662         // message is sent to itself.
 663         // Scheduling the heartbeat only once here because heartbeats do not
 664         // need to be sent if there are other messages being sent to the remote
 665         // actor.
 666         heartbeatSchedule = context.getActorSystem().scheduler().scheduleOnce(
 667             interval, context.getActor(), new SendHeartBeat(),
 668             context.getActorSystem().dispatcher(), context.getActor());
 669     }
 670
 671     @Override
 672     public void close() throws Exception {
 673         stopHeartBeat();
 674     }
 675
 676     @Override
 677     public String getLeaderId() {
 678         return context.getId();
 679     }
 680
 681     protected boolean isLeaderIsolated() {
 682         int minPresent = minIsolatedLeaderPeerCount;
 683         for (FollowerLogInformation followerLogInformation : followerToLog.values()) {
 684             if (followerLogInformation.isFollowerActive()) {
 685                 --minPresent;
 686                 if (minPresent == 0) {
 687                     break;
 688                 }
 689             }
 690         }
 691         return (minPresent != 0);
 692     }
 693
 694     /**
 695      * Encapsulates the snapshot bytestring and handles the logic of sending
 696      * snapshot chunks
 697      */
 698     protected class FollowerToSnapshot {
 699         private final ByteString snapshotBytes;
 700         private int offset = 0;
 701         // the next snapshot chunk is sent only if the replyReceivedForOffset matches offset
 702         private int replyReceivedForOffset;
 703         // if replyStatus is false, the previous chunk is attempted
 704         private boolean replyStatus = false;
 705         private int chunkIndex;
 706         private final int totalChunks;
 707         private int lastChunkHashCode = AbstractLeader.INITIAL_LAST_CHUNK_HASH_CODE;
 708         private int nextChunkHashCode = AbstractLeader.INITIAL_LAST_CHUNK_HASH_CODE;
 709
 710         public FollowerToSnapshot(ByteString snapshotBytes) {
 711             this.snapshotBytes = snapshotBytes;
 712             int size = snapshotBytes.size();
 713             totalChunks = ( size / context.getConfigParams().getSnapshotChunkSize()) +
 714                 ((size % context.getConfigParams().getSnapshotChunkSize()) > 0 ? 1 : 0);
 715             if(LOG.isDebugEnabled()) {
 716                 LOG.debug("{}: Snapshot {} bytes, total chunks to send:{}",
 717                         logName(), size, totalChunks);
 718             }
 719             replyReceivedForOffset = -1;
 720             chunkIndex = AbstractLeader.FIRST_CHUNK_INDEX;
 721         }
 722
 723         public ByteString getSnapshotBytes() {
 724             return snapshotBytes;
 725         }
 726
 727         public int incrementOffset() {
 728             if(replyStatus) {
 729                 // if prev chunk failed, we would want to sent the same chunk again
 730                 offset = offset + context.getConfigParams().getSnapshotChunkSize();
 731             }
 732             return offset;
 733         }
 734
 735         public int incrementChunkIndex() {
 736             if (replyStatus) {
 737                 // if prev chunk failed, we would want to sent the same chunk again
 738                 chunkIndex =  chunkIndex + 1;
 739             }
 740             return chunkIndex;
 741         }
 742
 743         public int getChunkIndex() {
 744             return chunkIndex;
 745         }
 746
 747         public int getTotalChunks() {
 748             return totalChunks;
 749         }
 750
 751         public boolean canSendNextChunk() {
 752             // we only send a false if a chunk is sent but we have not received a reply yet
 753             return replyReceivedForOffset == offset;
 754         }
 755
 756         public boolean isLastChunk(int chunkIndex) {
 757             return totalChunks == chunkIndex;
 758         }
 759
 760         public void markSendStatus(boolean success) {
 761             if (success) {
 762                 // if the chunk sent was successful
 763                 replyReceivedForOffset = offset;
 764                 replyStatus = true;
 765                 lastChunkHashCode = nextChunkHashCode;
 766             } else {
 767                 // if the chunk sent was failure
 768                 replyReceivedForOffset = offset;
 769                 replyStatus = false;
 770             }
 771         }
 772
 773         public ByteString getNextChunk() {
 774             int snapshotLength = getSnapshotBytes().size();
 775             int start = incrementOffset();
 776             int size = context.getConfigParams().getSnapshotChunkSize();
 777             if (context.getConfigParams().getSnapshotChunkSize() > snapshotLength) {
 778                 size = snapshotLength;
 779             } else {
 780                 if ((start + context.getConfigParams().getSnapshotChunkSize()) > snapshotLength) {
 781                     size = snapshotLength - start;
 782                 }
 783             }
 784
 785
 786             LOG.debug("{}: Next chunk: length={}, offset={},size={}", logName(),
 787                     snapshotLength, start, size);
 788
 789             ByteString substring = getSnapshotBytes().substring(start, start + size);
 790             nextChunkHashCode = substring.hashCode();
 791             return substring;
 792         }
 793
 794         /**
 795          * reset should be called when the Follower needs to be sent the snapshot from the beginning
 796          */
 797         public void reset(){
 798             offset = 0;
 799             replyStatus = false;
 800             replyReceivedForOffset = offset;
 801             chunkIndex = AbstractLeader.FIRST_CHUNK_INDEX;
 802             lastChunkHashCode = AbstractLeader.INITIAL_LAST_CHUNK_HASH_CODE;
 803         }
 804
 805         public int getLastChunkHashCode() {
 806             return lastChunkHashCode;
 807         }
 808     }
 809
 810     // called from example-actor for printing the follower-states
 811     public String printFollowerStates() {
 812         final StringBuilder sb = new StringBuilder();
 813
 814         sb.append('[');
 815         for (FollowerLogInformation followerLogInformation : followerToLog.values()) {
 816             sb.append('{');
 817             sb.append(followerLogInformation.getId());
 818             sb.append(" state:");
 819             sb.append(followerLogInformation.isFollowerActive());
 820             sb.append("},");
 821         }
 822         sb.append(']');
 823
 824         return sb.toString();
 825     }
 826
 827     @VisibleForTesting
 828     public FollowerLogInformation getFollower(String followerId) {
 829         return followerToLog.get(followerId);
 830     }
 831
 832     @VisibleForTesting
 833     protected void setFollowerSnapshot(String followerId, FollowerToSnapshot snapshot) {
 834         mapFollowerToSnapshot.put(followerId, snapshot);
 835     }
 836
 837     @VisibleForTesting
 838     public int followerSnapshotSize() {
 839         return mapFollowerToSnapshot.size();
 840     }
 841
 842     @VisibleForTesting
 843     public int followerLogSize() {
 844         return followerToLog.size();
 845     }
 846 }