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);
  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: {}", context.getId(), 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);
 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     private Optional<ByteString> getSnapshot() {
 141         return snapshot;
 142     }
 143
 144     @VisibleForTesting
 145     void setSnapshot(Optional<ByteString> snapshot) {
 146         this.snapshot = snapshot;
 147     }
 148
 149     @Override
 150     protected RaftActorBehavior handleAppendEntries(ActorRef sender,
 151         AppendEntries appendEntries) {
 152
 153         if(LOG.isDebugEnabled()) {
 154             LOG.debug("{}: handleAppendEntries: {}", context.getId(), appendEntries);
 155         }
 156
 157         return this;
 158     }
 159
 160     @Override
 161     protected RaftActorBehavior handleAppendEntriesReply(ActorRef sender,
 162         AppendEntriesReply appendEntriesReply) {
 163
 164         if(! appendEntriesReply.isSuccess()) {
 165             if(LOG.isDebugEnabled()) {
 166                 LOG.debug("{}: handleAppendEntriesReply: {}", context.getId(), appendEntriesReply);
 167             }
 168         }
 169
 170         // Update the FollowerLogInformation
 171         String followerId = appendEntriesReply.getFollowerId();
 172         FollowerLogInformation followerLogInformation =
 173             followerToLog.get(followerId);
 174
 175         if(followerLogInformation == null){
 176             LOG.error("{}: handleAppendEntriesReply - unknown follower {}", context.getId(), followerId);
 177             return this;
 178         }
 179
 180         followerLogInformation.markFollowerActive();
 181
 182         if (appendEntriesReply.isSuccess()) {
 183             followerLogInformation
 184                 .setMatchIndex(appendEntriesReply.getLogLastIndex());
 185             followerLogInformation
 186                 .setNextIndex(appendEntriesReply.getLogLastIndex() + 1);
 187         } else {
 188
 189             // TODO: When we find that the follower is out of sync with the
 190             // Leader we simply decrement that followers next index by 1.
 191             // Would it be possible to do better than this? The RAFT spec
 192             // does not explicitly deal with it but may be something for us to
 193             // think about
 194
 195             followerLogInformation.decrNextIndex();
 196         }
 197
 198         // Now figure out if this reply warrants a change in the commitIndex
 199         // If there exists an N such that N > commitIndex, a majority
 200         // of matchIndex[i] ≥ N, and log[N].term == currentTerm:
 201         // set commitIndex = N (§5.3, §5.4).
 202         for (long N = context.getCommitIndex() + 1; ; N++) {
 203             int replicatedCount = 1;
 204
 205             for (FollowerLogInformation info : followerToLog.values()) {
 206                 if (info.getMatchIndex() >= N) {
 207                     replicatedCount++;
 208                 }
 209             }
 210
 211             if (replicatedCount >= minReplicationCount) {
 212                 ReplicatedLogEntry replicatedLogEntry = context.getReplicatedLog().get(N);
 213                 if (replicatedLogEntry != null &&
 214                     replicatedLogEntry.getTerm() == currentTerm()) {
 215                     context.setCommitIndex(N);
 216                 }
 217             } else {
 218                 break;
 219             }
 220         }
 221
 222         // Apply the change to the state machine
 223         if (context.getCommitIndex() > context.getLastApplied()) {
 224             applyLogToStateMachine(context.getCommitIndex());
 225         }
 226
 227         if (!context.isSnapshotCaptureInitiated()) {
 228             purgeInMemoryLog();
 229         }
 230
 231         //Send the next log entry immediately, if possible, no need to wait for heartbeat to trigger that event
 232         sendUpdatesToFollower(followerId, followerLogInformation, false);
 233         return this;
 234     }
 235
 236     private void purgeInMemoryLog() {
 237         //find the lowest index across followers which has been replicated to all.
 238         // lastApplied if there are no followers, so that we keep clearing the log for single-node
 239         // we would delete the in-mem log from that index on, in-order to minimize mem usage
 240         // we would also share this info thru AE with the followers so that they can delete their log entries as well.
 241         long minReplicatedToAllIndex = followerToLog.isEmpty() ? context.getLastApplied() : Long.MAX_VALUE;
 242         for (FollowerLogInformation info : followerToLog.values()) {
 243             minReplicatedToAllIndex = Math.min(minReplicatedToAllIndex, info.getMatchIndex());
 244         }
 245
 246         super.performSnapshotWithoutCapture(minReplicatedToAllIndex);
 247     }
 248
 249     @Override
 250     protected ClientRequestTracker removeClientRequestTracker(long logIndex) {
 251         final Iterator<ClientRequestTracker> it = trackerList.iterator();
 252         while (it.hasNext()) {
 253             final ClientRequestTracker t = it.next();
 254             if (t.getIndex() == logIndex) {
 255                 it.remove();
 256                 return t;
 257             }
 258         }
 259
 260         return null;
 261     }
 262
 263     @Override
 264     protected ClientRequestTracker findClientRequestTracker(long logIndex) {
 265         for (ClientRequestTracker tracker : trackerList) {
 266             if (tracker.getIndex() == logIndex) {
 267                 return tracker;
 268             }
 269         }
 270         return null;
 271     }
 272
 273     @Override
 274     protected RaftActorBehavior handleRequestVoteReply(ActorRef sender,
 275         RequestVoteReply requestVoteReply) {
 276         return this;
 277     }
 278
 279     @Override
 280     public RaftState state() {
 281         return RaftState.Leader;
 282     }
 283
 284     @Override
 285     public RaftActorBehavior handleMessage(ActorRef sender, Object originalMessage) {
 286         Preconditions.checkNotNull(sender, "sender should not be null");
 287
 288         Object message = fromSerializableMessage(originalMessage);
 289
 290         if (message instanceof RaftRPC) {
 291             RaftRPC rpc = (RaftRPC) message;
 292             // If RPC request or response contains term T > currentTerm:
 293             // set currentTerm = T, convert to follower (§5.1)
 294             // This applies to all RPC messages and responses
 295             if (rpc.getTerm() > context.getTermInformation().getCurrentTerm()) {
 296                 LOG.debug("{}: Term {} in \"{}\" message is greater than leader's term {}", context.getId(),
 297                         rpc.getTerm(), rpc, context.getTermInformation().getCurrentTerm());
 298
 299                 context.getTermInformation().updateAndPersist(rpc.getTerm(), null);
 300
 301                 return switchBehavior(new Follower(context));
 302             }
 303         }
 304
 305         try {
 306             if (message instanceof SendHeartBeat) {
 307                 sendHeartBeat();
 308                 return this;
 309
 310             } else if(message instanceof SendInstallSnapshot) {
 311                 // received from RaftActor
 312                 setSnapshot(Optional.of(((SendInstallSnapshot) message).getSnapshot()));
 313                 sendInstallSnapshot();
 314
 315             } else if (message instanceof Replicate) {
 316                 replicate((Replicate) message);
 317
 318             } else if (message instanceof InstallSnapshotReply){
 319                 handleInstallSnapshotReply((InstallSnapshotReply) message);
 320
 321             }
 322         } finally {
 323             scheduleHeartBeat(context.getConfigParams().getHeartBeatInterval());
 324         }
 325
 326         return super.handleMessage(sender, message);
 327     }
 328
 329     private void handleInstallSnapshotReply(InstallSnapshotReply reply) {
 330         String followerId = reply.getFollowerId();
 331         FollowerToSnapshot followerToSnapshot = mapFollowerToSnapshot.get(followerId);
 332
 333         if (followerToSnapshot == null) {
 334             LOG.error("{}: FollowerId {} in InstallSnapshotReply not known to Leader",
 335                     context.getId(), followerId);
 336             return;
 337         }
 338
 339         FollowerLogInformation followerLogInformation = followerToLog.get(followerId);
 340         followerLogInformation.markFollowerActive();
 341
 342         if (followerToSnapshot.getChunkIndex() == reply.getChunkIndex()) {
 343             boolean wasLastChunk = false;
 344             if (reply.isSuccess()) {
 345                 if(followerToSnapshot.isLastChunk(reply.getChunkIndex())) {
 346                     //this was the last chunk reply
 347                     if(LOG.isDebugEnabled()) {
 348                         LOG.debug("{}: InstallSnapshotReply received, " +
 349                                 "last chunk received, Chunk:{}. Follower:{} Setting nextIndex:{}",
 350                                 context.getId(), reply.getChunkIndex(), followerId,
 351                             context.getReplicatedLog().getSnapshotIndex() + 1
 352                         );
 353                     }
 354
 355                     followerLogInformation.setMatchIndex(
 356                         context.getReplicatedLog().getSnapshotIndex());
 357                     followerLogInformation.setNextIndex(
 358                         context.getReplicatedLog().getSnapshotIndex() + 1);
 359                     mapFollowerToSnapshot.remove(followerId);
 360
 361                     if(LOG.isDebugEnabled()) {
 362                         LOG.debug("{}: followerToLog.get(followerId).getNextIndex()=" +
 363                                 context.getId(), followerToLog.get(followerId).getNextIndex());
 364                     }
 365
 366                     if (mapFollowerToSnapshot.isEmpty()) {
 367                         // once there are no pending followers receiving snapshots
 368                         // we can remove snapshot from the memory
 369                         setSnapshot(Optional.<ByteString>absent());
 370                     }
 371                     wasLastChunk = true;
 372
 373                 } else {
 374                     followerToSnapshot.markSendStatus(true);
 375                 }
 376             } else {
 377                 LOG.info("{}: InstallSnapshotReply received sending snapshot chunk failed, Will retry, Chunk: {}",
 378                         context.getId(), reply.getChunkIndex());
 379
 380                 followerToSnapshot.markSendStatus(false);
 381             }
 382
 383             if (!wasLastChunk && followerToSnapshot.canSendNextChunk()) {
 384                 ActorSelection followerActor = context.getPeerActorSelection(followerId);
 385                 if(followerActor != null) {
 386                     sendSnapshotChunk(followerActor, followerId);
 387                 }
 388             }
 389
 390         } else {
 391             LOG.error("{}: Chunk index {} in InstallSnapshotReply from follower {} does not match expected index {}",
 392                     context.getId(), reply.getChunkIndex(), followerId,
 393                     followerToSnapshot.getChunkIndex());
 394
 395             if(reply.getChunkIndex() == INVALID_CHUNK_INDEX){
 396                 // Since the Follower did not find this index to be valid we should reset the follower snapshot
 397                 // so that Installing the snapshot can resume from the beginning
 398                 followerToSnapshot.reset();
 399             }
 400         }
 401     }
 402
 403     private void replicate(Replicate replicate) {
 404         long logIndex = replicate.getReplicatedLogEntry().getIndex();
 405
 406         if(LOG.isDebugEnabled()) {
 407             LOG.debug("{}: Replicate message {}", context.getId(), logIndex);
 408         }
 409
 410         // Create a tracker entry we will use this later to notify the
 411         // client actor
 412         trackerList.add(
 413             new ClientRequestTrackerImpl(replicate.getClientActor(),
 414                 replicate.getIdentifier(),
 415                 logIndex)
 416         );
 417
 418         if (followerToLog.isEmpty()) {
 419             context.setCommitIndex(logIndex);
 420             applyLogToStateMachine(logIndex);
 421         } else {
 422             sendAppendEntries(0);
 423         }
 424     }
 425
 426     private void sendAppendEntries(long timeSinceLastActivityInterval) {
 427         // Send an AppendEntries to all followers
 428         for (Entry<String, FollowerLogInformation> e : followerToLog.entrySet()) {
 429             final String followerId = e.getKey();
 430             final FollowerLogInformation followerLogInformation = e.getValue();
 431             // This checks helps not to send a repeat message to the follower
 432             if(!followerLogInformation.isFollowerActive() ||
 433                     followerLogInformation.timeSinceLastActivity() >= timeSinceLastActivityInterval) {
 434                 sendUpdatesToFollower(followerId, followerLogInformation, true);
 435             }
 436         }
 437     }
 438
 439     /**
 440      *
 441      * This method checks if any update needs to be sent to the given follower. This includes append log entries,
 442      * sending next snapshot chunk, and initiating a snapshot.
 443      * @return true if any update is sent, false otherwise
 444      */
 445
 446     private void sendUpdatesToFollower(String followerId, FollowerLogInformation followerLogInformation,
 447                                           boolean sendHeartbeat) {
 448
 449         ActorSelection followerActor = context.getPeerActorSelection(followerId);
 450         if (followerActor != null) {
 451             long followerNextIndex = followerLogInformation.getNextIndex();
 452             boolean isFollowerActive = followerLogInformation.isFollowerActive();
 453
 454             if (mapFollowerToSnapshot.get(followerId) != null) {
 455                 // if install snapshot is in process , then sent next chunk if possible
 456                 if (isFollowerActive && mapFollowerToSnapshot.get(followerId).canSendNextChunk()) {
 457                     sendSnapshotChunk(followerActor, followerId);
 458                 } else if(sendHeartbeat) {
 459                     // we send a heartbeat even if we have not received a reply for the last chunk
 460                     sendAppendEntriesToFollower(followerActor, followerLogInformation.getNextIndex(),
 461                         Collections.<ReplicatedLogEntry>emptyList(), followerId);
 462                 }
 463             } else {
 464                 long leaderLastIndex = context.getReplicatedLog().lastIndex();
 465                 long leaderSnapShotIndex = context.getReplicatedLog().getSnapshotIndex();
 466                 if (isFollowerActive &&
 467                     context.getReplicatedLog().isPresent(followerNextIndex)) {
 468                     // FIXME : Sending one entry at a time
 469                     final List<ReplicatedLogEntry> entries = context.getReplicatedLog().getFrom(followerNextIndex, 1);
 470
 471                     sendAppendEntriesToFollower(followerActor, followerNextIndex, entries, followerId);
 472
 473                 } else if (isFollowerActive && followerNextIndex >= 0 &&
 474                     leaderLastIndex > followerNextIndex && !context.isSnapshotCaptureInitiated()) {
 475                     // if the followers next index is not present in the leaders log, and
 476                     // if the follower is just not starting and if leader's index is more than followers index
 477                     // then snapshot should be sent
 478
 479                     if (LOG.isDebugEnabled()) {
 480                         LOG.debug("InitiateInstallSnapshot to follower:{}," +
 481                                 "follower-nextIndex:{}, leader-snapshot-index:{},  " +
 482                                 "leader-last-index:{}", followerId,
 483                             followerNextIndex, leaderSnapShotIndex, leaderLastIndex
 484                         );
 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()) {
 512             LOG.debug("{}: Sending AppendEntries to follower {}: {}", context.getId(), 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:{}", context.getId(), 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         for (Entry<String, FollowerLogInformation> e : followerToLog.entrySet()) {
 574             ActorSelection followerActor = context.getPeerActorSelection(e.getKey());
 575
 576             if (followerActor != null) {
 577                 long nextIndex = e.getValue().getNextIndex();
 578
 579                 if (!context.getReplicatedLog().isPresent(nextIndex) &&
 580                     context.getReplicatedLog().isInSnapshot(nextIndex)) {
 581                     sendSnapshotChunk(followerActor, e.getKey());
 582                 }
 583             }
 584         }
 585     }
 586
 587     /**
 588      *  Sends a snapshot chunk to a given follower
 589      *  InstallSnapshot should qualify as a heartbeat too.
 590      */
 591     private void sendSnapshotChunk(ActorSelection followerActor, String followerId) {
 592         try {
 593             if (snapshot.isPresent()) {
 594                 ByteString nextSnapshotChunk = getNextSnapshotChunk(followerId,snapshot.get());
 595
 596                 // Note: the previous call to getNextSnapshotChunk has the side-effect of adding
 597                 // followerId to the followerToSnapshot map.
 598                 FollowerToSnapshot followerToSnapshot = mapFollowerToSnapshot.get(followerId);
 599
 600                 followerActor.tell(
 601                     new InstallSnapshot(currentTerm(), context.getId(),
 602                         context.getReplicatedLog().getSnapshotIndex(),
 603                         context.getReplicatedLog().getSnapshotTerm(),
 604                         nextSnapshotChunk,
 605                         followerToSnapshot.incrementChunkIndex(),
 606                         followerToSnapshot.getTotalChunks(),
 607                         Optional.of(followerToSnapshot.getLastChunkHashCode())
 608                     ).toSerializable(),
 609                     actor()
 610                 );
 611                 LOG.info("{}: InstallSnapshot sent to follower {}, Chunk: {}/{}",
 612                         context.getId(), followerActor.path(),
 613                         followerToSnapshot.getChunkIndex(),
 614                         followerToSnapshot.getTotalChunks());
 615             }
 616         } catch (IOException e) {
 617             LOG.error("{}: InstallSnapshot failed for Leader.", context.getId(), e);
 618         }
 619     }
 620
 621     /**
 622      * Acccepts snaphot as ByteString, enters into map for future chunks
 623      * creates and return a ByteString chunk
 624      */
 625     private ByteString getNextSnapshotChunk(String followerId, ByteString snapshotBytes) throws IOException {
 626         FollowerToSnapshot followerToSnapshot = mapFollowerToSnapshot.get(followerId);
 627         if (followerToSnapshot == null) {
 628             followerToSnapshot = new FollowerToSnapshot(snapshotBytes);
 629             mapFollowerToSnapshot.put(followerId, followerToSnapshot);
 630         }
 631         ByteString nextChunk = followerToSnapshot.getNextChunk();
 632         if (LOG.isDebugEnabled()) {
 633             LOG.debug("{}: Leader's snapshot nextChunk size:{}", context.getId(), nextChunk.size());
 634         }
 635         return nextChunk;
 636     }
 637
 638     private void sendHeartBeat() {
 639         if (!followerToLog.isEmpty()) {
 640             sendAppendEntries(context.getConfigParams().getHeartBeatInterval().toMillis());
 641         }
 642     }
 643
 644     private void stopHeartBeat() {
 645         if (heartbeatSchedule != null && !heartbeatSchedule.isCancelled()) {
 646             heartbeatSchedule.cancel();
 647         }
 648     }
 649
 650     private void scheduleHeartBeat(FiniteDuration interval) {
 651         if (followerToLog.isEmpty()) {
 652             // Optimization - do not bother scheduling a heartbeat as there are
 653             // no followers
 654             return;
 655         }
 656
 657         stopHeartBeat();
 658
 659         // Schedule a heartbeat. When the scheduler triggers a SendHeartbeat
 660         // message is sent to itself.
 661         // Scheduling the heartbeat only once here because heartbeats do not
 662         // need to be sent if there are other messages being sent to the remote
 663         // actor.
 664         heartbeatSchedule = context.getActorSystem().scheduler().scheduleOnce(
 665             interval, context.getActor(), new SendHeartBeat(),
 666             context.getActorSystem().dispatcher(), context.getActor());
 667     }
 668
 669     @Override
 670     public void close() throws Exception {
 671         stopHeartBeat();
 672     }
 673
 674     @Override
 675     public String getLeaderId() {
 676         return context.getId();
 677     }
 678
 679     protected boolean isLeaderIsolated() {
 680         int minPresent = minIsolatedLeaderPeerCount;
 681         for (FollowerLogInformation followerLogInformation : followerToLog.values()) {
 682             if (followerLogInformation.isFollowerActive()) {
 683                 --minPresent;
 684                 if (minPresent == 0) {
 685                     break;
 686                 }
 687             }
 688         }
 689         return (minPresent != 0);
 690     }
 691
 692     /**
 693      * Encapsulates the snapshot bytestring and handles the logic of sending
 694      * snapshot chunks
 695      */
 696     protected class FollowerToSnapshot {
 697         private final ByteString snapshotBytes;
 698         private int offset = 0;
 699         // the next snapshot chunk is sent only if the replyReceivedForOffset matches offset
 700         private int replyReceivedForOffset;
 701         // if replyStatus is false, the previous chunk is attempted
 702         private boolean replyStatus = false;
 703         private int chunkIndex;
 704         private final int totalChunks;
 705         private int lastChunkHashCode = AbstractLeader.INITIAL_LAST_CHUNK_HASH_CODE;
 706         private int nextChunkHashCode = AbstractLeader.INITIAL_LAST_CHUNK_HASH_CODE;
 707
 708         public FollowerToSnapshot(ByteString snapshotBytes) {
 709             this.snapshotBytes = snapshotBytes;
 710             int size = snapshotBytes.size();
 711             totalChunks = ( size / context.getConfigParams().getSnapshotChunkSize()) +
 712                 ((size % context.getConfigParams().getSnapshotChunkSize()) > 0 ? 1 : 0);
 713             if(LOG.isDebugEnabled()) {
 714                 LOG.debug("{}: Snapshot {} bytes, total chunks to send:{}",
 715                         context.getId(), size, totalChunks);
 716             }
 717             replyReceivedForOffset = -1;
 718             chunkIndex = AbstractLeader.FIRST_CHUNK_INDEX;
 719         }
 720
 721         public ByteString getSnapshotBytes() {
 722             return snapshotBytes;
 723         }
 724
 725         public int incrementOffset() {
 726             if(replyStatus) {
 727                 // if prev chunk failed, we would want to sent the same chunk again
 728                 offset = offset + context.getConfigParams().getSnapshotChunkSize();
 729             }
 730             return offset;
 731         }
 732
 733         public int incrementChunkIndex() {
 734             if (replyStatus) {
 735                 // if prev chunk failed, we would want to sent the same chunk again
 736                 chunkIndex =  chunkIndex + 1;
 737             }
 738             return chunkIndex;
 739         }
 740
 741         public int getChunkIndex() {
 742             return chunkIndex;
 743         }
 744
 745         public int getTotalChunks() {
 746             return totalChunks;
 747         }
 748
 749         public boolean canSendNextChunk() {
 750             // we only send a false if a chunk is sent but we have not received a reply yet
 751             return replyReceivedForOffset == offset;
 752         }
 753
 754         public boolean isLastChunk(int chunkIndex) {
 755             return totalChunks == chunkIndex;
 756         }
 757
 758         public void markSendStatus(boolean success) {
 759             if (success) {
 760                 // if the chunk sent was successful
 761                 replyReceivedForOffset = offset;
 762                 replyStatus = true;
 763                 lastChunkHashCode = nextChunkHashCode;
 764             } else {
 765                 // if the chunk sent was failure
 766                 replyReceivedForOffset = offset;
 767                 replyStatus = false;
 768             }
 769         }
 770
 771         public ByteString getNextChunk() {
 772             int snapshotLength = getSnapshotBytes().size();
 773             int start = incrementOffset();
 774             int size = context.getConfigParams().getSnapshotChunkSize();
 775             if (context.getConfigParams().getSnapshotChunkSize() > snapshotLength) {
 776                 size = snapshotLength;
 777             } else {
 778                 if ((start + context.getConfigParams().getSnapshotChunkSize()) > snapshotLength) {
 779                     size = snapshotLength - start;
 780                 }
 781             }
 782
 783             if(LOG.isDebugEnabled()) {
 784                 LOG.debug("{}: Next chunk: length={}, offset={},size={}", context.getId(),
 785                     snapshotLength, start, size);
 786             }
 787             ByteString substring = getSnapshotBytes().substring(start, start + size);
 788             nextChunkHashCode = substring.hashCode();
 789             return substring;
 790         }
 791
 792         /**
 793          * reset should be called when the Follower needs to be sent the snapshot from the beginning
 794          */
 795         public void reset(){
 796             offset = 0;
 797             replyStatus = false;
 798             replyReceivedForOffset = offset;
 799             chunkIndex = AbstractLeader.FIRST_CHUNK_INDEX;
 800             lastChunkHashCode = AbstractLeader.INITIAL_LAST_CHUNK_HASH_CODE;
 801         }
 802
 803         public int getLastChunkHashCode() {
 804             return lastChunkHashCode;
 805         }
 806     }
 807
 808     // called from example-actor for printing the follower-states
 809     public String printFollowerStates() {
 810         final StringBuilder sb = new StringBuilder();
 811
 812         sb.append('[');
 813         for (FollowerLogInformation followerLogInformation : followerToLog.values()) {
 814             sb.append('{');
 815             sb.append(followerLogInformation.getId());
 816             sb.append(" state:");
 817             sb.append(followerLogInformation.isFollowerActive());
 818             sb.append("},");
 819         }
 820         sb.append(']');
 821
 822         return sb.toString();
 823     }
 824
 825     @VisibleForTesting
 826     public FollowerLogInformation getFollower(String followerId) {
 827         return followerToLog.get(followerId);
 828     }
 829
 830     @VisibleForTesting
 831     protected void setFollowerSnapshot(String followerId, FollowerToSnapshot snapshot) {
 832         mapFollowerToSnapshot.put(followerId, snapshot);
 833     }
 834
 835     @VisibleForTesting
 836     public int followerSnapshotSize() {
 837         return mapFollowerToSnapshot.size();
 838     }
 839
 840     @VisibleForTesting
 841     public int followerLogSize() {
 842         return followerToLog.size();
 843     }
 844 }