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, -1, context);
 101
 102             ftlBuilder.put(followerId, followerLogInformation);
 103         }
 104         followerToLog = ftlBuilder.build();
 105
 106         leaderId = context.getId();
 107
 108         LOG.debug("{}: Election: Leader has following peers: {}", logName(), getFollowerIds());
 109
 110         minReplicationCount = getMajorityVoteCount(getFollowerIds().size());
 111
 112         // the isolated Leader peer count will be 1 less than the majority vote count.
 113         // this is because the vote count has the self vote counted in it
 114         // for e.g
 115         // 0 peers = 1 votesRequired , minIsolatedLeaderPeerCount = 0
 116         // 2 peers = 2 votesRequired , minIsolatedLeaderPeerCount = 1
 117         // 4 peers = 3 votesRequired, minIsolatedLeaderPeerCount = 2
 118         minIsolatedLeaderPeerCount = minReplicationCount > 0 ? (minReplicationCount - 1) : 0;
 119
 120         snapshot = Optional.absent();
 121
 122         // Immediately schedule a heartbeat
 123         // Upon election: send initial empty AppendEntries RPCs
 124         // (heartbeat) to each server; repeat during idle periods to
 125         // prevent election timeouts (§5.2)
 126         sendAppendEntries(0, false);
 127
 128         // It is important to schedule this heartbeat here
 129         scheduleHeartBeat(context.getConfigParams().getHeartBeatInterval());
 130     }
 131
 132     /**
 133      * Return an immutable collection of follower identifiers.
 134      *
 135      * @return Collection of follower IDs
 136      */
 137     protected final Collection<String> getFollowerIds() {
 138         return followerToLog.keySet();
 139     }
 140
 141     @VisibleForTesting
 142     void setSnapshot(Optional<ByteString> snapshot) {
 143         this.snapshot = snapshot;
 144     }
 145
 146     @Override
 147     protected RaftActorBehavior handleAppendEntries(ActorRef sender,
 148         AppendEntries appendEntries) {
 149
 150         LOG.debug("{}: handleAppendEntries: {}", logName(), appendEntries);
 151
 152         return this;
 153     }
 154
 155     @Override
 156     protected RaftActorBehavior handleAppendEntriesReply(ActorRef sender,
 157         AppendEntriesReply appendEntriesReply) {
 158
 159         if(LOG.isTraceEnabled()) {
 160             LOG.trace("{}: handleAppendEntriesReply: {}", logName(), appendEntriesReply);
 161         } else if(LOG.isDebugEnabled() && !appendEntriesReply.isSuccess()) {
 162             LOG.debug("{}: handleAppendEntriesReply: {}", logName(), appendEntriesReply);
 163         }
 164
 165         // Update the FollowerLogInformation
 166         String followerId = appendEntriesReply.getFollowerId();
 167         FollowerLogInformation followerLogInformation =
 168             followerToLog.get(followerId);
 169
 170         if(followerLogInformation == null){
 171             LOG.error("{}: handleAppendEntriesReply - unknown follower {}", logName(), followerId);
 172             return this;
 173         }
 174
 175         if(followerLogInformation.timeSinceLastActivity() >
 176                 context.getConfigParams().getElectionTimeOutInterval().toMillis()) {
 177             LOG.error("{} : handleAppendEntriesReply delayed beyond election timeout, " +
 178                             "appendEntriesReply : {}, timeSinceLastActivity : {}, lastApplied : {}, commitIndex : {}",
 179                     logName(), appendEntriesReply, followerLogInformation.timeSinceLastActivity(),
 180                     context.getLastApplied(), context.getCommitIndex());
 181         }
 182
 183         followerLogInformation.markFollowerActive();
 184
 185         if (appendEntriesReply.isSuccess()) {
 186             followerLogInformation
 187                 .setMatchIndex(appendEntriesReply.getLogLastIndex());
 188             followerLogInformation
 189                 .setNextIndex(appendEntriesReply.getLogLastIndex() + 1);
 190         } else {
 191
 192             // TODO: When we find that the follower is out of sync with the
 193             // Leader we simply decrement that followers next index by 1.
 194             // Would it be possible to do better than this? The RAFT spec
 195             // does not explicitly deal with it but may be something for us to
 196             // think about
 197
 198             followerLogInformation.decrNextIndex();
 199         }
 200
 201         // Now figure out if this reply warrants a change in the commitIndex
 202         // If there exists an N such that N > commitIndex, a majority
 203         // of matchIndex[i] ≥ N, and log[N].term == currentTerm:
 204         // set commitIndex = N (§5.3, §5.4).
 205         for (long N = context.getCommitIndex() + 1; ; N++) {
 206             int replicatedCount = 1;
 207
 208             for (FollowerLogInformation info : followerToLog.values()) {
 209                 if (info.getMatchIndex() >= N) {
 210                     replicatedCount++;
 211                 }
 212             }
 213
 214             if (replicatedCount >= minReplicationCount) {
 215                 ReplicatedLogEntry replicatedLogEntry = context.getReplicatedLog().get(N);
 216                 if (replicatedLogEntry != null &&
 217                     replicatedLogEntry.getTerm() == currentTerm()) {
 218                     context.setCommitIndex(N);
 219                 }
 220             } else {
 221                 break;
 222             }
 223         }
 224
 225         // Apply the change to the state machine
 226         if (context.getCommitIndex() > context.getLastApplied()) {
 227             LOG.debug("{}: handleAppendEntriesReply: applying to log - commitIndex: {}, lastAppliedIndex: {}",
 228                     logName(), context.getCommitIndex(), context.getLastApplied());
 229
 230             applyLogToStateMachine(context.getCommitIndex());
 231         }
 232
 233         if (!context.isSnapshotCaptureInitiated()) {
 234             purgeInMemoryLog();
 235         }
 236
 237         //Send the next log entry immediately, if possible, no need to wait for heartbeat to trigger that event
 238         sendUpdatesToFollower(followerId, followerLogInformation, false, false);
 239         return this;
 240     }
 241
 242     private void purgeInMemoryLog() {
 243         //find the lowest index across followers which has been replicated to all.
 244         // lastApplied if there are no followers, so that we keep clearing the log for single-node
 245         // we would delete the in-mem log from that index on, in-order to minimize mem usage
 246         // we would also share this info thru AE with the followers so that they can delete their log entries as well.
 247         long minReplicatedToAllIndex = followerToLog.isEmpty() ? context.getLastApplied() : Long.MAX_VALUE;
 248         for (FollowerLogInformation info : followerToLog.values()) {
 249             minReplicatedToAllIndex = Math.min(minReplicatedToAllIndex, info.getMatchIndex());
 250         }
 251
 252         super.performSnapshotWithoutCapture(minReplicatedToAllIndex);
 253     }
 254
 255     @Override
 256     protected ClientRequestTracker removeClientRequestTracker(long logIndex) {
 257         final Iterator<ClientRequestTracker> it = trackerList.iterator();
 258         while (it.hasNext()) {
 259             final ClientRequestTracker t = it.next();
 260             if (t.getIndex() == logIndex) {
 261                 it.remove();
 262                 return t;
 263             }
 264         }
 265
 266         return null;
 267     }
 268
 269     @Override
 270     protected ClientRequestTracker findClientRequestTracker(long logIndex) {
 271         for (ClientRequestTracker tracker : trackerList) {
 272             if (tracker.getIndex() == logIndex) {
 273                 return tracker;
 274             }
 275         }
 276         return null;
 277     }
 278
 279     @Override
 280     protected RaftActorBehavior handleRequestVoteReply(ActorRef sender,
 281         RequestVoteReply requestVoteReply) {
 282         return this;
 283     }
 284
 285     protected void beforeSendHeartbeat(){}
 286
 287     @Override
 288     public RaftActorBehavior handleMessage(ActorRef sender, Object originalMessage) {
 289         Preconditions.checkNotNull(sender, "sender should not be null");
 290
 291         Object message = fromSerializableMessage(originalMessage);
 292
 293         if (message instanceof RaftRPC) {
 294             RaftRPC rpc = (RaftRPC) message;
 295             // If RPC request or response contains term T > currentTerm:
 296             // set currentTerm = T, convert to follower (§5.1)
 297             // This applies to all RPC messages and responses
 298             if (rpc.getTerm() > context.getTermInformation().getCurrentTerm()) {
 299                 LOG.debug("{}: Term {} in \"{}\" message is greater than leader's term {} - switching to Follower",
 300                         logName(), rpc.getTerm(), rpc, context.getTermInformation().getCurrentTerm());
 301
 302                 context.getTermInformation().updateAndPersist(rpc.getTerm(), null);
 303
 304                 return switchBehavior(new Follower(context));
 305             }
 306         }
 307
 308         if (message instanceof SendHeartBeat) {
 309             beforeSendHeartbeat();
 310             sendHeartBeat();
 311             scheduleHeartBeat(context.getConfigParams().getHeartBeatInterval());
 312             return this;
 313
 314         } else if(message instanceof SendInstallSnapshot) {
 315             // received from RaftActor
 316             setSnapshot(Optional.of(((SendInstallSnapshot) message).getSnapshot()));
 317             sendInstallSnapshot();
 318
 319         } else if (message instanceof Replicate) {
 320             replicate((Replicate) message);
 321
 322         } else if (message instanceof InstallSnapshotReply){
 323             handleInstallSnapshotReply((InstallSnapshotReply) message);
 324
 325         }
 326
 327
 328         return super.handleMessage(sender, message);
 329     }
 330
 331     private void handleInstallSnapshotReply(InstallSnapshotReply reply) {
 332         LOG.debug("{}: handleInstallSnapshotReply: {}", logName(), reply);
 333
 334         String followerId = reply.getFollowerId();
 335         FollowerToSnapshot followerToSnapshot = mapFollowerToSnapshot.get(followerId);
 336
 337         if (followerToSnapshot == null) {
 338             LOG.error("{}: FollowerId {} in InstallSnapshotReply not known to Leader",
 339                     logName(), followerId);
 340             return;
 341         }
 342
 343         FollowerLogInformation followerLogInformation = followerToLog.get(followerId);
 344         followerLogInformation.markFollowerActive();
 345
 346         if (followerToSnapshot.getChunkIndex() == reply.getChunkIndex()) {
 347             boolean wasLastChunk = false;
 348             if (reply.isSuccess()) {
 349                 if(followerToSnapshot.isLastChunk(reply.getChunkIndex())) {
 350                     //this was the last chunk reply
 351                     if(LOG.isDebugEnabled()) {
 352                         LOG.debug("{}: InstallSnapshotReply received, " +
 353                                 "last chunk received, Chunk: {}. Follower: {} Setting nextIndex: {}",
 354                                 logName(), reply.getChunkIndex(), followerId,
 355                             context.getReplicatedLog().getSnapshotIndex() + 1
 356                         );
 357                     }
 358
 359                     followerLogInformation.setMatchIndex(
 360                         context.getReplicatedLog().getSnapshotIndex());
 361                     followerLogInformation.setNextIndex(
 362                         context.getReplicatedLog().getSnapshotIndex() + 1);
 363                     mapFollowerToSnapshot.remove(followerId);
 364
 365                     LOG.debug("{}: follower: {}, matchIndex set to {}, nextIndex set to {}",
 366                                 logName(), followerId, followerLogInformation.getMatchIndex(),
 367                                 followerLogInformation.getNextIndex());
 368
 369                     if (mapFollowerToSnapshot.isEmpty()) {
 370                         // once there are no pending followers receiving snapshots
 371                         // we can remove snapshot from the memory
 372                         setSnapshot(Optional.<ByteString>absent());
 373                     }
 374                     wasLastChunk = true;
 375
 376                 } else {
 377                     followerToSnapshot.markSendStatus(true);
 378                 }
 379             } else {
 380                 LOG.info("{}: InstallSnapshotReply received sending snapshot chunk failed, Will retry, Chunk: {}",
 381                         logName(), reply.getChunkIndex());
 382
 383                 followerToSnapshot.markSendStatus(false);
 384             }
 385
 386             if (!wasLastChunk && followerToSnapshot.canSendNextChunk()) {
 387                 ActorSelection followerActor = context.getPeerActorSelection(followerId);
 388                 if(followerActor != null) {
 389                     sendSnapshotChunk(followerActor, followerId);
 390                 }
 391             }
 392
 393         } else {
 394             LOG.error("{}: Chunk index {} in InstallSnapshotReply from follower {} does not match expected index {}",
 395                     logName(), reply.getChunkIndex(), followerId,
 396                     followerToSnapshot.getChunkIndex());
 397
 398             if(reply.getChunkIndex() == INVALID_CHUNK_INDEX){
 399                 // Since the Follower did not find this index to be valid we should reset the follower snapshot
 400                 // so that Installing the snapshot can resume from the beginning
 401                 followerToSnapshot.reset();
 402             }
 403         }
 404     }
 405
 406     private void replicate(Replicate replicate) {
 407         long logIndex = replicate.getReplicatedLogEntry().getIndex();
 408
 409         LOG.debug("{}: Replicate message: identifier: {}, logIndex: {}", logName(),
 410                 replicate.getIdentifier(), logIndex);
 411
 412         // Create a tracker entry we will use this later to notify the
 413         // client actor
 414         trackerList.add(
 415             new ClientRequestTrackerImpl(replicate.getClientActor(),
 416                 replicate.getIdentifier(),
 417                 logIndex)
 418         );
 419
 420         if (followerToLog.isEmpty()) {
 421             context.setCommitIndex(logIndex);
 422             applyLogToStateMachine(logIndex);
 423         } else {
 424             sendAppendEntries(0, false);
 425         }
 426     }
 427
 428     private void sendAppendEntries(long timeSinceLastActivityInterval, boolean isHeartbeat) {
 429         // Send an AppendEntries to all followers
 430         for (Entry<String, FollowerLogInformation> e : followerToLog.entrySet()) {
 431             final String followerId = e.getKey();
 432             final FollowerLogInformation followerLogInformation = e.getValue();
 433             // This checks helps not to send a repeat message to the follower
 434             if(!followerLogInformation.isFollowerActive() ||
 435                     followerLogInformation.timeSinceLastActivity() >= timeSinceLastActivityInterval) {
 436                 sendUpdatesToFollower(followerId, followerLogInformation, true, isHeartbeat);
 437             }
 438         }
 439     }
 440
 441     /**
 442      *
 443      * This method checks if any update needs to be sent to the given follower. This includes append log entries,
 444      * sending next snapshot chunk, and initiating a snapshot.
 445      * @return true if any update is sent, false otherwise
 446      */
 447
 448     private void sendUpdatesToFollower(String followerId, FollowerLogInformation followerLogInformation,
 449                                        boolean sendHeartbeat, boolean isHeartbeat) {
 450
 451         ActorSelection followerActor = context.getPeerActorSelection(followerId);
 452         if (followerActor != null) {
 453             long followerNextIndex = followerLogInformation.getNextIndex();
 454             boolean isFollowerActive = followerLogInformation.isFollowerActive();
 455
 456             if (mapFollowerToSnapshot.get(followerId) != null) {
 457                 // if install snapshot is in process , then sent next chunk if possible
 458                 if (isFollowerActive && mapFollowerToSnapshot.get(followerId).canSendNextChunk()) {
 459                     sendSnapshotChunk(followerActor, followerId);
 460                 } else if(sendHeartbeat) {
 461                     // we send a heartbeat even if we have not received a reply for the last chunk
 462                     sendAppendEntriesToFollower(followerActor, followerLogInformation.getNextIndex(),
 463                         Collections.<ReplicatedLogEntry>emptyList(), followerId);
 464                 }
 465             } else {
 466                 long leaderLastIndex = context.getReplicatedLog().lastIndex();
 467                 long leaderSnapShotIndex = context.getReplicatedLog().getSnapshotIndex();
 468
 469                 if(!isHeartbeat || LOG.isTraceEnabled()) {
 470                     LOG.debug("{}: Checking sendAppendEntries for follower {}, leaderLastIndex: {}, leaderSnapShotIndex: {}",
 471                             logName(), followerId, leaderLastIndex, leaderSnapShotIndex);
 472                 }
 473
 474                 if (isFollowerActive && context.getReplicatedLog().isPresent(followerNextIndex)) {
 475
 476                     LOG.debug("{}: sendAppendEntries: {} is present for follower {}", logName(),
 477                             followerNextIndex, followerId);
 478
 479                     // FIXME : Sending one entry at a time
 480                     final List<ReplicatedLogEntry> entries = context.getReplicatedLog().getFrom(followerNextIndex, 1);
 481
 482                     sendAppendEntriesToFollower(followerActor, followerNextIndex, entries, followerId);
 483
 484                 } else if (isFollowerActive && followerNextIndex >= 0 &&
 485                     leaderLastIndex > followerNextIndex && !context.isSnapshotCaptureInitiated()) {
 486                     // if the followers next index is not present in the leaders log, and
 487                     // if the follower is just not starting and if leader's index is more than followers index
 488                     // then snapshot should be sent
 489
 490                     if (LOG.isDebugEnabled()) {
 491                         LOG.debug(String.format("%s: InitiateInstallSnapshot to follower: %s," +
 492                                     "follower-nextIndex: %d, leader-snapshot-index: %d,  " +
 493                                     "leader-last-index: %d", logName(), followerId,
 494                                     followerNextIndex, leaderSnapShotIndex, leaderLastIndex));
 495                     }
 496
 497                     // Send heartbeat to follower whenever install snapshot is initiated.
 498                     sendAppendEntriesToFollower(followerActor, followerLogInformation.getNextIndex(),
 499                             Collections.<ReplicatedLogEntry>emptyList(), followerId);
 500
 501                     initiateCaptureSnapshot(followerId, followerNextIndex);
 502
 503                 } else if(sendHeartbeat) {
 504                     //we send an AppendEntries, even if the follower is inactive
 505                     // in-order to update the followers timestamp, in case it becomes active again
 506                     sendAppendEntriesToFollower(followerActor, followerLogInformation.getNextIndex(),
 507                         Collections.<ReplicatedLogEntry>emptyList(), followerId);
 508                 }
 509
 510             }
 511         }
 512     }
 513
 514     private void sendAppendEntriesToFollower(ActorSelection followerActor, long followerNextIndex,
 515         List<ReplicatedLogEntry> entries, String followerId) {
 516         AppendEntries appendEntries = new AppendEntries(currentTerm(), context.getId(),
 517             prevLogIndex(followerNextIndex),
 518             prevLogTerm(followerNextIndex), entries,
 519             context.getCommitIndex(), super.getReplicatedToAllIndex());
 520
 521         if(!entries.isEmpty() || LOG.isTraceEnabled()) {
 522             LOG.debug("{}: Sending AppendEntries to follower {}: {}", logName(), followerId,
 523                     appendEntries);
 524         }
 525
 526         followerActor.tell(appendEntries.toSerializable(), actor());
 527     }
 528
 529     /**
 530      * Install Snapshot works as follows
 531      * 1. Leader initiates the capture snapshot by sending a CaptureSnapshot message to actor
 532      * 2. RaftActor on receipt of the CaptureSnapshotReply (from Shard), stores the received snapshot in the replicated log
 533      * and makes a call to Leader's handleMessage , with SendInstallSnapshot message.
 534      * 3. Leader , picks the snapshot from im-mem ReplicatedLog and sends it in chunks to the Follower
 535      * 4. On complete, Follower sends back a InstallSnapshotReply.
 536      * 5. On receipt of the InstallSnapshotReply for the last chunk, Leader marks the install complete for that follower
 537      * and replenishes the memory by deleting the snapshot in Replicated log.
 538      * 6. If another follower requires a snapshot and a snapshot has been collected (via CaptureSnapshotReply)
 539      * then send the existing snapshot in chunks to the follower.
 540      * @param followerId
 541      * @param followerNextIndex
 542      */
 543     private void initiateCaptureSnapshot(String followerId, long followerNextIndex) {
 544         if (!context.getReplicatedLog().isPresent(followerNextIndex) &&
 545                 context.getReplicatedLog().isInSnapshot(followerNextIndex)) {
 546
 547             if (snapshot.isPresent()) {
 548                 // if a snapshot is present in the memory, most likely another install is in progress
 549                 // no need to capture snapshot.
 550                 // This could happen if another follower needs an install when one is going on.
 551                 final ActorSelection followerActor = context.getPeerActorSelection(followerId);
 552                 sendSnapshotChunk(followerActor, followerId);
 553
 554             } else if (!context.isSnapshotCaptureInitiated()) {
 555
 556                 LOG.info("{}: Initiating Snapshot Capture to Install Snapshot, Leader:{}", logName(), getLeaderId());
 557                 ReplicatedLogEntry lastAppliedEntry = context.getReplicatedLog().get(context.getLastApplied());
 558                 long lastAppliedIndex = -1;
 559                 long lastAppliedTerm = -1;
 560
 561                 if (lastAppliedEntry != null) {
 562                     lastAppliedIndex = lastAppliedEntry.getIndex();
 563                     lastAppliedTerm = lastAppliedEntry.getTerm();
 564                 } else if (context.getReplicatedLog().getSnapshotIndex() > -1) {
 565                     lastAppliedIndex = context.getReplicatedLog().getSnapshotIndex();
 566                     lastAppliedTerm = context.getReplicatedLog().getSnapshotTerm();
 567                 }
 568
 569                 boolean isInstallSnapshotInitiated = true;
 570                 long replicatedToAllIndex = super.getReplicatedToAllIndex();
 571                 ReplicatedLogEntry replicatedToAllEntry = context.getReplicatedLog().get(replicatedToAllIndex);
 572                 actor().tell(new CaptureSnapshot(lastIndex(), lastTerm(), lastAppliedIndex, lastAppliedTerm,
 573                     (replicatedToAllEntry != null ? replicatedToAllEntry.getIndex() : -1),
 574                     (replicatedToAllEntry != null ? replicatedToAllEntry.getTerm() : -1),
 575                     isInstallSnapshotInitiated), actor());
 576                 context.setSnapshotCaptureInitiated(true);
 577             }
 578         }
 579     }
 580
 581
 582     private void sendInstallSnapshot() {
 583         LOG.debug("{}: sendInstallSnapshot", logName());
 584         for (Entry<String, FollowerLogInformation> e : followerToLog.entrySet()) {
 585             ActorSelection followerActor = context.getPeerActorSelection(e.getKey());
 586
 587             if (followerActor != null) {
 588                 long nextIndex = e.getValue().getNextIndex();
 589
 590                 if (!context.getReplicatedLog().isPresent(nextIndex) &&
 591                     context.getReplicatedLog().isInSnapshot(nextIndex)) {
 592                     sendSnapshotChunk(followerActor, e.getKey());
 593                 }
 594             }
 595         }
 596     }
 597
 598     /**
 599      *  Sends a snapshot chunk to a given follower
 600      *  InstallSnapshot should qualify as a heartbeat too.
 601      */
 602     private void sendSnapshotChunk(ActorSelection followerActor, String followerId) {
 603         try {
 604             if (snapshot.isPresent()) {
 605                 ByteString nextSnapshotChunk = getNextSnapshotChunk(followerId,snapshot.get());
 606
 607                 // Note: the previous call to getNextSnapshotChunk has the side-effect of adding
 608                 // followerId to the followerToSnapshot map.
 609                 FollowerToSnapshot followerToSnapshot = mapFollowerToSnapshot.get(followerId);
 610
 611                 followerActor.tell(
 612                     new InstallSnapshot(currentTerm(), context.getId(),
 613                         context.getReplicatedLog().getSnapshotIndex(),
 614                         context.getReplicatedLog().getSnapshotTerm(),
 615                         nextSnapshotChunk,
 616                         followerToSnapshot.incrementChunkIndex(),
 617                         followerToSnapshot.getTotalChunks(),
 618                         Optional.of(followerToSnapshot.getLastChunkHashCode())
 619                     ).toSerializable(),
 620                     actor()
 621                 );
 622                 LOG.info("{}: InstallSnapshot sent to follower {}, Chunk: {}/{}",
 623                         logName(), followerActor.path(),
 624                         followerToSnapshot.getChunkIndex(),
 625                         followerToSnapshot.getTotalChunks());
 626             }
 627         } catch (IOException e) {
 628             LOG.error("{}: InstallSnapshot failed for Leader.", logName(), e);
 629         }
 630     }
 631
 632     /**
 633      * Acccepts snaphot as ByteString, enters into map for future chunks
 634      * creates and return a ByteString chunk
 635      */
 636     private ByteString getNextSnapshotChunk(String followerId, ByteString snapshotBytes) throws IOException {
 637         FollowerToSnapshot followerToSnapshot = mapFollowerToSnapshot.get(followerId);
 638         if (followerToSnapshot == null) {
 639             followerToSnapshot = new FollowerToSnapshot(snapshotBytes);
 640             mapFollowerToSnapshot.put(followerId, followerToSnapshot);
 641         }
 642         ByteString nextChunk = followerToSnapshot.getNextChunk();
 643
 644         LOG.debug("{}: next snapshot chunk size for follower {}: {}", logName(), followerId, nextChunk.size());
 645
 646         return nextChunk;
 647     }
 648
 649     private void sendHeartBeat() {
 650         if (!followerToLog.isEmpty()) {
 651             LOG.trace("{}: Sending heartbeat", logName());
 652             sendAppendEntries(context.getConfigParams().getHeartBeatInterval().toMillis(), true);
 653         }
 654     }
 655
 656     private void stopHeartBeat() {
 657         if (heartbeatSchedule != null && !heartbeatSchedule.isCancelled()) {
 658             heartbeatSchedule.cancel();
 659         }
 660     }
 661
 662     private void scheduleHeartBeat(FiniteDuration interval) {
 663         if (followerToLog.isEmpty()) {
 664             // Optimization - do not bother scheduling a heartbeat as there are
 665             // no followers
 666             return;
 667         }
 668
 669         stopHeartBeat();
 670
 671         // Schedule a heartbeat. When the scheduler triggers a SendHeartbeat
 672         // message is sent to itself.
 673         // Scheduling the heartbeat only once here because heartbeats do not
 674         // need to be sent if there are other messages being sent to the remote
 675         // actor.
 676         heartbeatSchedule = context.getActorSystem().scheduler().scheduleOnce(
 677             interval, context.getActor(), new SendHeartBeat(),
 678             context.getActorSystem().dispatcher(), context.getActor());
 679     }
 680
 681     @Override
 682     public void close() throws Exception {
 683         stopHeartBeat();
 684     }
 685
 686     @Override
 687     public String getLeaderId() {
 688         return context.getId();
 689     }
 690
 691     protected boolean isLeaderIsolated() {
 692         int minPresent = minIsolatedLeaderPeerCount;
 693         for (FollowerLogInformation followerLogInformation : followerToLog.values()) {
 694             if (followerLogInformation.isFollowerActive()) {
 695                 --minPresent;
 696                 if (minPresent == 0) {
 697                     break;
 698                 }
 699             }
 700         }
 701         return (minPresent != 0);
 702     }
 703
 704     /**
 705      * Encapsulates the snapshot bytestring and handles the logic of sending
 706      * snapshot chunks
 707      */
 708     protected class FollowerToSnapshot {
 709         private final ByteString snapshotBytes;
 710         private int offset = 0;
 711         // the next snapshot chunk is sent only if the replyReceivedForOffset matches offset
 712         private int replyReceivedForOffset;
 713         // if replyStatus is false, the previous chunk is attempted
 714         private boolean replyStatus = false;
 715         private int chunkIndex;
 716         private final int totalChunks;
 717         private int lastChunkHashCode = AbstractLeader.INITIAL_LAST_CHUNK_HASH_CODE;
 718         private int nextChunkHashCode = AbstractLeader.INITIAL_LAST_CHUNK_HASH_CODE;
 719
 720         public FollowerToSnapshot(ByteString snapshotBytes) {
 721             this.snapshotBytes = snapshotBytes;
 722             int size = snapshotBytes.size();
 723             totalChunks = ( size / context.getConfigParams().getSnapshotChunkSize()) +
 724                 ((size % context.getConfigParams().getSnapshotChunkSize()) > 0 ? 1 : 0);
 725             if(LOG.isDebugEnabled()) {
 726                 LOG.debug("{}: Snapshot {} bytes, total chunks to send:{}",
 727                         logName(), size, totalChunks);
 728             }
 729             replyReceivedForOffset = -1;
 730             chunkIndex = AbstractLeader.FIRST_CHUNK_INDEX;
 731         }
 732
 733         public ByteString getSnapshotBytes() {
 734             return snapshotBytes;
 735         }
 736
 737         public int incrementOffset() {
 738             if(replyStatus) {
 739                 // if prev chunk failed, we would want to sent the same chunk again
 740                 offset = offset + context.getConfigParams().getSnapshotChunkSize();
 741             }
 742             return offset;
 743         }
 744
 745         public int incrementChunkIndex() {
 746             if (replyStatus) {
 747                 // if prev chunk failed, we would want to sent the same chunk again
 748                 chunkIndex =  chunkIndex + 1;
 749             }
 750             return chunkIndex;
 751         }
 752
 753         public int getChunkIndex() {
 754             return chunkIndex;
 755         }
 756
 757         public int getTotalChunks() {
 758             return totalChunks;
 759         }
 760
 761         public boolean canSendNextChunk() {
 762             // we only send a false if a chunk is sent but we have not received a reply yet
 763             return replyReceivedForOffset == offset;
 764         }
 765
 766         public boolean isLastChunk(int chunkIndex) {
 767             return totalChunks == chunkIndex;
 768         }
 769
 770         public void markSendStatus(boolean success) {
 771             if (success) {
 772                 // if the chunk sent was successful
 773                 replyReceivedForOffset = offset;
 774                 replyStatus = true;
 775                 lastChunkHashCode = nextChunkHashCode;
 776             } else {
 777                 // if the chunk sent was failure
 778                 replyReceivedForOffset = offset;
 779                 replyStatus = false;
 780             }
 781         }
 782
 783         public ByteString getNextChunk() {
 784             int snapshotLength = getSnapshotBytes().size();
 785             int start = incrementOffset();
 786             int size = context.getConfigParams().getSnapshotChunkSize();
 787             if (context.getConfigParams().getSnapshotChunkSize() > snapshotLength) {
 788                 size = snapshotLength;
 789             } else {
 790                 if ((start + context.getConfigParams().getSnapshotChunkSize()) > snapshotLength) {
 791                     size = snapshotLength - start;
 792                 }
 793             }
 794
 795
 796             LOG.debug("{}: Next chunk: length={}, offset={},size={}", logName(),
 797                     snapshotLength, start, size);
 798
 799             ByteString substring = getSnapshotBytes().substring(start, start + size);
 800             nextChunkHashCode = substring.hashCode();
 801             return substring;
 802         }
 803
 804         /**
 805          * reset should be called when the Follower needs to be sent the snapshot from the beginning
 806          */
 807         public void reset(){
 808             offset = 0;
 809             replyStatus = false;
 810             replyReceivedForOffset = offset;
 811             chunkIndex = AbstractLeader.FIRST_CHUNK_INDEX;
 812             lastChunkHashCode = AbstractLeader.INITIAL_LAST_CHUNK_HASH_CODE;
 813         }
 814
 815         public int getLastChunkHashCode() {
 816             return lastChunkHashCode;
 817         }
 818     }
 819
 820     // called from example-actor for printing the follower-states
 821     public String printFollowerStates() {
 822         final StringBuilder sb = new StringBuilder();
 823
 824         sb.append('[');
 825         for (FollowerLogInformation followerLogInformation : followerToLog.values()) {
 826             sb.append('{');
 827             sb.append(followerLogInformation.getId());
 828             sb.append(" state:");
 829             sb.append(followerLogInformation.isFollowerActive());
 830             sb.append("},");
 831         }
 832         sb.append(']');
 833
 834         return sb.toString();
 835     }
 836
 837     @VisibleForTesting
 838     public FollowerLogInformation getFollower(String followerId) {
 839         return followerToLog.get(followerId);
 840     }
 841
 842     @VisibleForTesting
 843     protected void setFollowerSnapshot(String followerId, FollowerToSnapshot snapshot) {
 844         mapFollowerToSnapshot.put(followerId, snapshot);
 845     }
 846
 847     @VisibleForTesting
 848     public int followerSnapshotSize() {
 849         return mapFollowerToSnapshot.size();
 850     }
 851
 852     @VisibleForTesting
 853     public int followerLogSize() {
 854         return followerToLog.size();
 855     }
 856 }