opendaylight/md-sal/sal-akka-raft/src/main/java/org/opendaylight/controller/cluster/raft/behaviors/Leader.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.base.Preconditions;
  15 import org.opendaylight.controller.cluster.raft.ClientRequestTracker;
  16 import org.opendaylight.controller.cluster.raft.ClientRequestTrackerImpl;
  17 import org.opendaylight.controller.cluster.raft.FollowerLogInformation;
  18 import org.opendaylight.controller.cluster.raft.FollowerLogInformationImpl;
  19 import org.opendaylight.controller.cluster.raft.RaftActorContext;
  20 import org.opendaylight.controller.cluster.raft.RaftState;
  21 import org.opendaylight.controller.cluster.raft.ReplicatedLogEntry;
  22 import org.opendaylight.controller.cluster.raft.internal.messages.ApplyState;
  23 import org.opendaylight.controller.cluster.raft.internal.messages.Replicate;
  24 import org.opendaylight.controller.cluster.raft.internal.messages.SendHeartBeat;
  25 import org.opendaylight.controller.cluster.raft.internal.messages.SendInstallSnapshot;
  26 import org.opendaylight.controller.cluster.raft.messages.AppendEntries;
  27 import org.opendaylight.controller.cluster.raft.messages.AppendEntriesReply;
  28 import org.opendaylight.controller.cluster.raft.messages.InstallSnapshot;
  29 import org.opendaylight.controller.cluster.raft.messages.InstallSnapshotReply;
  30 import org.opendaylight.controller.cluster.raft.messages.RaftRPC;
  31 import org.opendaylight.controller.cluster.raft.messages.RequestVoteReply;
  32 import scala.concurrent.duration.FiniteDuration;
  33
  34 import java.util.ArrayList;
  35 import java.util.Collections;
  36 import java.util.HashMap;
  37 import java.util.List;
  38 import java.util.Map;
  39 import java.util.concurrent.TimeUnit;
  40 import java.util.concurrent.atomic.AtomicLong;
  41
  42 /**
  43  * The behavior of a RaftActor when it is in the Leader state
  44  * <p/>
  45  * Leaders:
  46  * <ul>
  47  * <li> Upon election: send initial empty AppendEntries RPCs
  48  * (heartbeat) to each server; repeat during idle periods to
  49  * prevent election timeouts (§5.2)
  50  * <li> If command received from client: append entry to local log,
  51  * respond after entry applied to state machine (§5.3)
  52  * <li> If last log index ≥ nextIndex for a follower: send
  53  * AppendEntries RPC with log entries starting at nextIndex
  54  * <ul>
  55  * <li> If successful: update nextIndex and matchIndex for
  56  * follower (§5.3)
  57  * <li> If AppendEntries fails because of log inconsistency:
  58  * decrement nextIndex and retry (§5.3)
  59  * </ul>
  60  * <li> If there exists an N such that N > commitIndex, a majority
  61  * of matchIndex[i] ≥ N, and log[N].term == currentTerm:
  62  * set commitIndex = N (§5.3, §5.4).
  63  */
  64 public class Leader extends AbstractRaftActorBehavior {
  65
  66
  67     private final Map<String, FollowerLogInformation> followerToLog =
  68         new HashMap();
  69
  70     private final Map<String, ActorSelection> followerToActor = new HashMap<>();
  71
  72     private Cancellable heartbeatSchedule = null;
  73     private Cancellable appendEntriesSchedule = null;
  74     private Cancellable installSnapshotSchedule = null;
  75
  76     private List<ClientRequestTracker> trackerList = new ArrayList<>();
  77
  78     private final int minReplicationCount;
  79
  80     public Leader(RaftActorContext context) {
  81         super(context);
  82
  83         if (lastIndex() >= 0) {
  84             context.setCommitIndex(lastIndex());
  85         }
  86
  87         for (String followerId : context.getPeerAddresses().keySet()) {
  88             FollowerLogInformation followerLogInformation =
  89                 new FollowerLogInformationImpl(followerId,
  90                     new AtomicLong(lastIndex()),
  91                     new AtomicLong(-1));
  92
  93             followerToActor.put(followerId,
  94                 context.actorSelection(context.getPeerAddress(followerId)));
  95
  96             followerToLog.put(followerId, followerLogInformation);
  97         }
  98
  99         context.getLogger().debug("Election:Leader has following peers:"+followerToActor.keySet());
 100
 101         if (followerToActor.size() > 0) {
 102             minReplicationCount = (followerToActor.size() + 1) / 2 + 1;
 103         } else {
 104             minReplicationCount = 0;
 105         }
 106
 107
 108         // Immediately schedule a heartbeat
 109         // Upon election: send initial empty AppendEntries RPCs
 110         // (heartbeat) to each server; repeat during idle periods to
 111         // prevent election timeouts (§5.2)
 112         scheduleHeartBeat(new FiniteDuration(0, TimeUnit.SECONDS));
 113
 114         scheduleInstallSnapshotCheck(
 115             new FiniteDuration(HEART_BEAT_INTERVAL.length() * 1000,
 116                 HEART_BEAT_INTERVAL.unit())
 117         );
 118
 119     }
 120
 121     @Override protected RaftState handleAppendEntries(ActorRef sender,
 122         AppendEntries appendEntries) {
 123
 124         return state();
 125     }
 126
 127     @Override protected RaftState handleAppendEntriesReply(ActorRef sender,
 128         AppendEntriesReply appendEntriesReply) {
 129
 130         // Update the FollowerLogInformation
 131         String followerId = appendEntriesReply.getFollowerId();
 132         FollowerLogInformation followerLogInformation =
 133             followerToLog.get(followerId);
 134         if (appendEntriesReply.isSuccess()) {
 135             followerLogInformation
 136                 .setMatchIndex(appendEntriesReply.getLogLastIndex());
 137             followerLogInformation
 138                 .setNextIndex(appendEntriesReply.getLogLastIndex() + 1);
 139         } else {
 140             followerLogInformation.decrNextIndex();
 141         }
 142
 143         // Now figure out if this reply warrants a change in the commitIndex
 144         // If there exists an N such that N > commitIndex, a majority
 145         // of matchIndex[i] ≥ N, and log[N].term == currentTerm:
 146         // set commitIndex = N (§5.3, §5.4).
 147         for (long N = context.getCommitIndex() + 1; ; N++) {
 148             int replicatedCount = 1;
 149
 150             for (FollowerLogInformation info : followerToLog.values()) {
 151                 if (info.getMatchIndex().get() >= N) {
 152                     replicatedCount++;
 153                 }
 154             }
 155
 156             if (replicatedCount >= minReplicationCount) {
 157                 ReplicatedLogEntry replicatedLogEntry =
 158                     context.getReplicatedLog().get(N);
 159                 if (replicatedLogEntry != null
 160                     && replicatedLogEntry.getTerm()
 161                     == currentTerm()) {
 162                     context.setCommitIndex(N);
 163                 }
 164             } else {
 165                 break;
 166             }
 167         }
 168
 169         // Apply the change to the state machine
 170         if (context.getCommitIndex() > context.getLastApplied()) {
 171             applyLogToStateMachine(context.getCommitIndex());
 172         }
 173
 174         return state();
 175     }
 176
 177     protected ClientRequestTracker findClientRequestTracker(long logIndex) {
 178         for (ClientRequestTracker tracker : trackerList) {
 179             if (tracker.getIndex() == logIndex) {
 180                 return tracker;
 181             }
 182         }
 183
 184         return null;
 185     }
 186
 187     @Override protected RaftState handleRequestVoteReply(ActorRef sender,
 188         RequestVoteReply requestVoteReply) {
 189         return state();
 190     }
 191
 192     @Override public RaftState state() {
 193         return RaftState.Leader;
 194     }
 195
 196     @Override public RaftState handleMessage(ActorRef sender, Object message) {
 197         Preconditions.checkNotNull(sender, "sender should not be null");
 198
 199         if (message instanceof RaftRPC) {
 200             RaftRPC rpc = (RaftRPC) message;
 201             // If RPC request or response contains term T > currentTerm:
 202             // set currentTerm = T, convert to follower (§5.1)
 203             // This applies to all RPC messages and responses
 204             if (rpc.getTerm() > context.getTermInformation().getCurrentTerm()) {
 205                 context.getTermInformation().updateAndPersist(rpc.getTerm(), null);
 206                 return RaftState.Follower;
 207             }
 208         }
 209
 210         try {
 211             if (message instanceof SendHeartBeat) {
 212                 return sendHeartBeat();
 213             } else if(message instanceof SendInstallSnapshot) {
 214                 installSnapshotIfNeeded();
 215             } else if (message instanceof Replicate) {
 216                 replicate((Replicate) message);
 217             } else if (message instanceof InstallSnapshotReply){
 218                 // FIXME : Should I be checking the term here too?
 219                 handleInstallSnapshotReply(
 220                     (InstallSnapshotReply) message);
 221             }
 222         } finally {
 223             scheduleHeartBeat(HEART_BEAT_INTERVAL);
 224         }
 225
 226         return super.handleMessage(sender, message);
 227     }
 228
 229     private void handleInstallSnapshotReply(InstallSnapshotReply message) {
 230         InstallSnapshotReply reply = message;
 231         String followerId = reply.getFollowerId();
 232         FollowerLogInformation followerLogInformation =
 233             followerToLog.get(followerId);
 234
 235         followerLogInformation
 236             .setMatchIndex(context.getReplicatedLog().getSnapshotIndex());
 237         followerLogInformation
 238             .setNextIndex(context.getReplicatedLog().getSnapshotIndex() + 1);
 239     }
 240
 241     private void replicate(Replicate replicate) {
 242         long logIndex = replicate.getReplicatedLogEntry().getIndex();
 243
 244         context.getLogger().debug("Replicate message " + logIndex);
 245
 246         if (followerToActor.size() == 0) {
 247             context.setCommitIndex(
 248                 replicate.getReplicatedLogEntry().getIndex());
 249
 250             context.getActor()
 251                 .tell(new ApplyState(replicate.getClientActor(),
 252                         replicate.getIdentifier(),
 253                         replicate.getReplicatedLogEntry()),
 254                     context.getActor()
 255                 );
 256         } else {
 257
 258             // Create a tracker entry we will use this later to notify the
 259             // client actor
 260             trackerList.add(
 261                 new ClientRequestTrackerImpl(replicate.getClientActor(),
 262                     replicate.getIdentifier(),
 263                     logIndex)
 264             );
 265
 266             sendAppendEntries();
 267         }
 268     }
 269
 270     private void sendAppendEntries() {
 271         // Send an AppendEntries to all followers
 272         for (String followerId : followerToActor.keySet()) {
 273             ActorSelection followerActor =
 274                 followerToActor.get(followerId);
 275
 276             FollowerLogInformation followerLogInformation =
 277                 followerToLog.get(followerId);
 278
 279             long nextIndex = followerLogInformation.getNextIndex().get();
 280
 281             List<ReplicatedLogEntry> entries = Collections.emptyList();
 282
 283             if(context.getReplicatedLog().isPresent(nextIndex)){
 284                 entries =
 285                     context.getReplicatedLog().getFrom(nextIndex);
 286             }
 287
 288             followerActor.tell(
 289                 new AppendEntries(currentTerm(), context.getId(),
 290                     prevLogIndex(nextIndex), prevLogTerm(nextIndex),
 291                     entries, context.getCommitIndex()
 292                 ),
 293                 actor()
 294             );
 295         }
 296     }
 297
 298     private void installSnapshotIfNeeded(){
 299         for (String followerId : followerToActor.keySet()) {
 300             ActorSelection followerActor =
 301                 followerToActor.get(followerId);
 302
 303             FollowerLogInformation followerLogInformation =
 304                 followerToLog.get(followerId);
 305
 306             long nextIndex = followerLogInformation.getNextIndex().get();
 307
 308             if(!context.getReplicatedLog().isPresent(nextIndex) && context.getReplicatedLog().isInSnapshot(nextIndex)){
 309                 followerActor.tell(
 310                     new InstallSnapshot(currentTerm(), context.getId(),
 311                         context.getReplicatedLog().getSnapshotIndex(),
 312                         context.getReplicatedLog().getSnapshotTerm(),
 313                         context.getReplicatedLog().getSnapshot()
 314                     ),
 315                     actor()
 316                 );
 317             }
 318         }
 319     }
 320
 321     private RaftState sendHeartBeat() {
 322         if (followerToActor.size() > 0) {
 323             sendAppendEntries();
 324         }
 325         return state();
 326     }
 327
 328     private void stopHeartBeat() {
 329         if (heartbeatSchedule != null && !heartbeatSchedule.isCancelled()) {
 330             heartbeatSchedule.cancel();
 331         }
 332     }
 333
 334     private void stopInstallSnapshotSchedule() {
 335         if (installSnapshotSchedule != null && !installSnapshotSchedule.isCancelled()) {
 336             installSnapshotSchedule.cancel();
 337         }
 338     }
 339
 340     private void scheduleHeartBeat(FiniteDuration interval) {
 341         if(followerToActor.keySet().size() == 0){
 342             // Optimization - do not bother scheduling a heartbeat as there are
 343             // no followers
 344             return;
 345         }
 346
 347         stopHeartBeat();
 348
 349         // Schedule a heartbeat. When the scheduler triggers a SendHeartbeat
 350         // message is sent to itself.
 351         // Scheduling the heartbeat only once here because heartbeats do not
 352         // need to be sent if there are other messages being sent to the remote
 353         // actor.
 354         heartbeatSchedule =
 355             context.getActorSystem().scheduler().scheduleOnce(
 356                 interval,
 357                 context.getActor(), new SendHeartBeat(),
 358                 context.getActorSystem().dispatcher(), context.getActor());
 359     }
 360
 361
 362     private void scheduleInstallSnapshotCheck(FiniteDuration interval) {
 363         if(followerToActor.keySet().size() == 0){
 364             // Optimization - do not bother scheduling a heartbeat as there are
 365             // no followers
 366             return;
 367         }
 368
 369         stopInstallSnapshotSchedule();
 370
 371         // Schedule a message to send append entries to followers that can
 372         // accept an append entries with some data in it
 373         installSnapshotSchedule =
 374             context.getActorSystem().scheduler().scheduleOnce(
 375                 interval,
 376                 context.getActor(), new SendInstallSnapshot(),
 377                 context.getActorSystem().dispatcher(), context.getActor());
 378     }
 379
 380
 381
 382     @Override public void close() throws Exception {
 383         stopHeartBeat();
 384     }
 385
 386     @Override public String getLeaderId() {
 387         return context.getId();
 388     }
 389
 390 }