2 * Copyright (c) 2014 Cisco Systems, Inc. and others. All rights reserved.
3 * Copyright (c) 2015 Brocade Communications Systems, Inc. and others. All rights reserved.
5 * This program and the accompanying materials are made available under the
6 * terms of the Eclipse Public License v1.0 which accompanies this distribution,
7 * and is available at http://www.eclipse.org/legal/epl-v10.html
10 package org.opendaylight.controller.cluster.raft;
12 import akka.actor.ActorRef;
13 import akka.actor.ActorSelection;
14 import akka.actor.PoisonPill;
15 import akka.japi.Procedure;
16 import com.google.common.annotations.VisibleForTesting;
17 import com.google.common.base.Objects;
18 import com.google.common.base.Optional;
19 import com.google.common.base.Supplier;
20 import com.google.common.collect.Lists;
21 import java.io.Serializable;
22 import java.util.Collection;
23 import java.util.HashMap;
24 import java.util.List;
26 import java.util.concurrent.TimeUnit;
27 import javax.annotation.Nonnull;
28 import org.apache.commons.lang3.time.DurationFormatUtils;
29 import org.opendaylight.controller.cluster.DataPersistenceProvider;
30 import org.opendaylight.controller.cluster.DelegatingPersistentDataProvider;
31 import org.opendaylight.controller.cluster.NonPersistentDataProvider;
32 import org.opendaylight.controller.cluster.PersistentDataProvider;
33 import org.opendaylight.controller.cluster.common.actor.AbstractUntypedPersistentActor;
34 import org.opendaylight.controller.cluster.notifications.LeaderStateChanged;
35 import org.opendaylight.controller.cluster.notifications.RoleChanged;
36 import org.opendaylight.controller.cluster.raft.base.messages.ApplyJournalEntries;
37 import org.opendaylight.controller.cluster.raft.base.messages.ApplyState;
38 import org.opendaylight.controller.cluster.raft.base.messages.InitiateCaptureSnapshot;
39 import org.opendaylight.controller.cluster.raft.base.messages.LeaderTransitioning;
40 import org.opendaylight.controller.cluster.raft.base.messages.Replicate;
41 import org.opendaylight.controller.cluster.raft.base.messages.SwitchBehavior;
42 import org.opendaylight.controller.cluster.raft.behaviors.AbstractLeader;
43 import org.opendaylight.controller.cluster.raft.behaviors.DelegatingRaftActorBehavior;
44 import org.opendaylight.controller.cluster.raft.behaviors.Follower;
45 import org.opendaylight.controller.cluster.raft.behaviors.RaftActorBehavior;
46 import org.opendaylight.controller.cluster.raft.client.messages.FindLeader;
47 import org.opendaylight.controller.cluster.raft.client.messages.FindLeaderReply;
48 import org.opendaylight.controller.cluster.raft.client.messages.FollowerInfo;
49 import org.opendaylight.controller.cluster.raft.client.messages.GetOnDemandRaftState;
50 import org.opendaylight.controller.cluster.raft.client.messages.OnDemandRaftState;
51 import org.opendaylight.controller.cluster.raft.client.messages.Shutdown;
52 import org.opendaylight.controller.cluster.raft.protobuff.client.messages.Payload;
53 import org.slf4j.Logger;
54 import org.slf4j.LoggerFactory;
57 * RaftActor encapsulates a state machine that needs to be kept synchronized
58 * in a cluster. It implements the RAFT algorithm as described in the paper
59 * <a href='https://ramcloud.stanford.edu/wiki/download/attachments/11370504/raft.pdf'>
60 * In Search of an Understandable Consensus Algorithm</a>
62 * RaftActor has 3 states and each state has a certain behavior associated
63 * with it. A Raft actor can behave as,
66 * <li> A Follower (or) </li>
67 * <li> A Candidate </li>
71 * A RaftActor MUST be a Leader in order to accept requests from clients to
72 * change the state of it's encapsulated state machine. Once a RaftActor becomes
73 * a Leader it is also responsible for ensuring that all followers ultimately
74 * have the same log and therefore the same state machine as itself.
77 * The current behavior of a RaftActor determines how election for leadership
78 * is initiated and how peer RaftActors react to request for votes.
81 * Each RaftActor also needs to know the current election term. It uses this
82 * information for a couple of things. One is to simply figure out who it
83 * voted for in the last election. Another is to figure out if the message
84 * it received to update it's state is stale.
87 * The RaftActor uses akka-persistence to store it's replicated log.
88 * Furthermore through it's behaviors a Raft Actor determines
91 * <li> when a log entry should be persisted </li>
92 * <li> when a log entry should be applied to the state machine (and) </li>
93 * <li> when a snapshot should be saved </li>
96 public abstract class RaftActor extends AbstractUntypedPersistentActor {
98 private static final long APPLY_STATE_DELAY_THRESHOLD_IN_NANOS = TimeUnit.MILLISECONDS.toNanos(50L); // 50 millis
100 protected final Logger LOG = LoggerFactory.getLogger(getClass());
103 * The current state determines the current behavior of a RaftActor
104 * A Raft Actor always starts off in the Follower State
106 private final DelegatingRaftActorBehavior currentBehavior = new DelegatingRaftActorBehavior();
109 * This context should NOT be passed directly to any other actor it is
110 * only to be consumed by the RaftActorBehaviors
112 private final RaftActorContextImpl context;
114 private final DelegatingPersistentDataProvider delegatingPersistenceProvider;
116 private final PersistentDataProvider persistentProvider;
118 private RaftActorRecoverySupport raftRecovery;
120 private RaftActorSnapshotMessageSupport snapshotSupport;
122 private final BehaviorStateHolder reusableBehaviorStateHolder = new BehaviorStateHolder();
124 private final SwitchBehaviorSupplier reusableSwitchBehaviorSupplier = new SwitchBehaviorSupplier();
126 private RaftActorServerConfigurationSupport serverConfigurationSupport;
128 private RaftActorLeadershipTransferCohort leadershipTransferInProgress;
130 private boolean shuttingDown;
132 public RaftActor(String id, Map<String, String> peerAddresses,
133 Optional<ConfigParams> configParams, short payloadVersion) {
135 persistentProvider = new PersistentDataProvider(this);
136 delegatingPersistenceProvider = new RaftActorDelegatingPersistentDataProvider(null, persistentProvider);
138 context = new RaftActorContextImpl(this.getSelf(),
139 this.getContext(), id, new ElectionTermImpl(persistentProvider, id, LOG),
140 -1, -1, peerAddresses,
141 (configParams.isPresent() ? configParams.get(): new DefaultConfigParamsImpl()),
142 delegatingPersistenceProvider, LOG);
144 context.setPayloadVersion(payloadVersion);
145 context.setReplicatedLog(ReplicatedLogImpl.newInstance(context, currentBehavior));
149 public void preStart() throws Exception {
150 LOG.info("Starting recovery for {} with journal batch size {}", persistenceId(),
151 context.getConfigParams().getJournalRecoveryLogBatchSize());
155 snapshotSupport = newRaftActorSnapshotMessageSupport();
156 serverConfigurationSupport = new RaftActorServerConfigurationSupport(this);
160 public void postStop() {
161 if(currentBehavior.getDelegate() != null) {
163 currentBehavior.close();
164 } catch (Exception e) {
165 LOG.debug("{}: Error closing behavior {}", persistenceId(), currentBehavior.state());
173 public void handleRecover(Object message) {
174 if(raftRecovery == null) {
175 raftRecovery = newRaftActorRecoverySupport();
178 boolean recoveryComplete = raftRecovery.handleRecoveryMessage(message, persistentProvider);
179 if(recoveryComplete) {
180 onRecoveryComplete();
182 initializeBehavior();
186 if (context.getReplicatedLog().size() > 0) {
187 self().tell(new InitiateCaptureSnapshot(), self());
188 LOG.info("{}: Snapshot capture initiated after recovery", persistenceId());
190 LOG.info("{}: Snapshot capture NOT initiated after recovery, journal empty", persistenceId());
195 protected RaftActorRecoverySupport newRaftActorRecoverySupport() {
196 return new RaftActorRecoverySupport(context, currentBehavior, getRaftActorRecoveryCohort());
199 protected void initializeBehavior(){
200 changeCurrentBehavior(new Follower(context));
203 protected void changeCurrentBehavior(RaftActorBehavior newBehavior){
204 reusableBehaviorStateHolder.init(getCurrentBehavior());
205 setCurrentBehavior(newBehavior);
206 handleBehaviorChange(reusableBehaviorStateHolder, getCurrentBehavior());
210 public void handleCommand(final Object message) {
211 if(serverConfigurationSupport.handleMessage(message, getSender())) {
213 } else if (message instanceof ApplyState){
214 ApplyState applyState = (ApplyState) message;
216 long elapsedTime = (System.nanoTime() - applyState.getStartTime());
217 if(elapsedTime >= APPLY_STATE_DELAY_THRESHOLD_IN_NANOS){
218 LOG.warn("ApplyState took more time than expected. Elapsed Time = {} ms ApplyState = {}",
219 TimeUnit.NANOSECONDS.toMillis(elapsedTime), applyState);
222 if(LOG.isDebugEnabled()) {
223 LOG.debug("{}: Applying state for log index {} data {}",
224 persistenceId(), applyState.getReplicatedLogEntry().getIndex(),
225 applyState.getReplicatedLogEntry().getData());
228 applyState(applyState.getClientActor(), applyState.getIdentifier(),
229 applyState.getReplicatedLogEntry().getData());
231 if (!hasFollowers()) {
232 // for single node, the capture should happen after the apply state
233 // as we delete messages from the persistent journal which have made it to the snapshot
234 // capturing the snapshot before applying makes the persistent journal and snapshot out of sync
235 // and recovery shows data missing
236 context.getReplicatedLog().captureSnapshotIfReady(applyState.getReplicatedLogEntry());
238 context.getSnapshotManager().trimLog(context.getLastApplied(), currentBehavior);
241 } else if (message instanceof ApplyJournalEntries){
242 ApplyJournalEntries applyEntries = (ApplyJournalEntries) message;
243 if(LOG.isDebugEnabled()) {
244 LOG.debug("{}: Persisting ApplyLogEntries with index={}", persistenceId(), applyEntries.getToIndex());
247 persistence().persist(applyEntries, NoopProcedure.instance());
249 } else if (message instanceof FindLeader) {
251 new FindLeaderReply(getLeaderAddress()),
254 } else if(message instanceof GetOnDemandRaftState) {
255 onGetOnDemandRaftStats();
256 } else if(message instanceof InitiateCaptureSnapshot) {
258 } else if(message instanceof SwitchBehavior){
259 switchBehavior(((SwitchBehavior) message));
260 } else if(message instanceof LeaderTransitioning) {
261 onLeaderTransitioning();
262 } else if(message instanceof Shutdown) {
264 } else if(message instanceof Runnable) {
265 ((Runnable)message).run();
266 } else if(!snapshotSupport.handleSnapshotMessage(message, getSender())) {
267 switchBehavior(reusableSwitchBehaviorSupplier.handleMessage(getSender(), message));
271 private void initiateLeadershipTransfer(final RaftActorLeadershipTransferCohort.OnComplete onComplete) {
272 LOG.debug("{}: Initiating leader transfer", persistenceId());
274 if(leadershipTransferInProgress == null) {
275 leadershipTransferInProgress = new RaftActorLeadershipTransferCohort(this, getSender());
276 leadershipTransferInProgress.addOnComplete(new RaftActorLeadershipTransferCohort.OnComplete() {
278 public void onSuccess(ActorRef raftActorRef, ActorRef replyTo) {
279 leadershipTransferInProgress = null;
283 public void onFailure(ActorRef raftActorRef, ActorRef replyTo) {
284 leadershipTransferInProgress = null;
288 leadershipTransferInProgress.addOnComplete(onComplete);
289 leadershipTransferInProgress.init();
291 LOG.debug("{}: prior leader transfer in progress - adding callback", persistenceId());
292 leadershipTransferInProgress.addOnComplete(onComplete);
296 private void onShutDown() {
297 LOG.debug("{}: onShutDown", persistenceId());
304 if(currentBehavior.state() == RaftState.Leader && context.hasFollowers()) {
305 initiateLeadershipTransfer(new RaftActorLeadershipTransferCohort.OnComplete() {
307 public void onSuccess(ActorRef raftActorRef, ActorRef replyTo) {
308 LOG.debug("{}: leader transfer succeeded - sending PoisonPill", persistenceId());
309 raftActorRef.tell(PoisonPill.getInstance(), raftActorRef);
313 public void onFailure(ActorRef raftActorRef, ActorRef replyTo) {
314 LOG.debug("{}: leader transfer failed - sending PoisonPill", persistenceId());
315 raftActorRef.tell(PoisonPill.getInstance(), raftActorRef);
318 } else if(currentBehavior.state() == RaftState.Leader) {
319 pauseLeader(new TimedRunnable(context.getConfigParams().getElectionTimeOutInterval(), this) {
321 protected void doRun() {
322 self().tell(PoisonPill.getInstance(), self());
326 protected void doCancel() {
327 self().tell(PoisonPill.getInstance(), self());
331 self().tell(PoisonPill.getInstance(), self());
335 private void onLeaderTransitioning() {
336 LOG.debug("{}: onLeaderTransitioning", persistenceId());
337 Optional<ActorRef> roleChangeNotifier = getRoleChangeNotifier();
338 if(currentBehavior.state() == RaftState.Follower && roleChangeNotifier.isPresent()) {
339 roleChangeNotifier.get().tell(newLeaderStateChanged(getId(), null,
340 currentBehavior.getLeaderPayloadVersion()), getSelf());
344 private void switchBehavior(SwitchBehavior message) {
345 if(!getRaftActorContext().getRaftPolicy().automaticElectionsEnabled()) {
346 RaftState newState = message.getNewState();
347 if( newState == RaftState.Leader || newState == RaftState.Follower) {
348 switchBehavior(reusableSwitchBehaviorSupplier.handleMessage(getSender(), message));
349 getRaftActorContext().getTermInformation().updateAndPersist(message.getNewTerm(), "");
351 LOG.warn("Switching to behavior : {} - not supported", newState);
356 private void switchBehavior(Supplier<RaftActorBehavior> supplier){
357 reusableBehaviorStateHolder.init(getCurrentBehavior());
359 setCurrentBehavior(supplier.get());
361 handleBehaviorChange(reusableBehaviorStateHolder, getCurrentBehavior());
364 protected RaftActorSnapshotMessageSupport newRaftActorSnapshotMessageSupport() {
365 return new RaftActorSnapshotMessageSupport(context, currentBehavior,
366 getRaftActorSnapshotCohort());
369 private void onGetOnDemandRaftStats() {
370 // Debugging message to retrieve raft stats.
372 Map<String, String> peerAddresses = new HashMap<>();
373 for(String peerId: context.getPeerIds()) {
374 peerAddresses.put(peerId, context.getPeerAddress(peerId));
377 OnDemandRaftState.Builder builder = OnDemandRaftState.builder()
378 .commitIndex(context.getCommitIndex())
379 .currentTerm(context.getTermInformation().getCurrentTerm())
380 .inMemoryJournalDataSize(replicatedLog().dataSize())
381 .inMemoryJournalLogSize(replicatedLog().size())
382 .isSnapshotCaptureInitiated(context.getSnapshotManager().isCapturing())
383 .lastApplied(context.getLastApplied())
384 .lastIndex(replicatedLog().lastIndex())
385 .lastTerm(replicatedLog().lastTerm())
386 .leader(getLeaderId())
387 .raftState(currentBehavior.state().toString())
388 .replicatedToAllIndex(currentBehavior.getReplicatedToAllIndex())
389 .snapshotIndex(replicatedLog().getSnapshotIndex())
390 .snapshotTerm(replicatedLog().getSnapshotTerm())
391 .votedFor(context.getTermInformation().getVotedFor())
392 .peerAddresses(peerAddresses)
393 .customRaftPolicyClassName(context.getConfigParams().getCustomRaftPolicyImplementationClass());
395 ReplicatedLogEntry lastLogEntry = getLastLogEntry();
396 if (lastLogEntry != null) {
397 builder.lastLogIndex(lastLogEntry.getIndex());
398 builder.lastLogTerm(lastLogEntry.getTerm());
401 if(getCurrentBehavior() instanceof AbstractLeader) {
402 AbstractLeader leader = (AbstractLeader)getCurrentBehavior();
403 Collection<String> followerIds = leader.getFollowerIds();
404 List<FollowerInfo> followerInfoList = Lists.newArrayListWithCapacity(followerIds.size());
405 for(String id: followerIds) {
406 final FollowerLogInformation info = leader.getFollower(id);
407 followerInfoList.add(new FollowerInfo(id, info.getNextIndex(), info.getMatchIndex(),
408 info.isFollowerActive(), DurationFormatUtils.formatDurationHMS(info.timeSinceLastActivity())));
411 builder.followerInfoList(followerInfoList);
414 sender().tell(builder.build(), self());
418 private void handleBehaviorChange(BehaviorStateHolder oldBehaviorState, RaftActorBehavior currentBehavior) {
419 RaftActorBehavior oldBehavior = oldBehaviorState.getBehavior();
421 if (oldBehavior != currentBehavior){
425 String lastValidLeaderId = oldBehavior == null ? null : oldBehaviorState.getLastValidLeaderId();
426 String oldBehaviorStateName = oldBehavior == null ? null : oldBehavior.state().name();
428 // it can happen that the state has not changed but the leader has changed.
429 Optional<ActorRef> roleChangeNotifier = getRoleChangeNotifier();
430 if(!Objects.equal(lastValidLeaderId, currentBehavior.getLeaderId()) ||
431 oldBehaviorState.getLeaderPayloadVersion() != currentBehavior.getLeaderPayloadVersion()) {
432 if(roleChangeNotifier.isPresent()) {
433 roleChangeNotifier.get().tell(newLeaderStateChanged(getId(), currentBehavior.getLeaderId(),
434 currentBehavior.getLeaderPayloadVersion()), getSelf());
437 onLeaderChanged(lastValidLeaderId, currentBehavior.getLeaderId());
439 if(leadershipTransferInProgress != null) {
440 leadershipTransferInProgress.onNewLeader(currentBehavior.getLeaderId());
444 if (roleChangeNotifier.isPresent() &&
445 (oldBehavior == null || (oldBehavior.state() != currentBehavior.state()))) {
446 roleChangeNotifier.get().tell(new RoleChanged(getId(), oldBehaviorStateName ,
447 currentBehavior.state().name()), getSelf());
451 protected LeaderStateChanged newLeaderStateChanged(String memberId, String leaderId, short leaderPayloadVersion) {
452 return new LeaderStateChanged(memberId, leaderId, leaderPayloadVersion);
456 public long snapshotSequenceNr() {
457 // When we do a snapshot capture, we also capture and save the sequence-number of the persistent journal,
458 // so that we can delete the persistent journal based on the saved sequence-number
459 // However , when akka replays the journal during recovery, it replays it from the sequence number when the snapshot
460 // was saved and not the number we saved.
461 // We would want to override it , by asking akka to use the last-sequence number known to us.
462 return context.getSnapshotManager().getLastSequenceNumber();
466 * When a derived RaftActor needs to persist something it must call
473 protected void persistData(final ActorRef clientActor, final String identifier,
474 final Payload data) {
476 ReplicatedLogEntry replicatedLogEntry = new ReplicatedLogImplEntry(
477 context.getReplicatedLog().lastIndex() + 1,
478 context.getTermInformation().getCurrentTerm(), data);
480 if(LOG.isDebugEnabled()) {
481 LOG.debug("{}: Persist data {}", persistenceId(), replicatedLogEntry);
484 final RaftActorContext raftContext = getRaftActorContext();
486 replicatedLog().appendAndPersist(replicatedLogEntry, new Procedure<ReplicatedLogEntry>() {
488 public void apply(ReplicatedLogEntry replicatedLogEntry) {
489 if (!hasFollowers()){
490 // Increment the Commit Index and the Last Applied values
491 raftContext.setCommitIndex(replicatedLogEntry.getIndex());
492 raftContext.setLastApplied(replicatedLogEntry.getIndex());
494 // Apply the state immediately.
495 self().tell(new ApplyState(clientActor, identifier, replicatedLogEntry), self());
497 // Send a ApplyJournalEntries message so that we write the fact that we applied
498 // the state to durable storage
499 self().tell(new ApplyJournalEntries(replicatedLogEntry.getIndex()), self());
501 } else if (clientActor != null) {
502 context.getReplicatedLog().captureSnapshotIfReady(replicatedLogEntry);
504 // Send message for replication
505 currentBehavior.handleMessage(getSelf(),
506 new Replicate(clientActor, identifier, replicatedLogEntry));
512 private ReplicatedLog replicatedLog() {
513 return context.getReplicatedLog();
516 protected String getId() {
517 return context.getId();
521 void setCurrentBehavior(RaftActorBehavior behavior) {
522 currentBehavior.setDelegate(behavior);
525 protected RaftActorBehavior getCurrentBehavior() {
526 return currentBehavior.getDelegate();
530 * Derived actors can call the isLeader method to check if the current
531 * RaftActor is the Leader or not
533 * @return true it this RaftActor is a Leader false otherwise
535 protected boolean isLeader() {
536 return context.getId().equals(currentBehavior.getLeaderId());
539 protected boolean isLeaderActive() {
540 return currentBehavior.state() != RaftState.IsolatedLeader && !shuttingDown &&
541 !isLeadershipTransferInProgress();
544 private boolean isLeadershipTransferInProgress() {
545 return leadershipTransferInProgress != null && leadershipTransferInProgress.isTransferring();
549 * Derived actor can call getLeader if they need a reference to the Leader.
550 * This would be useful for example in forwarding a request to an actor
551 * which is the leader
553 * @return A reference to the leader if known, null otherwise
555 protected ActorSelection getLeader(){
556 String leaderAddress = getLeaderAddress();
558 if(leaderAddress == null){
562 return context.actorSelection(leaderAddress);
567 * @return the current leader's id
569 protected String getLeaderId(){
570 return currentBehavior.getLeaderId();
573 protected RaftState getRaftState() {
574 return currentBehavior.state();
577 protected ReplicatedLogEntry getLastLogEntry() {
578 return replicatedLog().last();
581 protected Long getCurrentTerm(){
582 return context.getTermInformation().getCurrentTerm();
585 protected Long getCommitIndex(){
586 return context.getCommitIndex();
589 protected Long getLastApplied(){
590 return context.getLastApplied();
593 protected RaftActorContext getRaftActorContext() {
597 protected void updateConfigParams(ConfigParams configParams) {
599 // obtain the RaftPolicy for oldConfigParams and the updated one.
600 String oldRaftPolicy = context.getConfigParams().
601 getCustomRaftPolicyImplementationClass();
602 String newRaftPolicy = configParams.
603 getCustomRaftPolicyImplementationClass();
605 LOG.debug("{}: RaftPolicy used with prev.config {}, RaftPolicy used with newConfig {}", persistenceId(),
606 oldRaftPolicy, newRaftPolicy);
607 context.setConfigParams(configParams);
608 if (!Objects.equal(oldRaftPolicy, newRaftPolicy)) {
609 // The RaftPolicy was modified. If the current behavior is Follower then re-initialize to Follower
610 // but transfer the previous leaderId so it doesn't immediately try to schedule an election. This
611 // avoids potential disruption. Otherwise, switch to Follower normally.
612 RaftActorBehavior behavior = currentBehavior.getDelegate();
613 if(behavior instanceof Follower) {
614 String previousLeaderId = ((Follower)behavior).getLeaderId();
616 LOG.debug("{}: Re-initializing to Follower with previous leaderId {}", persistenceId(), previousLeaderId);
618 changeCurrentBehavior(new Follower(context, previousLeaderId));
620 initializeBehavior();
625 public final DataPersistenceProvider persistence() {
626 return delegatingPersistenceProvider.getDelegate();
629 public void setPersistence(DataPersistenceProvider provider) {
630 delegatingPersistenceProvider.setDelegate(provider);
633 protected void setPersistence(boolean persistent) {
635 setPersistence(new PersistentDataProvider(this));
637 setPersistence(new NonPersistentDataProvider() {
639 * The way snapshotting works is,
641 * <li> RaftActor calls createSnapshot on the Shard
642 * <li> Shard sends a CaptureSnapshotReply and RaftActor then calls saveSnapshot
643 * <li> When saveSnapshot is invoked on the akka-persistence API it uses the SnapshotStore to save
644 * the snapshot. The SnapshotStore sends SaveSnapshotSuccess or SaveSnapshotFailure. When the
645 * RaftActor gets SaveSnapshot success it commits the snapshot to the in-memory journal. This
646 * commitSnapshot is mimicking what is done in SaveSnapshotSuccess.
650 public void saveSnapshot(Object o) {
651 // Make saving Snapshot successful
652 // Committing the snapshot here would end up calling commit in the creating state which would
653 // be a state violation. That's why now we send a message to commit the snapshot.
654 self().tell(RaftActorSnapshotMessageSupport.COMMIT_SNAPSHOT, self());
661 * setPeerAddress sets the address of a known peer at a later time.
663 * This is to account for situations where a we know that a peer
664 * exists but we do not know an address up-front. This may also be used in
665 * situations where a known peer starts off in a different location and we
666 * need to change it's address
668 * Note that if the peerId does not match the list of peers passed to
669 * this actor during construction an IllegalStateException will be thrown.
674 protected void setPeerAddress(String peerId, String peerAddress){
675 context.setPeerAddress(peerId, peerAddress);
679 * The applyState method will be called by the RaftActor when some data
680 * needs to be applied to the actor's state
682 * @param clientActor A reference to the client who sent this message. This
683 * is the same reference that was passed to persistData
684 * by the derived actor. clientActor may be null when
685 * the RaftActor is behaving as a follower or during
687 * @param identifier The identifier of the persisted data. This is also
688 * the same identifier that was passed to persistData by
689 * the derived actor. identifier may be null when
690 * the RaftActor is behaving as a follower or during
692 * @param data A piece of data that was persisted by the persistData call.
693 * This should NEVER be null.
695 protected abstract void applyState(ActorRef clientActor, String identifier,
699 * Returns the RaftActorRecoveryCohort to participate in persistence recovery.
702 protected abstract RaftActorRecoveryCohort getRaftActorRecoveryCohort();
705 * This method is called when recovery is complete.
707 protected abstract void onRecoveryComplete();
710 * Returns the RaftActorSnapshotCohort to participate in persistence recovery.
713 protected abstract RaftActorSnapshotCohort getRaftActorSnapshotCohort();
716 * This method will be called by the RaftActor when the state of the
717 * RaftActor changes. The derived actor can then use methods like
718 * isLeader or getLeader to do something useful
720 protected abstract void onStateChanged();
723 * Notifier Actor for this RaftActor to notify when a role change happens
724 * @return ActorRef - ActorRef of the notifier or Optional.absent if none.
726 protected abstract Optional<ActorRef> getRoleChangeNotifier();
729 * This method is called prior to operations such as leadership transfer and actor shutdown when the leader
730 * must pause or stop its duties. This method allows derived classes to gracefully pause or finish current
731 * work prior to performing the operation. On completion of any work, the run method must be called on the
732 * given Runnable to proceed with the given operation. <b>Important:</b> the run method must be called on
733 * this actor's thread dispatcher as as it modifies internal state.
735 * The default implementation immediately runs the operation.
737 * @param operation the operation to run
739 protected void pauseLeader(Runnable operation) {
743 protected void onLeaderChanged(String oldLeader, String newLeader){};
745 private String getLeaderAddress(){
747 return getSelf().path().toString();
749 String leaderId = currentBehavior.getLeaderId();
750 if (leaderId == null) {
753 String peerAddress = context.getPeerAddress(leaderId);
754 if(LOG.isDebugEnabled()) {
755 LOG.debug("{}: getLeaderAddress leaderId = {} peerAddress = {}",
756 persistenceId(), leaderId, peerAddress);
762 protected boolean hasFollowers(){
763 return getRaftActorContext().hasFollowers();
766 private void captureSnapshot() {
767 SnapshotManager snapshotManager = context.getSnapshotManager();
769 if(!snapshotManager.isCapturing()) {
770 LOG.debug("Take a snapshot of current state. lastReplicatedLog is {} and replicatedToAllIndex is {}",
771 replicatedLog().last(), currentBehavior.getReplicatedToAllIndex());
773 snapshotManager.capture(replicatedLog().last(), currentBehavior.getReplicatedToAllIndex());
778 * @deprecated Deprecated in favor of {@link org.opendaylight.controller.cluster.raft.base.messages.DeleteEntries}
779 * whose type for fromIndex is long instead of int. This class was kept for backwards
780 * compatibility with Helium.
782 // Suppressing this warning as we can't set serialVersionUID to maintain backwards compatibility.
783 @SuppressWarnings("serial")
785 static class DeleteEntries implements Serializable {
786 private final int fromIndex;
788 public DeleteEntries(int fromIndex) {
789 this.fromIndex = fromIndex;
792 public int getFromIndex() {
798 * @deprecated Deprecated in favor of non-inner class {@link org.opendaylight.controller.cluster.raft.base.messages.UpdateElectionTerm}
799 * which has serialVersionUID set. This class was kept for backwards compatibility with Helium.
801 // Suppressing this warning as we can't set serialVersionUID to maintain backwards compatibility.
802 @SuppressWarnings("serial")
804 static class UpdateElectionTerm implements Serializable {
805 private final long currentTerm;
806 private final String votedFor;
808 public UpdateElectionTerm(long currentTerm, String votedFor) {
809 this.currentTerm = currentTerm;
810 this.votedFor = votedFor;
813 public long getCurrentTerm() {
817 public String getVotedFor() {
822 private static class BehaviorStateHolder {
823 private RaftActorBehavior behavior;
824 private String lastValidLeaderId;
825 private short leaderPayloadVersion;
827 void init(RaftActorBehavior behavior) {
828 this.behavior = behavior;
829 this.leaderPayloadVersion = behavior != null ? behavior.getLeaderPayloadVersion() : -1;
831 String behaviorLeaderId = behavior != null ? behavior.getLeaderId() : null;
832 if(behaviorLeaderId != null) {
833 this.lastValidLeaderId = behaviorLeaderId;
837 RaftActorBehavior getBehavior() {
841 String getLastValidLeaderId() {
842 return lastValidLeaderId;
845 short getLeaderPayloadVersion() {
846 return leaderPayloadVersion;
850 private class SwitchBehaviorSupplier implements Supplier<RaftActorBehavior> {
851 private Object message;
852 private ActorRef sender;
854 public SwitchBehaviorSupplier handleMessage(ActorRef sender, Object message){
855 this.sender = sender;
856 this.message = message;
861 public RaftActorBehavior get() {
862 if(this.message instanceof SwitchBehavior){
863 return ((SwitchBehavior) message).getNewState().createBehavior(getRaftActorContext());
865 return currentBehavior.handleMessage(sender, message);