2 * Copyright (c) 2016 2015 Brocade Communications Systems, Inc. and others. All rights reserved.
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
8 package org.opendaylight.controller.cluster.raft;
10 import static org.junit.Assert.assertEquals;
11 import static org.junit.Assert.fail;
12 import static org.opendaylight.controller.cluster.raft.utils.MessageCollectorActor.assertNoneMatching;
13 import static org.opendaylight.controller.cluster.raft.utils.MessageCollectorActor.clearMessages;
14 import static org.opendaylight.controller.cluster.raft.utils.MessageCollectorActor.expectFirstMatching;
15 import static org.opendaylight.controller.cluster.raft.utils.MessageCollectorActor.expectMatching;
16 import static org.opendaylight.controller.cluster.raft.utils.MessageCollectorActor.getAllMatching;
18 import akka.actor.Actor;
19 import akka.actor.ActorRef;
20 import akka.actor.Props;
21 import akka.testkit.TestActorRef;
22 import com.google.common.collect.ImmutableMap;
23 import com.google.common.collect.Lists;
24 import java.util.List;
25 import java.util.concurrent.TimeUnit;
26 import org.junit.Test;
27 import org.opendaylight.controller.cluster.notifications.RoleChanged;
28 import org.opendaylight.controller.cluster.raft.MockRaftActorContext.MockPayload;
29 import org.opendaylight.controller.cluster.raft.base.messages.ApplyState;
30 import org.opendaylight.controller.cluster.raft.base.messages.TimeoutNow;
31 import org.opendaylight.controller.cluster.raft.messages.AppendEntries;
32 import org.opendaylight.controller.cluster.raft.messages.AppendEntriesReply;
33 import org.opendaylight.controller.cluster.raft.messages.InstallSnapshot;
34 import org.opendaylight.controller.cluster.raft.messages.RequestVote;
35 import org.opendaylight.controller.cluster.raft.utils.MessageCollectorActor;
36 import scala.concurrent.duration.FiniteDuration;
39 * Tests isolation of nodes end-to-end.
41 * @author Thomas Pantelis
43 public class IsolationScenarioTest extends AbstractRaftActorIntegrationTest {
44 private TestActorRef<Actor> follower1NotifierActor;
45 private TestActorRef<Actor> leaderNotifierActor;
48 * Isolates the leader after all initial payload entries have been committed and applied on all nodes. While
49 * isolated, the majority partition elects a new leader and both sides of the partition attempt to commit one entry
50 * independently. After isolation is removed, the entry will conflict and both sides should reconcile their logs
54 public void testLeaderIsolationWithAllPriorEntriesCommitted() throws Exception {
55 testLog.info("testLeaderIsolationWithAllPriorEntriesCommitted starting");
59 // Send an initial payloads and verify replication.
61 final MockPayload payload0 = sendPayloadData(leaderActor, "zero");
62 final MockPayload payload1 = sendPayloadData(leaderActor, "one");
63 verifyApplyJournalEntries(leaderCollectorActor, 1);
64 verifyApplyJournalEntries(follower1CollectorActor, 1);
65 verifyApplyJournalEntries(follower2CollectorActor, 1);
69 // Send a payload to the isolated leader so it has an uncommitted log entry with index 2.
71 testLog.info("Sending payload to isolated leader");
73 final MockPayload isolatedLeaderPayload2 = sendPayloadData(leaderActor, "two");
75 // Wait for the isolated leader to send AppendEntries to follower1 with the entry at index 2. Note the message
76 // is collected but not forwarded to the follower RaftActor.
78 AppendEntries appendEntries = expectFirstMatching(follower1CollectorActor, AppendEntries.class);
79 assertEquals("getTerm", currentTerm, appendEntries.getTerm());
80 assertEquals("getLeaderId", leaderId, appendEntries.getLeaderId());
81 assertEquals("getEntries().size()", 1, appendEntries.getEntries().size());
82 verifyReplicatedLogEntry(appendEntries.getEntries().get(0), currentTerm, 2, isolatedLeaderPayload2);
84 // The leader should transition to IsolatedLeader.
86 expectFirstMatching(leaderNotifierActor, RoleChanged.class,
87 rc -> rc.getNewRole().equals(RaftState.IsolatedLeader.name()));
89 forceElectionOnFollower1();
91 // Send a payload to the new leader follower1 with index 2 and verify it's replicated to follower2
94 testLog.info("Sending payload to new leader");
96 final MockPayload newLeaderPayload2 = sendPayloadData(follower1Actor, "two-new");
97 verifyApplyJournalEntries(follower1CollectorActor, 2);
98 verifyApplyJournalEntries(follower2CollectorActor, 2);
100 assertEquals("Follower 1 journal last term", currentTerm, follower1Context.getReplicatedLog().lastTerm());
101 assertEquals("Follower 1 journal last index", 2, follower1Context.getReplicatedLog().lastIndex());
102 assertEquals("Follower 1 commit index", 2, follower1Context.getCommitIndex());
103 verifyReplicatedLogEntry(follower1Context.getReplicatedLog().get(2), currentTerm, 2, newLeaderPayload2);
105 assertEquals("Follower 1 state", Lists.newArrayList(payload0, payload1, newLeaderPayload2),
106 follower1Actor.underlyingActor().getState());
110 // Previous leader should switch to follower b/c it will receive either an AppendEntries or AppendEntriesReply
111 // with a higher term.
113 expectFirstMatching(leaderNotifierActor, RoleChanged.class,
114 rc -> rc.getNewRole().equals(RaftState.Follower.name()));
116 // The previous leader has a conflicting log entry at index 2 with a different term which should get
117 // replaced by the new leader's index 1 entry.
119 verifyApplyJournalEntries(leaderCollectorActor, 2);
121 assertEquals("Prior leader journal last term", currentTerm, leaderContext.getReplicatedLog().lastTerm());
122 assertEquals("Prior leader journal last index", 2, leaderContext.getReplicatedLog().lastIndex());
123 assertEquals("Prior leader commit index", 2, leaderContext.getCommitIndex());
124 verifyReplicatedLogEntry(leaderContext.getReplicatedLog().get(2), currentTerm, 2, newLeaderPayload2);
126 assertEquals("Prior leader state", Lists.newArrayList(payload0, payload1, newLeaderPayload2),
127 leaderActor.underlyingActor().getState());
129 testLog.info("testLeaderIsolationWithAllPriorEntriesCommitted ending");
133 * Isolates the leader with a payload entry that's replicated to all followers and committed on the leader but
134 * uncommitted on the followers. While isolated, the majority partition elects a new leader and both sides of the
135 * partition attempt to commit one entry independently. After isolation is removed, the entry will conflict and both
136 * sides should reconcile their logs appropriately.
139 public void testLeaderIsolationWithPriorUncommittedEntryAndOneConflictingEntry() throws Exception {
140 testLog.info("testLeaderIsolationWithPriorUncommittedEntryAndOneConflictingEntry starting");
144 // Submit an initial payload that is committed/applied on all nodes.
146 final MockPayload payload0 = sendPayloadData(leaderActor, "zero");
147 verifyApplyJournalEntries(leaderCollectorActor, 0);
148 verifyApplyJournalEntries(follower1CollectorActor, 0);
149 verifyApplyJournalEntries(follower2CollectorActor, 0);
151 // Submit another payload that is replicated to all followers and committed on the leader but the leader is
152 // isolated before the entry is committed on the followers. To accomplish this we drop the AppendEntries
153 // with the updated leader commit index.
155 follower1Actor.underlyingActor().startDropMessages(AppendEntries.class, ae -> ae.getLeaderCommit() == 1);
156 follower2Actor.underlyingActor().startDropMessages(AppendEntries.class, ae -> ae.getLeaderCommit() == 1);
158 MockPayload payload1 = sendPayloadData(leaderActor, "one");
160 // Wait for the isolated leader to send AppendEntries to the followers with the new entry with index 1. This
161 // message is forwarded to the followers.
163 expectFirstMatching(follower1CollectorActor, AppendEntries.class, ae ->
164 ae.getEntries().size() == 1 && ae.getEntries().get(0).getIndex() == 1
165 && ae.getEntries().get(0).getData().equals(payload1));
167 expectFirstMatching(follower2CollectorActor, AppendEntries.class, ae ->
168 ae.getEntries().size() == 1 && ae.getEntries().get(0).getIndex() == 1
169 && ae.getEntries().get(0).getData().equals(payload1));
171 verifyApplyJournalEntries(leaderCollectorActor, 1);
175 // Send a payload to the isolated leader so it has an uncommitted log entry with index 2.
177 testLog.info("Sending payload to isolated leader");
179 final MockPayload isolatedLeaderPayload2 = sendPayloadData(leaderActor, "two");
181 // Wait for the isolated leader to send AppendEntries to follower1 with the entry at index 2. Note the message
182 // is collected but not forwarded to the follower RaftActor.
184 AppendEntries appendEntries = expectFirstMatching(follower1CollectorActor, AppendEntries.class);
185 assertEquals("getTerm", currentTerm, appendEntries.getTerm());
186 assertEquals("getLeaderId", leaderId, appendEntries.getLeaderId());
187 assertEquals("getEntries().size()", 1, appendEntries.getEntries().size());
188 verifyReplicatedLogEntry(appendEntries.getEntries().get(0), currentTerm, 2, isolatedLeaderPayload2);
190 // The leader should transition to IsolatedLeader.
192 expectFirstMatching(leaderNotifierActor, RoleChanged.class,
193 rc -> rc.getNewRole().equals(RaftState.IsolatedLeader.name()));
195 forceElectionOnFollower1();
197 // Send a payload to the new leader follower1 and verify it's replicated to follower2 and committed. Since the
198 // entry with index 1 from the previous term was uncommitted, the new leader should've also committed a
199 // NoopPayload entry with index 2 in the PreLeader state. Thus the new payload will have index 3.
201 testLog.info("Sending payload to new leader");
203 final MockPayload newLeaderPayload2 = sendPayloadData(follower1Actor, "two-new");
204 verifyApplyJournalEntries(follower1CollectorActor, 3);
205 verifyApplyJournalEntries(follower2CollectorActor, 3);
207 assertEquals("Follower 1 journal last term", currentTerm, follower1Context.getReplicatedLog().lastTerm());
208 assertEquals("Follower 1 journal last index", 3, follower1Context.getReplicatedLog().lastIndex());
209 assertEquals("Follower 1 commit index", 3, follower1Context.getCommitIndex());
210 verifyReplicatedLogEntry(follower1Context.getReplicatedLog().get(3), currentTerm, 3, newLeaderPayload2);
212 assertEquals("Follower 1 state", Lists.newArrayList(payload0, payload1, newLeaderPayload2),
213 follower1Actor.underlyingActor().getState());
217 // Previous leader should switch to follower b/c it will receive either an AppendEntries or AppendEntriesReply
218 // with a higher term.
220 expectFirstMatching(leaderNotifierActor, RoleChanged.class,
221 rc -> rc.getNewRole().equals(RaftState.Follower.name()));
223 // The previous leader has a conflicting log entry at index 2 with a different term which should get
224 // replaced by the new leader's entry.
226 verifyApplyJournalEntries(leaderCollectorActor, 3);
228 verifyRaftState(leaderActor, raftState -> {
229 assertEquals("Prior leader journal last term", currentTerm, leaderContext.getReplicatedLog().lastTerm());
230 assertEquals("Prior leader journal last index", 3, leaderContext.getReplicatedLog().lastIndex());
231 assertEquals("Prior leader commit index", 3, leaderContext.getCommitIndex());
234 assertEquals("Prior leader state", Lists.newArrayList(payload0, payload1, newLeaderPayload2),
235 leaderActor.underlyingActor().getState());
237 // Ensure the prior leader didn't apply its conflicting entry with index 2, term 1.
239 List<ApplyState> applyState = getAllMatching(leaderCollectorActor, ApplyState.class);
240 for (ApplyState as: applyState) {
241 if (as.getReplicatedLogEntry().getIndex() == 2 && as.getReplicatedLogEntry().getTerm() == 1) {
242 fail("Got unexpected ApplyState: " + as);
246 // The prior leader should not have needed a snapshot installed in order to get it synced.
248 assertNoneMatching(leaderCollectorActor, InstallSnapshot.class);
250 testLog.info("testLeaderIsolationWithPriorUncommittedEntryAndOneConflictingEntry ending");
254 * Isolates the leader with a payload entry that's replicated to all followers and committed on the leader but
255 * uncommitted on the followers. While isolated, the majority partition elects a new leader and both sides of the
256 * partition attempt to commit multiple entries independently. After isolation is removed, the entries will conflict
257 * and both sides should reconcile their logs appropriately.
260 public void testLeaderIsolationWithPriorUncommittedEntryAndMultipleConflictingEntries() throws Exception {
261 testLog.info("testLeaderIsolationWithPriorUncommittedEntryAndMultipleConflictingEntries starting");
265 // Submit an initial payload that is committed/applied on all nodes.
267 final MockPayload payload0 = sendPayloadData(leaderActor, "zero");
268 verifyApplyJournalEntries(leaderCollectorActor, 0);
269 verifyApplyJournalEntries(follower1CollectorActor, 0);
270 verifyApplyJournalEntries(follower2CollectorActor, 0);
272 // Submit another payload that is replicated to all followers and committed on the leader but the leader is
273 // isolated before the entry is committed on the followers. To accomplish this we drop the AppendEntries
274 // with the updated leader commit index.
276 follower1Actor.underlyingActor().startDropMessages(AppendEntries.class, ae -> ae.getLeaderCommit() == 1);
277 follower2Actor.underlyingActor().startDropMessages(AppendEntries.class, ae -> ae.getLeaderCommit() == 1);
279 MockPayload payload1 = sendPayloadData(leaderActor, "one");
281 // Wait for the isolated leader to send AppendEntries to the followers with the new entry with index 1. This
282 // message is forwarded to the followers.
284 expectFirstMatching(follower1CollectorActor, AppendEntries.class, ae ->
285 ae.getEntries().size() == 1 && ae.getEntries().get(0).getIndex() == 1
286 && ae.getEntries().get(0).getData().equals(payload1));
288 expectFirstMatching(follower2CollectorActor, AppendEntries.class, ae ->
289 ae.getEntries().size() == 1 && ae.getEntries().get(0).getIndex() == 1
290 && ae.getEntries().get(0).getData().equals(payload1));
292 verifyApplyJournalEntries(leaderCollectorActor, 1);
296 // Send 3 payloads to the isolated leader so it has uncommitted log entries.
298 testLog.info("Sending 3 payloads to isolated leader");
300 sendPayloadData(leaderActor, "two");
301 sendPayloadData(leaderActor, "three");
302 sendPayloadData(leaderActor, "four");
304 // Wait for the isolated leader to send AppendEntries to follower1 for each new entry. Note the messages
305 // are collected but not forwarded to the follower RaftActor.
307 expectFirstMatching(follower1CollectorActor, AppendEntries.class, ae -> {
308 for (ReplicatedLogEntry e: ae.getEntries()) {
309 if (e.getIndex() == 4) {
316 // The leader should transition to IsolatedLeader.
318 expectFirstMatching(leaderNotifierActor, RoleChanged.class,
319 rc -> rc.getNewRole().equals(RaftState.IsolatedLeader.name()));
321 forceElectionOnFollower1();
323 // Send 3 payloads to the new leader follower1 and verify they're replicated to follower2 and committed. Since
324 // the entry with index 1 from the previous term was uncommitted, the new leader should've also committed a
325 // NoopPayload entry with index 2 in the PreLeader state. Thus the new payload indices will start at 3.
327 testLog.info("Sending 3 payloads to new leader");
329 final MockPayload newLeaderPayload2 = sendPayloadData(follower1Actor, "two-new");
330 final MockPayload newLeaderPayload3 = sendPayloadData(follower1Actor, "three-new");
331 final MockPayload newLeaderPayload4 = sendPayloadData(follower1Actor, "four-new");
332 verifyApplyJournalEntries(follower1CollectorActor, 5);
333 verifyApplyJournalEntries(follower2CollectorActor, 5);
335 assertEquals("Follower 1 journal last term", currentTerm, follower1Context.getReplicatedLog().lastTerm());
336 assertEquals("Follower 1 journal last index", 5, follower1Context.getReplicatedLog().lastIndex());
337 assertEquals("Follower 1 commit index", 5, follower1Context.getCommitIndex());
338 verifyReplicatedLogEntry(follower1Context.getReplicatedLog().get(5), currentTerm, 5, newLeaderPayload4);
340 assertEquals("Follower 1 state", Lists.newArrayList(payload0, payload1, newLeaderPayload2, newLeaderPayload3,
341 newLeaderPayload4), follower1Actor.underlyingActor().getState());
345 // Previous leader should switch to follower b/c it will receive either an AppendEntries or AppendEntriesReply
346 // with a higher term.
348 expectFirstMatching(leaderNotifierActor, RoleChanged.class,
349 rc -> rc.getNewRole().equals(RaftState.Follower.name()));
351 // The previous leader has conflicting log entries starting at index 2 with different terms which should get
352 // replaced by the new leader's entries.
354 verifyApplyJournalEntries(leaderCollectorActor, 5);
356 verifyRaftState(leaderActor, raftState -> {
357 assertEquals("Prior leader journal last term", currentTerm, leaderContext.getReplicatedLog().lastTerm());
358 assertEquals("Prior leader journal last index", 5, leaderContext.getReplicatedLog().lastIndex());
359 assertEquals("Prior leader commit index", 5, leaderContext.getCommitIndex());
362 assertEquals("Prior leader state", Lists.newArrayList(payload0, payload1, newLeaderPayload2, newLeaderPayload3,
363 newLeaderPayload4), leaderActor.underlyingActor().getState());
365 // Ensure the prior leader didn't apply any of its conflicting entries with term 1.
367 List<ApplyState> applyState = getAllMatching(leaderCollectorActor, ApplyState.class);
368 for (ApplyState as: applyState) {
369 if (as.getReplicatedLogEntry().getTerm() == 1) {
370 fail("Got unexpected ApplyState: " + as);
374 // The prior leader should not have needed a snapshot installed in order to get it synced.
376 assertNoneMatching(leaderCollectorActor, InstallSnapshot.class);
378 testLog.info("testLeaderIsolationWithPriorUncommittedEntryAndMultipleConflictingEntries ending");
381 private void removeIsolation() {
382 testLog.info("Removing isolation");
384 clearMessages(leaderNotifierActor);
385 clearMessages(leaderCollectorActor);
387 leaderActor.underlyingActor().stopDropMessages(AppendEntries.class);
388 leaderActor.underlyingActor().stopDropMessages(RequestVote.class);
389 follower1Actor.underlyingActor().stopDropMessages(AppendEntries.class);
390 follower2Actor.underlyingActor().stopDropMessages(AppendEntries.class);
393 private void forceElectionOnFollower1() {
394 // Force follower1 to start an election. follower2 should grant the vote.
396 testLog.info("Forcing election on {}", follower1Id);
398 follower1Actor.tell(TimeoutNow.INSTANCE, ActorRef.noSender());
400 expectFirstMatching(follower1NotifierActor, RoleChanged.class,
401 rc -> rc.getNewRole().equals(RaftState.Leader.name()));
403 currentTerm = follower1Context.getTermInformation().getCurrentTerm();
406 private void isolateLeader() {
407 // Isolate the leader by dropping AppendEntries to the followers and incoming messages from the followers.
409 testLog.info("Isolating the leader");
411 leaderActor.underlyingActor().startDropMessages(AppendEntries.class);
412 leaderActor.underlyingActor().startDropMessages(RequestVote.class);
414 follower1Actor.underlyingActor().startDropMessages(AppendEntries.class,
415 ae -> ae.getLeaderId().equals(leaderId));
416 follower2Actor.underlyingActor().startDropMessages(AppendEntries.class,
417 ae -> ae.getLeaderId().equals(leaderId));
419 clearMessages(follower1CollectorActor);
420 clearMessages(follower1NotifierActor);
421 clearMessages(leaderNotifierActor);
424 private void createRaftActors() {
425 testLog.info("createRaftActors starting");
427 follower1NotifierActor = factory.createTestActor(Props.create(MessageCollectorActor.class),
428 factory.generateActorId(follower1Id + "-notifier"));
430 DefaultConfigParamsImpl followerConfigParams = new DefaultConfigParamsImpl();
431 followerConfigParams.setHeartBeatInterval(new FiniteDuration(100, TimeUnit.MILLISECONDS));
432 followerConfigParams.setElectionTimeoutFactor(1000);
433 follower1Actor = newTestRaftActor(follower1Id, TestRaftActor.newBuilder().peerAddresses(
434 ImmutableMap.of(leaderId, testActorPath(leaderId), follower2Id, testActorPath(follower2Id)))
435 .config(followerConfigParams).roleChangeNotifier(follower1NotifierActor));
437 follower2Actor = newTestRaftActor(follower2Id, ImmutableMap.of(leaderId, testActorPath(leaderId),
438 follower1Id, testActorPath(follower1Id)), followerConfigParams);
440 peerAddresses = ImmutableMap.<String, String>builder()
441 .put(follower1Id, follower1Actor.path().toString())
442 .put(follower2Id, follower2Actor.path().toString()).build();
444 leaderConfigParams = newLeaderConfigParams();
445 leaderConfigParams.setIsolatedLeaderCheckInterval(new FiniteDuration(500, TimeUnit.MILLISECONDS));
447 leaderNotifierActor = factory.createTestActor(Props.create(MessageCollectorActor.class),
448 factory.generateActorId(leaderId + "-notifier"));
450 leaderActor = newTestRaftActor(leaderId, TestRaftActor.newBuilder().peerAddresses(peerAddresses)
451 .config(leaderConfigParams).roleChangeNotifier(leaderNotifierActor));
453 follower1CollectorActor = follower1Actor.underlyingActor().collectorActor();
454 follower2CollectorActor = follower2Actor.underlyingActor().collectorActor();
455 leaderCollectorActor = leaderActor.underlyingActor().collectorActor();
457 leaderActor.tell(TimeoutNow.INSTANCE, ActorRef.noSender());
458 waitUntilLeader(leaderActor);
460 expectMatching(leaderCollectorActor, AppendEntriesReply.class, 2);
463 clearMessages(leaderCollectorActor);
464 clearMessages(follower1CollectorActor);
465 clearMessages(follower2CollectorActor);
467 leaderContext = leaderActor.underlyingActor().getRaftActorContext();
468 currentTerm = leaderContext.getTermInformation().getCurrentTerm();
470 follower1Context = follower1Actor.underlyingActor().getRaftActorContext();
471 follower2Context = follower2Actor.underlyingActor().getRaftActorContext();
473 testLog.info("createRaftActors ending");