SERVER-45081 Add a unit test case for awaitable isMaster after stepup and config term increment.

1 files changed, 108 insertions, 0 deletions

diff --git a/src/mongo/db/repl/replication_coordinator_impl_test.cpp b/src/mongo/db/repl/replication_coordinator_impl_test.cpp
index 5e0f6574ce9..3eed301094a 100644
--- a/src/mongo/db/repl/replication_coordinator_impl_test.cpp
+++ b/src/mongo/db/repl/replication_coordinator_impl_test.cpp
@@ -3877,6 +3877,114 @@ TEST_F(ReplCoordTest, AwaitIsMasterResponseReturnsOnElectionWin) {
     getIsMasterThread.join();
 }
 
+TEST_F(ReplCoordTest, AwaitIsMasterResponseReturnsOnElectionWinWithReconfig) {
+    // The config has a "term" field, so step-up will increment the config term
+    // via reconfig. As a result, step-up triggers three topology changes.
+    init();
+    assertStartSuccess(BSON("_id"
+                            << "mySet"
+                            << "version" << 2 << "term" << 0 << "members"
+                            << BSON_ARRAY(BSON("host"
+                                               << "node1:12345"
+                                               << "_id" << 0)
+                                          << BSON("host"
+                                                  << "node2:12345"
+                                                  << "_id" << 1))),
+                       HostAndPort("node1", 12345));
+
+    // Become secondary.
+    ASSERT_OK(getReplCoord()->setFollowerMode(MemberState::RS_SECONDARY));
+    replCoordSetMyLastAppliedOpTime(OpTimeWithTermOne(100, 1), Date_t() + Seconds(100));
+    replCoordSetMyLastDurableOpTime(OpTimeWithTermOne(100, 1), Date_t() + Seconds(100));
+    ASSERT(getReplCoord()->getMemberState().secondary());
+
+    auto maxAwaitTime = Milliseconds(50000);
+    auto deadline = getNet()->now() + maxAwaitTime;
+
+    auto currentTopologyVersion = getTopoCoord().getTopologyVersion();
+    auto expectedProcessId = currentTopologyVersion.getProcessId();
+    // A topology change should increment the TopologyVersion counter.
+    auto expectedCounter = currentTopologyVersion.getCounter() + 1;
+
+    auto opCtx = makeOperationContext();
+    // Calling isMaster without a TopologyVersion field should return immediately.
+    const auto response =
+        getReplCoord()->awaitIsMasterResponse(opCtx.get(), {}, boost::none, boost::none);
+    ASSERT_FALSE(response->isMaster());
+    ASSERT_TRUE(response->isSecondary());
+    ASSERT_FALSE(response->hasPrimary());
+
+    auto waitForIsMasterFailPoint = globalFailPointRegistry().find("waitForIsMasterResponse");
+    auto hangAfterReconfigFailPoint =
+        globalFailPointRegistry().find("hangAfterReconfigOnDrainComplete");
+
+    auto timesEnteredFailPoint = waitForIsMasterFailPoint->setMode(FailPoint::alwaysOn);
+
+    // awaitIsMasterResponse blocks and waits on a future when the request TopologyVersion equals
+    // the current TopologyVersion of the server.
+    stdx::thread getIsMasterThread([&] {
+        const auto responseAfterElection = getReplCoord()->awaitIsMasterResponse(
+            opCtx.get(), {}, currentTopologyVersion, deadline);
+
+        const auto topologyVersionAfterElection = responseAfterElection->getTopologyVersion();
+        ASSERT_EQUALS(topologyVersionAfterElection->getCounter(), expectedCounter);
+        ASSERT_EQUALS(topologyVersionAfterElection->getProcessId(), expectedProcessId);
+
+        // We expect the server to increment the TopologyVersion and respond to waiting IsMasters
+        // once an election is won even if we have yet to signal drain completion.
+        ASSERT_FALSE(responseAfterElection->isMaster());
+        ASSERT_TRUE(responseAfterElection->isSecondary());
+        ASSERT_TRUE(responseAfterElection->hasPrimary());
+        ASSERT_EQUALS(responseAfterElection->getPrimary().host(), "node1");
+        ASSERT(getReplCoord()->getMemberState().primary());
+
+        // The server TopologyVersion will increment once we finish reconfig.
+        expectedCounter = topologyVersionAfterElection->getCounter() + 1;
+        const auto responseAfterReconfig = getReplCoord()->awaitIsMasterResponse(
+            opCtx.get(), {}, topologyVersionAfterElection, deadline);
+        const auto topologyVersionAfterReconfig = responseAfterReconfig->getTopologyVersion();
+        ASSERT_EQUALS(topologyVersionAfterReconfig->getCounter(), expectedCounter);
+        ASSERT_EQUALS(topologyVersionAfterReconfig->getProcessId(), expectedProcessId);
+
+        ASSERT_FALSE(responseAfterReconfig->isMaster());
+        ASSERT_TRUE(responseAfterReconfig->isSecondary());
+        ASSERT_TRUE(responseAfterReconfig->hasPrimary());
+        ASSERT_EQUALS(responseAfterReconfig->getPrimary().host(), "node1");
+        ASSERT(getReplCoord()->getMemberState().primary());
+
+        hangAfterReconfigFailPoint->setMode(FailPoint::off);
+        // The server TopologyVersion will increment again once we exit drain mode.
+        expectedCounter = topologyVersionAfterReconfig->getCounter() + 1;
+        const auto responseAfterDrainComplete = getReplCoord()->awaitIsMasterResponse(
+            opCtx.get(), {}, topologyVersionAfterReconfig, deadline);
+        const auto topologyVersionAfterDrainComplete =
+            responseAfterDrainComplete->getTopologyVersion();
+        ASSERT_EQUALS(topologyVersionAfterDrainComplete->getCounter(), expectedCounter);
+        ASSERT_EQUALS(topologyVersionAfterDrainComplete->getProcessId(), expectedProcessId);
+
+        ASSERT_TRUE(responseAfterDrainComplete->isMaster());
+        ASSERT_FALSE(responseAfterDrainComplete->isSecondary());
+        ASSERT_TRUE(responseAfterDrainComplete->hasPrimary());
+        ASSERT_EQUALS(responseAfterDrainComplete->getPrimary().host(), "node1");
+        ASSERT(getReplCoord()->getMemberState().primary());
+    });
+
+    // Ensure that awaitIsMasterResponse() is called before finishing the election.
+    waitForIsMasterFailPoint->waitForTimesEntered(timesEnteredFailPoint + 1);
+    auto electionTimeoutWhen = getReplCoord()->getElectionTimeout_forTest();
+    ASSERT_NOT_EQUALS(Date_t(), electionTimeoutWhen);
+    LOGV2(4508104,
+          "Election timeout scheduled at {electionTimeoutWhen} (simulator time)",
+          "electionTimeoutWhen"_attr = electionTimeoutWhen);
+    simulateSuccessfulV1ElectionWithoutExitingDrainMode(electionTimeoutWhen, opCtx.get());
+
+    waitForIsMasterFailPoint->waitForTimesEntered(timesEnteredFailPoint + 2);
+    signalDrainComplete(opCtx.get());
+    ASSERT(getReplCoord()->getApplierState() == ReplicationCoordinator::ApplierState::Stopped);
+
+    getIsMasterThread.join();
+}
+
 TEST_F(ReplCoordTest, IsMasterResponseMentionsLackOfReplicaSetConfig) {
     start();