[Remove] Deprecated Zen1 Discovery (#1216)

Zen1 discovery was deprecated in Legacy 7.x for eventual removal. OpenSearch 1.x
carries this deprecation. This commit completely removes all support for Zen1
discovery in favor of Zen2.

Signed-off-by: Nicholas Walter Knize <nknize@apache.org>

server/src/internalClusterTest/java/org/opensearch/cluster/coordination/UnsafeBootstrapAndDetachCommandIT.java

@@ -39,13 +39,13 @@
 import org.opensearch.cluster.metadata.IndexMetadata;
 import org.opensearch.cluster.metadata.Metadata;
 import org.opensearch.common.settings.Settings;
-import org.opensearch.discovery.DiscoverySettings;
 import org.opensearch.env.Environment;
 import org.opensearch.env.NodeEnvironment;
 import org.opensearch.env.TestEnvironment;
 import org.opensearch.gateway.GatewayMetaState;
 import org.opensearch.gateway.PersistedClusterStateService;
 import org.opensearch.indices.IndicesService;
+import org.opensearch.node.Node.DiscoverySettings;
 import org.opensearch.test.InternalTestCluster;
 import org.opensearch.test.OpenSearchIntegTestCase;
 

server/src/internalClusterTest/java/org/opensearch/cluster/coordination/ZenDiscoveryIT.java

@@ -47,7 +47,6 @@
 import org.opensearch.common.xcontent.XContentFactory;
 import org.opensearch.discovery.Discovery;
 import org.opensearch.discovery.DiscoveryStats;
-import org.opensearch.discovery.zen.FaultDetection;
 import org.opensearch.test.OpenSearchIntegTestCase;
 import org.opensearch.test.TestCustomMetadata;
 import org.opensearch.transport.RemoteTransportException;
@@ -73,10 +72,6 @@
 public class ZenDiscoveryIT extends OpenSearchIntegTestCase {
 
     public void testNoShardRelocationsOccurWhenElectedMasterNodeFails() throws Exception {
-        Settings defaultSettings = Settings.builder()
-            .put(FaultDetection.PING_TIMEOUT_SETTING.getKey(), "1s")
-            .put(FaultDetection.PING_RETRIES_SETTING.getKey(), "1")
-            .build();
 
         Settings masterNodeSettings = masterOnlyNode();
         internalCluster().startNodes(2, masterNodeSettings);

server/src/internalClusterTest/java/org/opensearch/discovery/DiscoveryDisruptionIT.java

@@ -39,9 +39,6 @@
 import org.opensearch.cluster.node.DiscoveryNodes;
 import org.opensearch.cluster.service.ClusterService;
 import org.opensearch.common.settings.Settings;
-import org.opensearch.discovery.zen.MembershipAction;
-import org.opensearch.discovery.zen.PublishClusterStateAction;
-import org.opensearch.discovery.zen.ZenDiscovery;
 import org.opensearch.test.OpenSearchIntegTestCase;
 import org.opensearch.test.disruption.NetworkDisruption;
 import org.opensearch.test.disruption.ServiceDisruptionScheme;
@@ -56,7 +53,6 @@
 
 import static org.opensearch.cluster.metadata.IndexMetadata.INDEX_NUMBER_OF_REPLICAS_SETTING;
 import static org.opensearch.cluster.metadata.IndexMetadata.INDEX_NUMBER_OF_SHARDS_SETTING;
-import static org.hamcrest.Matchers.greaterThanOrEqualTo;
 
 /**
  * Tests for discovery during disruptions.
@@ -97,23 +93,15 @@ public void testClusterJoinDespiteOfPublishingIssues() throws Exception {
             discoveryNodes.getLocalNode().getName()
         );
         if (randomBoolean()) {
-            masterTransportService.addFailToSendNoConnectRule(
-                localTransportService,
-                PublishClusterStateAction.SEND_ACTION_NAME,
-                PublicationTransportHandler.PUBLISH_STATE_ACTION_NAME
-            );
+            masterTransportService.addFailToSendNoConnectRule(localTransportService, PublicationTransportHandler.PUBLISH_STATE_ACTION_NAME);
         } else {
-            masterTransportService.addFailToSendNoConnectRule(
-                localTransportService,
-                PublishClusterStateAction.COMMIT_ACTION_NAME,
-                PublicationTransportHandler.COMMIT_STATE_ACTION_NAME
-            );
+            masterTransportService.addFailToSendNoConnectRule(localTransportService, PublicationTransportHandler.COMMIT_STATE_ACTION_NAME);
         }
 
         logger.info("allowing requests from non master [{}] to master [{}], waiting for two join request", nonMasterNode, masterNode);
         final CountDownLatch countDownLatch = new CountDownLatch(2);
         nonMasterTransportService.addSendBehavior(masterTransportService, (connection, requestId, action, request, options) -> {
-            if (action.equals(MembershipAction.DISCOVERY_JOIN_ACTION_NAME) || action.equals(JoinHelper.JOIN_ACTION_NAME)) {
+            if (action.equals(JoinHelper.JOIN_ACTION_NAME)) {
                 countDownLatch.countDown();
             }
             connection.sendRequest(requestId, action, request, options);
@@ -139,7 +127,7 @@ public void testClusterFormingWithASlowNode() {
         SlowClusterStateProcessing disruption = new SlowClusterStateProcessing(random(), 0, 0, 1000, 2000);
 
         // don't wait for initial state, we want to add the disruption while the cluster is forming
-        internalCluster().startNodes(3, Settings.builder().put(DiscoverySettings.PUBLISH_TIMEOUT_SETTING.getKey(), "3s").build());
+        internalCluster().startNodes(3);
 
         logger.info("applying disruption while cluster is forming ...");
 
@@ -167,14 +155,6 @@ public void testElectMasterWithLatestVersion() throws Exception {
         ensureStableCluster(3);
         final String preferredMasterName = internalCluster().getMasterName();
         final DiscoveryNode preferredMaster = internalCluster().clusterService(preferredMasterName).localNode();
-        final Discovery discovery = internalCluster().getInstance(Discovery.class);
-        // only Zen1 guarantees that node with lowest id is elected
-        if (discovery instanceof ZenDiscovery) {
-            for (String node : nodes) {
-                DiscoveryNode discoveryNode = internalCluster().clusterService(node).localNode();
-                assertThat(discoveryNode.getId(), greaterThanOrEqualTo(preferredMaster.getId()));
-            }
-        }
 
         logger.info("--> preferred master is {}", preferredMaster);
         final Set<String> nonPreferredNodes = new HashSet<>(nodes);

server/src/internalClusterTest/java/org/opensearch/discovery/MasterDisruptionIT.java

@@ -42,7 +42,6 @@
 import org.opensearch.common.settings.Settings;
 import org.opensearch.common.unit.TimeValue;
 import org.opensearch.common.xcontent.XContentType;
-import org.opensearch.discovery.zen.ZenDiscovery;
 import org.opensearch.test.OpenSearchIntegTestCase;
 import org.opensearch.test.disruption.BlockMasterServiceOnMaster;
 import org.opensearch.test.disruption.IntermittentLongGCDisruption;
@@ -55,7 +54,6 @@
 import java.util.HashSet;
 import java.util.List;
 import java.util.Set;
-import java.util.concurrent.TimeUnit;
 
 import static org.opensearch.test.hamcrest.OpenSearchAssertions.assertAcked;
 import static org.hamcrest.Matchers.equalTo;
@@ -92,9 +90,6 @@ public void testMasterNodeGCs() throws Exception {
         logger.info("waiting for nodes to elect a new master");
         ensureStableCluster(2, oldNonMasterNodes.get(0));
 
-        logger.info("waiting for any pinging to stop");
-        assertDiscoveryCompleted(oldNonMasterNodes);
-
         // restore GC
         masterNodeDisruption.stopDisrupting();
         final TimeValue waitTime = new TimeValue(DISRUPTION_HEALING_OVERHEAD.millis() + masterNodeDisruption.expectedTimeToHeal().millis());
@@ -323,15 +318,4 @@ public void testMappingTimeout() throws Exception {
         });
 
     }
-
-    private void assertDiscoveryCompleted(List<String> nodes) throws Exception {
-        for (final String node : nodes) {
-            assertBusy(() -> {
-                final Discovery discovery = internalCluster().getInstance(Discovery.class, node);
-                if (discovery instanceof ZenDiscovery) {
-                    assertFalse("node [" + node + "] is still joining master", ((ZenDiscovery) discovery).joiningCluster());
-                }
-            }, 30, TimeUnit.SECONDS);
-        }
-    }
 }

server/src/internalClusterTest/java/org/opensearch/discovery/single/SingleNodeDiscoveryIT.java

@@ -41,7 +41,7 @@
 import org.opensearch.cluster.service.ClusterService;
 import org.opensearch.common.logging.Loggers;
 import org.opensearch.common.settings.Settings;
-import org.opensearch.discovery.DiscoverySettings;
+import org.opensearch.node.Node.DiscoverySettings;
 import org.opensearch.test.OpenSearchIntegTestCase;
 import org.opensearch.test.InternalTestCluster;
 import org.opensearch.test.MockHttpTransport;

server/src/main/java/org/opensearch/action/admin/cluster/state/TransportClusterStateAction.java

@@ -166,7 +166,6 @@ private ClusterStateResponse buildResponse(final ClusterStateRequest request, fi
         ClusterState.Builder builder = ClusterState.builder(currentState.getClusterName());
         builder.version(currentState.version());
         builder.stateUUID(currentState.stateUUID());
-        builder.minimumMasterNodesOnPublishingMaster(currentState.getMinimumMasterNodesOnPublishingMaster());
 
         if (request.nodes()) {
             builder.nodes(currentState.nodes());

server/src/main/java/org/opensearch/bootstrap/BootstrapChecks.java

@@ -762,11 +762,6 @@ boolean isAllPermissionGranted() {
     static class DiscoveryConfiguredCheck implements BootstrapCheck {
         @Override
         public BootstrapCheckResult check(BootstrapContext context) {
-            if (DiscoveryModule.ZEN_DISCOVERY_TYPE.equals(DiscoveryModule.DISCOVERY_TYPE_SETTING.get(context.settings()))) {
-                return BootstrapCheckResult.failure(
-                    String.format(Locale.ROOT, "discovery type [%s] is unsuitable for production use", DiscoveryModule.ZEN_DISCOVERY_TYPE)
-                );
-            }
 
             if (DiscoveryModule.ZEN2_DISCOVERY_TYPE.equals(DiscoveryModule.DISCOVERY_TYPE_SETTING.get(context.settings())) == false) {
                 return BootstrapCheckResult.success();

server/src/main/java/org/opensearch/cluster/ClusterState.java

@@ -323,23 +323,6 @@ public Set<VotingConfigExclusion> getVotingConfigExclusions() {
         return coordinationMetadata().getVotingConfigExclusions();
     }
 
-    /**
-     * The node-level `discovery.zen.minimum_master_nodes` setting on the master node that published this cluster state, for use in rolling
-     * upgrades from 6.x to 7.x. Once all the 6.x master-eligible nodes have left the cluster, the 7.x nodes use this value to determine how
-     * many master-eligible nodes must be discovered before the cluster can be bootstrapped. Note that this method returns the node-level
-     * value of this setting, and ignores any cluster-level override that was set via the API. Callers are expected to combine this value
-     * with any value set in the cluster-level settings. This should be removed once we no longer need support for
-     * {@link LegacyESVersion#V_6_7_0}.
-     */
-    public int getMinimumMasterNodesOnPublishingMaster() {
-        return minimumMasterNodesOnPublishingMaster;
-    }
-
-    // Used for testing and logging to determine how this cluster state was send over the wire
-    public boolean wasReadFromDiff() {
-        return wasReadFromDiff;
-    }
-
     /**
      * Returns a built (on demand) routing nodes view of the routing table.
      */

server/src/main/java/org/opensearch/cluster/coordination/Coordinator.java

@@ -82,7 +82,6 @@
 import org.opensearch.discovery.PeerFinder;
 import org.opensearch.discovery.SeedHostsProvider;
 import org.opensearch.discovery.SeedHostsResolver;
-import org.opensearch.discovery.zen.PendingClusterStateStats;
 import org.opensearch.monitor.NodeHealthService;
 import org.opensearch.monitor.StatusInfo;
 import org.opensearch.threadpool.Scheduler;
@@ -166,7 +165,6 @@ public class Coordinator extends AbstractLifecycleComponent implements Discovery
     private long maxTermSeen;
     private final Reconfigurator reconfigurator;
     private final ClusterBootstrapService clusterBootstrapService;
-    private final DiscoveryUpgradeService discoveryUpgradeService;
     private final LagDetector lagDetector;
     private final ClusterFormationFailureHelper clusterFormationFailureHelper;
 
@@ -266,14 +264,6 @@ public Coordinator(
             this::isInitialConfigurationSet,
             this::setInitialConfiguration
         );
-        this.discoveryUpgradeService = new DiscoveryUpgradeService(
-            settings,
-            transportService,
-            this::isInitialConfigurationSet,
-            joinHelper,
-            peerFinder::getFoundPeers,
-            this::setInitialConfiguration
-        );
         this.lagDetector = new LagDetector(
             settings,
             transportService.getThreadPool(),
@@ -670,10 +660,6 @@ void becomeCandidate(String method) {
             peerFinder.activate(coordinationState.get().getLastAcceptedState().nodes());
             clusterFormationFailureHelper.start();
 
-            if (getCurrentTerm() == ZEN1_BWC_TERM) {
-                discoveryUpgradeService.activate(lastKnownLeader, coordinationState.get().getLastAcceptedState());
-            }
-
             leaderChecker.setCurrentNodes(DiscoveryNodes.EMPTY_NODES);
             leaderChecker.updateLeader(null);
 
@@ -713,7 +699,6 @@ void becomeLeader(String method) {
 
         lastKnownLeader = Optional.of(getLocalNode());
         peerFinder.deactivate(getLocalNode());
-        discoveryUpgradeService.deactivate();
         clusterFormationFailureHelper.stop();
         closePrevotingAndElectionScheduler();
         preVoteCollector.update(getPreVoteResponse(), getLocalNode());
@@ -751,7 +736,6 @@ void becomeFollower(String method, DiscoveryNode leaderNode) {
 
         lastKnownLeader = Optional.of(leaderNode);
         peerFinder.deactivate(leaderNode);
-        discoveryUpgradeService.deactivate();
         clusterFormationFailureHelper.stop();
         closePrevotingAndElectionScheduler();
         cancelActivePublication("become follower: " + method);