vldpyatkov commented on code in PR #1648: URL: https://github.com/apache/ignite-3/pull/1648#discussion_r1100617690
########## modules/placement-driver/src/main/java/org/apache/ignite/internal/raft/client/TopologyAwareRaftGroupService.java: ########## @@ -0,0 +1,443 @@ +/* + * Licensed to the Apache Software Foundation (ASF) under one or more + * contributor license agreements. See the NOTICE file distributed with + * this work for additional information regarding copyright ownership. + * The ASF licenses this file to You under the Apache License, Version 2.0 + * (the "License"); you may not use this file except in compliance with + * the License. You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +package org.apache.ignite.internal.raft.client; + +import java.io.IOException; +import java.util.Collection; +import java.util.List; +import java.util.concurrent.CompletableFuture; +import java.util.concurrent.CompletionException; +import java.util.concurrent.ExecutionException; +import java.util.concurrent.ScheduledExecutorService; +import java.util.concurrent.TimeoutException; +import java.util.function.Consumer; +import org.apache.ignite.internal.cluster.management.topology.api.LogicalTopologyEventListener; +import org.apache.ignite.internal.cluster.management.topology.api.LogicalTopologyService; +import org.apache.ignite.internal.cluster.management.topology.api.LogicalTopologySnapshot; +import org.apache.ignite.internal.logger.IgniteLogger; +import org.apache.ignite.internal.logger.Loggers; +import org.apache.ignite.internal.raft.Command; +import org.apache.ignite.internal.raft.Peer; +import org.apache.ignite.internal.raft.PeersAndLearners; +import org.apache.ignite.internal.raft.RaftGroupServiceImpl; +import org.apache.ignite.internal.raft.configuration.RaftConfiguration; +import org.apache.ignite.internal.raft.service.LeaderWithTerm; +import org.apache.ignite.internal.raft.service.RaftGroupService; +import org.apache.ignite.internal.replicator.ReplicationGroupId; +import org.apache.ignite.lang.ErrorGroups.Common; +import org.apache.ignite.lang.IgniteException; +import org.apache.ignite.network.ClusterNode; +import org.apache.ignite.network.ClusterService; +import org.apache.ignite.raft.jraft.RaftMessageGroup; +import org.apache.ignite.raft.jraft.RaftMessagesFactory; +import org.apache.ignite.raft.jraft.rpc.CliRequests.SubscriptionLeaderChangeRequest; +import org.apache.ignite.raft.jraft.rpc.CliRequests.SubscriptionLeaderChangeResponse; +import org.jetbrains.annotations.Nullable; + +/** + * RAFT client aware of a logical topology to handle distributed events. + */ +public class TopologyAwareRaftGroupService implements RaftGroupService { + + /** The logger. */ + private static final IgniteLogger LOG = Loggers.forClass(TopologyAwareRaftGroupService.class); + + /** Raft message factory. */ + private final RaftMessagesFactory factory; + + /** Cluster service. */ + private final ClusterService clusterService; + + /** RPC RAFT client. */ + private final RaftGroupService raftClient; + + /** Logical topology service. */ + private final LogicalTopologyService logicalTopologyService; + + /** Leader election handler. */ + private final ServerEventHandler serverEventHandler; + + /** Executor to invoke RPC requests. */ + private final ScheduledExecutorService executor; + + /** RAFT configuration. */ + private final RaftConfiguration raftConfiguration; + + + /** + * The constructor. + * + * @param cluster Cluster service. + * @param factory Message factory. + * @param executor RPC executor. + * @param raftClient RPC RAFT client. + * @param logicalTopologyService Logical topology. + */ + private TopologyAwareRaftGroupService( + ClusterService cluster, + RaftMessagesFactory factory, + ScheduledExecutorService executor, + RaftConfiguration raftConfiguration, + RaftGroupService raftClient, + LogicalTopologyService logicalTopologyService + ) { + this.clusterService = cluster; + this.factory = factory; + this.executor = executor; + this.raftConfiguration = raftConfiguration; + this.raftClient = raftClient; + this.logicalTopologyService = logicalTopologyService; + this.serverEventHandler = new ServerEventHandler(); + + cluster.messagingService().addMessageHandler(RaftMessageGroup.class, (message, senderConsistentId, correlationId) -> { + if (message instanceof SubscriptionLeaderChangeResponse) { + var msg = (SubscriptionLeaderChangeResponse) message; + + serverEventHandler.onLeaderElected(clusterService.topologyService().getByConsistentId(senderConsistentId), msg.term()); + } + }); + + logicalTopologyService.addEventListener(new LogicalTopologyEventListener() { + @Override + public void onAppeared(ClusterNode appearedNode, LogicalTopologySnapshot newTopology) { + for (Peer peer : peers()) { + if (serverEventHandler.isSubscribed() && appearedNode.name().equals(peer.consistentId())) { + LOG.info("New peer will be sending a leader elected notification [grpId={}, consistentId={}]", groupId(), + peer.consistentId()); + + sendSubscribeMessage(appearedNode, TopologyAwareRaftGroupService.this.factory.subscriptionLeaderChangeRequest() + .groupId(groupId()) + .subscribe(true) + .build()) + .thenComposeAsync(couldLeaderChange -> { + if (couldLeaderChange) { + return refreshAndGetLeaderWithTerm() + .thenAcceptAsync(leaderWithTerm -> { + if (leaderWithTerm.leader() != null + && appearedNode.name().equals(leaderWithTerm.leader().consistentId())) { + serverEventHandler.onLeaderElected(appearedNode, leaderWithTerm.term()); + } + }, executor); + } + + return CompletableFuture.completedFuture(null); + }, executor); + } + } + } + }); + } + + /** + * Starts an instance of topology aware RAFT client. + * + * @param groupId Replication group id. + * @param cluster Cluster service. + * @param factory Message factory. + * @param raftConfiguration RAFT configuration. + * @param configuration Group configuration. + * @param getLeader True to get the group's leader upon service creation. + * @param executor RPC executor. + * @param logicalTopologyService Logical topology service. + * @return Future to create a raft client. + */ + public static CompletableFuture<RaftGroupService> start( + ReplicationGroupId groupId, + ClusterService cluster, + RaftMessagesFactory factory, + RaftConfiguration raftConfiguration, + PeersAndLearners configuration, + boolean getLeader, + ScheduledExecutorService executor, + LogicalTopologyService logicalTopologyService + ) { + return RaftGroupServiceImpl.start(groupId, cluster, factory, raftConfiguration, configuration, getLeader, executor) + .thenApply(raftGroupService -> new TopologyAwareRaftGroupService(cluster, factory, executor, raftConfiguration, + raftGroupService, logicalTopologyService)); + } + + /** + * Sends a subscribe message to a specific node of the cluster. + * + * @param node Node. + * @param msg Subscribe message. + * @return A future that will complete when the message is sent. + */ + private CompletableFuture<Boolean> sendSubscribeMessage(ClusterNode node, SubscriptionLeaderChangeRequest msg) { + var msgSendFut = new CompletableFuture<Boolean>(); + + sendWithRetry(node, msg, msgSendFut); + + return msgSendFut; + } + + /** + * Tries to send a subscribe message until the node leaves of the cluster. + * + * @param node Node. + * @param msg Subscribe message to send. + * @param msgSendFut Future that will completed when the message is send or an issue prevent that. + */ + private void sendWithRetry(ClusterNode node, SubscriptionLeaderChangeRequest msg, CompletableFuture<Boolean> msgSendFut) { + clusterService.messagingService().invoke(node, msg, raftConfiguration.responseTimeout().value()).whenCompleteAsync((unused, th) -> { + if (th != null) { + if (recoverable(th)) { Review Comment: I think the loop will be internally in network service layer in the future. They handle it more correctly, but currently I think the node which is catching an exception due to cannot get a topology has no cause except the particular exception. Moreover, we do not an exception when we retry send a message and node still available in logical topology. Why do we have to write an exception into cause if we cannot get a logical topology?. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
