SzyWilliam commented on code in PR #15014: URL: https://github.com/apache/iotdb/pull/15014#discussion_r2004663213
########## iotdb-core/confignode/src/main/java/org/apache/iotdb/confignode/manager/load/service/TopologyService.java: ########## @@ -0,0 +1,275 @@ +/* + * 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.iotdb.confignode.manager.load.service; + +import org.apache.iotdb.common.rpc.thrift.TDataNodeConfiguration; +import org.apache.iotdb.common.rpc.thrift.TDataNodeLocation; +import org.apache.iotdb.common.rpc.thrift.TNodeLocations; +import org.apache.iotdb.common.rpc.thrift.TTestConnectionResp; +import org.apache.iotdb.common.rpc.thrift.TTestConnectionResult; +import org.apache.iotdb.commons.cluster.NodeStatus; +import org.apache.iotdb.commons.concurrent.IoTDBThreadPoolFactory; +import org.apache.iotdb.commons.concurrent.ThreadName; +import org.apache.iotdb.confignode.client.async.CnToDnAsyncRequestType; +import org.apache.iotdb.confignode.client.async.CnToDnInternalServiceAsyncRequestManager; +import org.apache.iotdb.confignode.client.async.handlers.DataNodeAsyncRequestContext; +import org.apache.iotdb.confignode.conf.ConfigNodeConfig; +import org.apache.iotdb.confignode.conf.ConfigNodeDescriptor; +import org.apache.iotdb.confignode.manager.IManager; +import org.apache.iotdb.confignode.manager.load.cache.AbstractHeartbeatSample; +import org.apache.iotdb.confignode.manager.load.cache.IFailureDetector; +import org.apache.iotdb.confignode.manager.load.cache.detector.FixedDetector; +import org.apache.iotdb.confignode.manager.load.cache.detector.PhiAccrualDetector; +import org.apache.iotdb.confignode.manager.load.cache.node.NodeHeartbeatSample; +import org.apache.iotdb.confignode.manager.load.cache.node.NodeStatistics; +import org.apache.iotdb.confignode.manager.load.subscriber.IClusterStatusSubscriber; +import org.apache.iotdb.confignode.manager.load.subscriber.NodeStatisticsChangeEvent; + +import org.apache.ratis.util.AwaitForSignal; +import org.apache.tsfile.utils.Pair; +import org.slf4j.Logger; +import org.slf4j.LoggerFactory; + +import java.util.ArrayList; +import java.util.Collections; +import java.util.HashMap; +import java.util.HashSet; +import java.util.LinkedList; +import java.util.List; +import java.util.Map; +import java.util.Objects; +import java.util.Optional; +import java.util.Set; +import java.util.concurrent.CopyOnWriteArrayList; +import java.util.concurrent.ExecutorService; +import java.util.concurrent.TimeUnit; +import java.util.concurrent.atomic.AtomicBoolean; +import java.util.function.Consumer; +import java.util.stream.Collectors; + +public class TopologyService implements Runnable, IClusterStatusSubscriber { + private static final Logger LOGGER = LoggerFactory.getLogger(TopologyService.class); + private static final long PROBING_INTERVAL_MS = 5_000L; + private static final long PROBING_TIMEOUT_MS = PROBING_INTERVAL_MS; + private static final int SAMPLING_WINDOW_SIZE = 100; + + private final ExecutorService topologyThread = + IoTDBThreadPoolFactory.newSingleThreadExecutor( + ThreadName.CONFIG_NODE_TOPOLOGY_SERVICE.getName()); + private final Consumer<Map<Integer, Set<Integer>>> topologyChangeListener; + + private final AwaitForSignal awaitForSignal; + private final IManager configManager; + + private final AtomicBoolean shouldRun; + + /* (fromDataNodeId, toDataNodeId) -> heartbeat history */ + private final Map<Pair<Integer, Integer>, List<AbstractHeartbeatSample>> heartbeats; + private final List<Integer> startingDataNodes = new CopyOnWriteArrayList<>(); + + private final IFailureDetector failureDetector; + private static final ConfigNodeConfig CONF = ConfigNodeDescriptor.getInstance().getConf(); + + public TopologyService( + IManager configManager, Consumer<Map<Integer, Set<Integer>>> topologyChangeListener) { + this.configManager = configManager; + this.topologyChangeListener = topologyChangeListener; + this.heartbeats = new HashMap<>(); + this.shouldRun = new AtomicBoolean(false); + this.awaitForSignal = new AwaitForSignal(this.getClass().getSimpleName()); + + // here we use the same failure + switch (CONF.getFailureDetector()) { + case IFailureDetector.PHI_ACCRUAL_DETECTOR: + this.failureDetector = + new PhiAccrualDetector( + CONF.getFailureDetectorPhiThreshold(), + CONF.getFailureDetectorPhiAcceptablePauseInMs() * 1000_000L, + CONF.getHeartbeatIntervalInMs() * 200_000L, + IFailureDetector.PHI_COLD_START_THRESHOLD, + new FixedDetector(CONF.getFailureDetectorFixedThresholdInMs() * 1000_000L)); + break; + case IFailureDetector.FIXED_DETECTOR: + default: + this.failureDetector = + new FixedDetector(CONF.getFailureDetectorFixedThresholdInMs() * 1000_000L); + } + } + + public synchronized void startTopologyService() { + shouldRun.set(true); + topologyThread.submit(this); + LOGGER.info("Topology Probing has started successfully"); + } + + public synchronized void stopTopologyService() { + shouldRun.set(false); + topologyThread.shutdown(); + try { + topologyThread.awaitTermination(PROBING_INTERVAL_MS, TimeUnit.MILLISECONDS); + } catch (InterruptedException e) { + Thread.currentThread().interrupt(); + } + LOGGER.info("Topology Probing has stopped successfully"); + } + + /** + * Schedule the {@link #topologyProbing} task either: 1. every PROBING_INTERVAL_MS interval. 2. + * Manually triggered by outside events (node restart / register, etc.). + */ + private void mayWait() { + try { + this.awaitForSignal.await(PROBING_INTERVAL_MS, TimeUnit.MILLISECONDS); + } catch (InterruptedException e) { + Thread.currentThread().interrupt(); + } + } + + @Override + public void run() { + for (; shouldRun.get(); mayWait()) { + topologyProbing(); + } + } + + private void topologyProbing() { + // 1. get the latest datanode list + final List<TDataNodeLocation> dataNodeLocations = new ArrayList<>(); + final Set<Integer> dataNodeIds = new HashSet<>(); + for (final TDataNodeConfiguration dataNodeConf : + configManager.getNodeManager().getRegisteredDataNodes()) { + final TDataNodeLocation location = dataNodeConf.getLocation(); + if (startingDataNodes.contains(location.getDataNodeId())) { + continue; // we shall wait for internal endpoint to be ready + } + dataNodeLocations.add(location); + dataNodeIds.add(location.getDataNodeId()); + } + + // 2. send the verify connection RPC to all datanode + final TNodeLocations nodeLocations = new TNodeLocations(); + nodeLocations.setDataNodeLocations(dataNodeLocations); + nodeLocations.setConfigNodeLocations(Collections.emptyList()); + final Map<Integer, TDataNodeLocation> dataNodeLocationMap = + configManager.getNodeManager().getRegisteredDataNodes().stream() + .map(TDataNodeConfiguration::getLocation) + .collect(Collectors.toMap(TDataNodeLocation::getDataNodeId, location -> location)); + final DataNodeAsyncRequestContext<TNodeLocations, TTestConnectionResp> + dataNodeAsyncRequestContext = + new DataNodeAsyncRequestContext<>( + CnToDnAsyncRequestType.SUBMIT_TEST_DN_INTERNAL_CONNECTION_TASK, + nodeLocations, + dataNodeLocationMap); + CnToDnInternalServiceAsyncRequestManager.getInstance() + .sendAsyncRequestWithTimeoutInMs(dataNodeAsyncRequestContext, PROBING_TIMEOUT_MS); + final List<TTestConnectionResult> results = new ArrayList<>(); + dataNodeAsyncRequestContext + .getResponseMap() + .forEach( + (nodeId, resp) -> { + if (resp.isSetResultList()) { + results.addAll(resp.getResultList()); + } + }); + + // 3. collect results and update the heartbeat timestamps + for (final TTestConnectionResult result : results) { + final int fromDataNodeId = + Optional.ofNullable(result.getSender().getDataNodeLocation()) + .map(TDataNodeLocation::getDataNodeId) + .orElse(-1); + final int toDataNodeId = result.getServiceProvider().getNodeId(); + if (result.isSuccess() + && dataNodeIds.contains(fromDataNodeId) + && dataNodeIds.contains(toDataNodeId)) { + // testAllDataNodeConnectionWithTimeout ensures the heartbeats are Dn-Dn internally. Here we + // just double-check. + final List<AbstractHeartbeatSample> heartbeatHistory = + heartbeats.computeIfAbsent( + new Pair<>(fromDataNodeId, toDataNodeId), p -> new LinkedList<>()); + heartbeatHistory.add(new NodeHeartbeatSample(NodeStatus.Running)); + if (heartbeatHistory.size() > SAMPLING_WINDOW_SIZE) { + heartbeatHistory.remove(0); + } + } + } + + // 4. use failure detector to identify potential network partitions + final Map<Integer, Set<Integer>> latestTopology = + dataNodeLocations.stream() + .collect(Collectors.toMap(TDataNodeLocation::getDataNodeId, k -> new HashSet<>())); + for (final Map.Entry<Pair<Integer, Integer>, List<AbstractHeartbeatSample>> entry : + heartbeats.entrySet()) { + final int fromId = entry.getKey().getLeft(); + final int toId = entry.getKey().getRight(); + if (!entry.getValue().isEmpty() + && !failureDetector.isAvailable(entry.getKey(), entry.getValue())) { + LOGGER.debug("Connection from DataNode {} to DataNode {} is broken", fromId, toId); + } else { + latestTopology.get(fromId).add(toId); + } + } + + // 5. notify the listeners on topology change + topologyChangeListener.accept(latestTopology); + } + + /** We only listen to datanode remove / restart / register events */ + @Override + public void onNodeStatisticsChanged(NodeStatisticsChangeEvent event) { + final Set<Integer> datanodeIds = + configManager.getNodeManager().getRegisteredDataNodeLocations().keySet(); + final Map<Integer, Pair<NodeStatistics, NodeStatistics>> changes = + event.getDifferentNodeStatisticsMap(); + for (final Map.Entry<Integer, Pair<NodeStatistics, NodeStatistics>> entry : + changes.entrySet()) { + final Integer nodeId = entry.getKey(); + final Pair<NodeStatistics, NodeStatistics> changeEvent = entry.getValue(); + if (!datanodeIds.contains(nodeId)) { + continue; + } + if (changeEvent.getLeft() == null) { + // if a new datanode registered, DO NOT trigger probing immediately + startingDataNodes.add(nodeId); + continue; + } else { + startingDataNodes.remove(nodeId); + } + + if (changeEvent.getRight() == null) { + // datanode removed from cluster, clean up probing history + final Set<Pair<Integer, Integer>> toRemove = + heartbeats.keySet().stream() + .filter( + pair -> + Objects.equals(pair.getLeft(), nodeId) + || Objects.equals(pair.getRight(), nodeId)) + .collect(Collectors.toSet()); + toRemove.forEach(heartbeats::remove); Review Comment: I‘ll use `ConcurrentHashMap` then -- 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]
