nacos RaftCore中signalPublish的原理和使用方法

发布时间：2021-06-24 10:33:48 作者：chen
来源：亿速云阅读：192

这篇文章主要讲解了“nacos RaftCore中signalPublish的原理和使用方法”，文中的讲解内容简单清晰，易于学习与理解，下面请大家跟着小编的思路慢慢深入，一起来研究和学习“nacos RaftCore中signalPublish的原理和使用方法”吧！

序

本文主要研究一下nacos RaftCore的signalPublish

RaftCore

nacos-1.1.3/naming/src/main/java/com/alibaba/nacos/naming/consistency/persistent/raft/RaftCore.java

@Component
public class RaftCore {

    public static final String API_VOTE = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/vote";

    public static final String API_BEAT = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/beat";

    public static final String API_PUB = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/datum";

    public static final String API_DEL = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/datum";

    public static final String API_GET = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/datum";

    public static final String API_ON_PUB = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/datum/commit";

    public static final String API_ON_DEL = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/datum/commit";

    public static final String API_GET_PEER = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/peer";

    private ScheduledExecutorService executor = new ScheduledThreadPoolExecutor(1, new ThreadFactory() {
        @Override
        public Thread newThread(Runnable r) {
            Thread t = new Thread(r);

            t.setDaemon(true);
            t.setName("com.alibaba.nacos.naming.raft.notifier");

            return t;
        }
    });

    public static final Lock OPERATE_LOCK = new ReentrantLock();

    public static final int PUBLISH_TERM_INCREASE_COUNT = 100;

    private volatile Map<String, List<RecordListener>> listeners = new ConcurrentHashMap<>();

    private volatile ConcurrentMap<String, Datum> datums = new ConcurrentHashMap<>();

    @Autowired
    private RaftPeerSet peers;

    @Autowired
    private SwitchDomain switchDomain;

    @Autowired
    private GlobalConfig globalConfig;

    @Autowired
    private RaftProxy raftProxy;

    @Autowired
    private RaftStore raftStore;

    public volatile Notifier notifier = new Notifier();

    private boolean initialized = false;

    @PostConstruct
    public void init() throws Exception {

        Loggers.RAFT.info("initializing Raft sub-system");

        executor.submit(notifier);

        long start = System.currentTimeMillis();

        raftStore.loadDatums(notifier, datums);

        setTerm(NumberUtils.toLong(raftStore.loadMeta().getProperty("term"), 0L));

        Loggers.RAFT.info("cache loaded, datum count: {}, current term: {}", datums.size(), peers.getTerm());

        while (true) {
            if (notifier.tasks.size() <= 0) {
                break;
            }
            Thread.sleep(1000L);
        }

        initialized = true;

        Loggers.RAFT.info("finish to load data from disk, cost: {} ms.", (System.currentTimeMillis() - start));

        GlobalExecutor.registerMasterElection(new MasterElection());
        GlobalExecutor.registerHeartbeat(new HeartBeat());

        Loggers.RAFT.info("timer started: leader timeout ms: {}, heart-beat timeout ms: {}",
            GlobalExecutor.LEADER_TIMEOUT_MS, GlobalExecutor.HEARTBEAT_INTERVAL_MS);
    }

    public Map<String, List<RecordListener>> getListeners() {
        return listeners;
    }

    public void signalPublish(String key, Record value) throws Exception {

        if (!isLeader()) {
            JSONObject params = new JSONObject();
            params.put("key", key);
            params.put("value", value);
            Map<String, String> parameters = new HashMap<>(1);
            parameters.put("key", key);

            raftProxy.proxyPostLarge(getLeader().ip, API_PUB, params.toJSONString(), parameters);
            return;
        }

        try {
            OPERATE_LOCK.lock();
            long start = System.currentTimeMillis();
            final Datum datum = new Datum();
            datum.key = key;
            datum.value = value;
            if (getDatum(key) == null) {
                datum.timestamp.set(1L);
            } else {
                datum.timestamp.set(getDatum(key).timestamp.incrementAndGet());
            }

            JSONObject json = new JSONObject();
            json.put("datum", datum);
            json.put("source", peers.local());

            onPublish(datum, peers.local());

            final String content = JSON.toJSONString(json);

            final CountDownLatch latch = new CountDownLatch(peers.majorityCount());
            for (final String server : peers.allServersIncludeMyself()) {
                if (isLeader(server)) {
                    latch.countDown();
                    continue;
                }
                final String url = buildURL(server, API_ON_PUB);
                HttpClient.asyncHttpPostLarge(url, Arrays.asList("key=" + key), content, new AsyncCompletionHandler<Integer>() {
                    @Override
                    public Integer onCompleted(Response response) throws Exception {
                        if (response.getStatusCode() != HttpURLConnection.HTTP_OK) {
                            Loggers.RAFT.warn("[RAFT] failed to publish data to peer, datumId={}, peer={}, http code={}",
                                datum.key, server, response.getStatusCode());
                            return 1;
                        }
                        latch.countDown();
                        return 0;
                    }

                    @Override
                    public STATE onContentWriteCompleted() {
                        return STATE.CONTINUE;
                    }
                });

            }

            if (!latch.await(UtilsAndCommons.RAFT_PUBLISH_TIMEOUT, TimeUnit.MILLISECONDS)) {
                // only majority servers return success can we consider this update success
                Loggers.RAFT.error("data publish failed, caused failed to notify majority, key={}", key);
                throw new IllegalStateException("data publish failed, caused failed to notify majority, key=" + key);
            }

            long end = System.currentTimeMillis();
            Loggers.RAFT.info("signalPublish cost {} ms, key: {}", (end - start), key);
        } finally {
            OPERATE_LOCK.unlock();
        }
    }

    //......
}

signalPublish方法判断当前节点是否是leader，如果不是则转发publish到leader节点的/v1/ns/raft/datum接口
如果是leader则构造datum以及peers.majorityCount()大小的CountDownLatch，然后遍历peers.allServersIncludeMyself()，对于leader节点直接latch.countDown()，对于非leader节点则发送异步请求，请求/v1/ns/raft/datum/commit接口，在onCompleted的时候，如果请求成功执行latch.countDown()
最后对于CountDownLatch未能在RAFT_PUBLISH_TIMEOUT返回的，抛出IllegalStateException

RaftController

nacos-1.1.3/naming/src/main/java/com/alibaba/nacos/naming/controllers/RaftController.java

@RestController
@RequestMapping({UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft",
    UtilsAndCommons.NACOS_SERVER_CONTEXT + UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft"})
public class RaftController {

    @Autowired
    private RaftConsistencyServiceImpl raftConsistencyService;

    @Autowired
    private ServiceManager serviceManager;

    @Autowired
    private RaftCore raftCore;

    //......

    @NeedAuth
    @RequestMapping(value = "/datum/commit", method = RequestMethod.POST)
    public String onPublish(HttpServletRequest request, HttpServletResponse response) throws Exception {

        response.setHeader("Content-Type", "application/json; charset=" + getAcceptEncoding(request));
        response.setHeader("Cache-Control", "no-cache");
        response.setHeader("Content-Encode", "gzip");

        String entity = IOUtils.toString(request.getInputStream(), "UTF-8");
        String value = URLDecoder.decode(entity, "UTF-8");
        JSONObject jsonObject = JSON.parseObject(value);
        String key = "key";

        RaftPeer source = JSON.parseObject(jsonObject.getString("source"), RaftPeer.class);
        JSONObject datumJson = jsonObject.getJSONObject("datum");

        Datum datum = null;
        if (KeyBuilder.matchInstanceListKey(datumJson.getString(key))) {
            datum = JSON.parseObject(jsonObject.getString("datum"), new TypeReference<Datum<Instances>>() {
            });
        } else if (KeyBuilder.matchSwitchKey(datumJson.getString(key))) {
            datum = JSON.parseObject(jsonObject.getString("datum"), new TypeReference<Datum<SwitchDomain>>() {
            });
        } else if (KeyBuilder.matchServiceMetaKey(datumJson.getString(key))) {
            datum = JSON.parseObject(jsonObject.getString("datum"), new TypeReference<Datum<Service>>() {
            });
        }

        raftConsistencyService.onPut(datum, source);
        return "ok";
    }

    //......
}

onPublish方法主要是执行raftConsistencyService.onPut(datum, source)

RaftConsistencyServiceImpl

nacos-1.1.3/naming/src/main/java/com/alibaba/nacos/naming/consistency/persistent/raft/RaftConsistencyServiceImpl.java

@Service
public class RaftConsistencyServiceImpl implements PersistentConsistencyService {

    @Autowired
    private RaftCore raftCore;

    @Autowired
    private RaftPeerSet peers;

    @Autowired
    private SwitchDomain switchDomain;

    //......

    public void onPut(Datum datum, RaftPeer source) throws NacosException {
        try {
            raftCore.onPublish(datum, source);
        } catch (Exception e) {
            Loggers.RAFT.error("Raft onPut failed.", e);
            throw new NacosException(NacosException.SERVER_ERROR, "Raft onPut failed, datum:" + datum + ", source: " + source, e);
        }
    }

    //......
}

onPut方法执行的是raftCore.onPublish(datum, source)

RaftCore

nacos-1.1.3/naming/src/main/java/com/alibaba/nacos/naming/consistency/persistent/raft/RaftCore.java

@Component
public class RaftCore {

    public static final String API_VOTE = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/vote";

    public static final String API_BEAT = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/beat";

    public static final String API_PUB = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/datum";

    public static final String API_DEL = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/datum";

    public static final String API_GET = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/datum";

    public static final String API_ON_PUB = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/datum/commit";

    public static final String API_ON_DEL = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/datum/commit";

    public static final String API_GET_PEER = UtilsAndCommons.NACOS_NAMING_CONTEXT + "/raft/peer";

    private ScheduledExecutorService executor = new ScheduledThreadPoolExecutor(1, new ThreadFactory() {
        @Override
        public Thread newThread(Runnable r) {
            Thread t = new Thread(r);

            t.setDaemon(true);
            t.setName("com.alibaba.nacos.naming.raft.notifier");

            return t;
        }
    });

    public static final Lock OPERATE_LOCK = new ReentrantLock();

    public static final int PUBLISH_TERM_INCREASE_COUNT = 100;

    private volatile Map<String, List<RecordListener>> listeners = new ConcurrentHashMap<>();

    private volatile ConcurrentMap<String, Datum> datums = new ConcurrentHashMap<>();

    @Autowired
    private RaftPeerSet peers;

    @Autowired
    private SwitchDomain switchDomain;

    @Autowired
    private GlobalConfig globalConfig;

    @Autowired
    private RaftProxy raftProxy;

    @Autowired
    private RaftStore raftStore;

    public volatile Notifier notifier = new Notifier();

    private boolean initialized = false;

    //......

    public void onPublish(Datum datum, RaftPeer source) throws Exception {
        RaftPeer local = peers.local();
        if (datum.value == null) {
            Loggers.RAFT.warn("received empty datum");
            throw new IllegalStateException("received empty datum");
        }

        if (!peers.isLeader(source.ip)) {
            Loggers.RAFT.warn("peer {} tried to publish data but wasn't leader, leader: {}",
                JSON.toJSONString(source), JSON.toJSONString(getLeader()));
            throw new IllegalStateException("peer(" + source.ip + ") tried to publish " +
                "data but wasn't leader");
        }

        if (source.term.get() < local.term.get()) {
            Loggers.RAFT.warn("out of date publish, pub-term: {}, cur-term: {}",
                JSON.toJSONString(source), JSON.toJSONString(local));
            throw new IllegalStateException("out of date publish, pub-term:"
                + source.term.get() + ", cur-term: " + local.term.get());
        }

        local.resetLeaderDue();

        // if data should be persistent, usually this is always true:
        if (KeyBuilder.matchPersistentKey(datum.key)) {
            raftStore.write(datum);
        }

        datums.put(datum.key, datum);

        if (isLeader()) {
            local.term.addAndGet(PUBLISH_TERM_INCREASE_COUNT);
        } else {
            if (local.term.get() + PUBLISH_TERM_INCREASE_COUNT > source.term.get()) {
                //set leader term:
                getLeader().term.set(source.term.get());
                local.term.set(getLeader().term.get());
            } else {
                local.term.addAndGet(PUBLISH_TERM_INCREASE_COUNT);
            }
        }
        raftStore.updateTerm(local.term.get());

        notifier.addTask(datum.key, ApplyAction.CHANGE);

        Loggers.RAFT.info("data added/updated, key={}, term={}", datum.key, local.term);
    }

    //......
}

onPublish方法首先判断请求的节点是否是leader，不是则抛出IllegalStateException；对于source.term小于local.term的抛出IllegalStateException
之后执行local.resetLeaderDue()，以及raftStore.write(datum)，datums.put(datum.key, datum)；对于leader节点执行local.term.addAndGet(PUBLISH_TERM_INCREASE_COUNT)，非leader节点则更新leader term以及local.term
最后执行raftStore.updateTerm(local.term.get())以及notifier.addTask(datum.key, ApplyAction.CHANGE)

小结

signalPublish方法判断当前节点是否是leader，如果不是则转发publish到leader节点的/v1/ns/raft/datum接口
如果是leader则构造datum以及peers.majorityCount()大小的CountDownLatch，然后遍历peers.allServersIncludeMyself()，对于leader节点直接latch.countDown()，对于非leader节点则发送异步请求，请求/v1/ns/raft/datum/commit接口，在onCompleted的时候，如果请求成功执行latch.countDown()
最后对于CountDownLatch未能在RAFT_PUBLISH_TIMEOUT返回的，抛出IllegalStateException

感谢各位的阅读，以上就是“nacos RaftCore中signalPublish的原理和使用方法”的内容了，经过本文的学习后，相信大家对nacos RaftCore中signalPublish的原理和使用方法这一问题有了更深刻的体会，具体使用情况还需要大家实践验证。这里是亿速云，小编将为大家推送更多相关知识点的文章，欢迎关注！

nacos RaftCore中signalPublish的原理和使用方法

序

RaftCore

RaftController

RaftConsistencyServiceImpl

RaftCore

小结

相关阅读