在看具体源码之前我们先回顾一下之前在《服务治理:Spring-cloud Eureka入门实例详解》所实现的内容。首先,对于服务注册中心、服务提供者、服务消费者这三个主要元素来说,后两者(也就是Eurake客户端)在整个运行机制中是大部分通信行为的主动发起者,而注册中心主要是处理请求的接受者,所以,我们可以从Eurake的客户端作为入口看看它是如何完成这些主要通信行为的。
我们将一个普通的Spring Boot应用注册到Eureka Server或是从Eureka Server中获取服务列表时,主要做了以下两件事:
现在我们来看一下@EnableDiscoveryClient的源码
/**
* Annotation to enable a DiscoveryClient implementation.
* @author Spencer Gibb
*/
@Target(ElementType.TYPE)
@Retention(RetentionPolicy.RUNTIME)
@Documented
@Inherited
@Import(EnableDiscoveryClientImportSelector.class)
public @interface EnableDiscoveryClient {
/**
* If true, the ServiceRegistry will automatically register the local server.
* 如果为真,ServiceRegistry将自动注册本地服务器
*/
boolean autoRegister() default true;
}
通过注释我们可以知道它主要用来开启DiscoveryClient的实例。接下来我们来看一下DiscoveryClient相关的类与接口。
其中,左边的org.springframework.cloud.client.discovery.DiscoveryClient是Spring Cloud的接口,它定义了用来发现服务的常用抽象方法,通过该接口可以有效地屏蔽服务治理的实现细节。org.springframework.cloud.netflix.eureka.EurekaDiscoveryClient是对该接口的实现,实现的是对Eureka发现服务的封装。右边的接口和类均来自com.netflix.discovery包,EurekaDiscoveryClient依赖了Netflix Eurake的com.netflix.discovery.EurakeClient接口,EurekaClient继承了LookuoServer接口,它们都是Netflix开源包的内容,主要定义了针对Eureka的发现服务的抽象方法,而真正实现发现服务的则是Netflix包中的com.netflix.discovery.DiscoveryClient类。
根据com.netflix.discovery.DiscoveryClient类的头部注释,我们可以知道以下信息:
这个类用于帮助Eurake Server互相协作。
Eureka Client负责下面的任务:
- 向Eureka Server注册服务实例
- 向Eureka Server服务续约
- 当服务关闭期间,向Eureka Server取消租约
- 查询Eureka Server中的服务实例列表
Eureka Client还需要配置一个Eureka Server的URL列表
在具体看Eureka Client负责完成的任务之前,我们先看看在哪里对Eureka Server的URL列表进行配置,根据我们配置的属性名eureka.client.serviceUrl.defaultZone,通过serviceUrl可以找到该属性相关的加载属性,但是在SR5版本中它们都被@Deprecated标注为不再建议使用,并@link到替代类com.netflix.discovery.endpoint.EndpointUtils,所以我们在com.netflix.discovery.endpoint.EndpointUtils类下找这个方法
/**
* Get the list of all eureka service urls from properties file for the eureka client to talk to.
* 从属性文件中获取所有eureka服务url的列表,以便与eureka客户机通信
* @param clientConfig the clientConfig to use
* @param instanceZone The zone in which the client resides
* @param preferSameZone true if we have to prefer the same zone as the client, false otherwise
* @return The list of all eureka service urls for the eureka client to talk to
*/
public static List<String> getServiceUrlsFromConfig(EurekaClientConfig clientConfig, String instanceZone, boolean preferSameZone) {
List<String> orderedUrls = new ArrayList<String>();
//重点部分
String region = getRegion(clientConfig);
String[] availZones = clientConfig.getAvailabilityZones(clientConfig.getRegion());
if (availZones == null || availZones.length == 0) {
availZones = new String[1];
availZones[0] = DEFAULT_ZONE;
}
logger.debug("The availability zone for the given region {} are {}", region, availZones);
//重点部分
int myZoneOffset = getZoneOffset(instanceZone, preferSameZone, availZones);
List<String> serviceUrls = clientConfig.getEurekaServerServiceUrls(availZones[myZoneOffset]);
if (serviceUrls != null) {
orderedUrls.addAll(serviceUrls);
}
int currentOffset = myZoneOffset == (availZones.length - 1) ? 0 : (myZoneOffset + 1);
while (currentOffset != myZoneOffset) {
serviceUrls = clientConfig.getEurekaServerServiceUrls(availZones[currentOffset]);
if (serviceUrls != null) {
orderedUrls.addAll(serviceUrls);
}
if (currentOffset == (availZones.length - 1)) {
currentOffset = 0;
} else {
currentOffset++;
}
}
if (orderedUrls.size() < 1) {
throw new IllegalArgumentException("DiscoveryClient: invalid serviceUrl specified!");
}
return orderedUrls;
}
在上面的函数中,可以发现,客户端依次加载了两个内容,第一个是Region,第二个是Zone;
/**
* Get the region that this particular instance is in.
* 获取这个特定实例所在的区域。
* @return - The region in which the particular instance belongs to.
*/
public static String getRegion(EurekaClientConfig clientConfig) {
String region = clientConfig.getRegion();
if (region == null) {
region = DEFAULT_REGION;
}
region = region.trim().toLowerCase();
return region;
}
@Override
public String[] getAvailabilityZones(String region) {
String value = this.availabilityZones.get(region);
if (value == null) {
value = DEFAULT_ZONE;
}
return value.split(",");
}
在获取了Region和Zone的信息之后,开始真正加载Eureka Server的具体地址,它根据传入的参数按一定算法确定加载位于哪一个Zone配置的serviceUrls
int myZoneOffset = getZoneOffset(instanceZone, preferSameZone, availZones);
List<String> serviceUrls = clientConfig.getEurekaServerServiceUrls(availZones[myZoneOffset]);
具体获取serviceUrls的实现,我们可以看getEurekaServerServiceUrls函数的具体实现类EurekaClientConfigBean,该类是EurekaClientConfig和EurekaConstants接口的实现,用来加载配置文件中的内容。
@Override
public List<String> getEurekaServerServiceUrls(String myZone) {
String serviceUrls = this.serviceUrl.get(myZone);
if (serviceUrls == null || serviceUrls.isEmpty()) {
serviceUrls = this.serviceUrl.get(DEFAULT_ZONE);
}
if (!StringUtils.isEmpty(serviceUrls)) {
final String[] serviceUrlsSplit = StringUtils.commaDelimitedListToStringArray(serviceUrls);
List<String> eurekaServiceUrls = new ArrayList<>(serviceUrlsSplit.length);
for (String eurekaServiceUrl : serviceUrlsSplit) {
if (!endsWithSlash(eurekaServiceUrl)) {
eurekaServiceUrl += "/";
}
eurekaServiceUrls.add(eurekaServiceUrl);
}
return eurekaServiceUrls;
}
return new ArrayList<>();
}
当我们在为微服务应用中使用Ribbon来实现服务调用时,对于Zone的设置可以在负载均衡时实现区域亲和特性:Ribbon的默认策略会优先访问同客户端处于一个Zone中的服务端实例,只有当同一个Zone中没有可用服务端实例的时候才会访问其他Zone中的实例。
在理解了多个服务注册中心信息的加载后,我们再回头看看DiscoveryClient类是如何实现“服务注册”行为的,来看一下它的构造函数
@Inject
DiscoveryClient(ApplicationInfoManager applicationInfoManager, EurekaClientConfig config, AbstractDiscoveryClientOptionalArgs args,
Provider<BackupRegistry> backupRegistryProvider) {
if (args != null) {
this.healthCheckHandlerProvider = args.healthCheckHandlerProvider;
this.healthCheckCallbackProvider = args.healthCheckCallbackProvider;
this.eventListeners.addAll(args.getEventListeners());
this.preRegistrationHandler = args.preRegistrationHandler;
} else {
this.healthCheckCallbackProvider = null;
this.healthCheckHandlerProvider = null;
this.preRegistrationHandler = null;
}
this.applicationInfoManager = applicationInfoManager;
InstanceInfo myInfo = applicationInfoManager.getInfo();
clientConfig = config;
staticClientConfig = clientConfig;
transportConfig = config.getTransportConfig();
instanceInfo = myInfo;
if (myInfo != null) {
appPathIdentifier = instanceInfo.getAppName() + "/" + instanceInfo.getId();
} else {
logger.warn("Setting instanceInfo to a passed in null value");
}
this.backupRegistryProvider = backupRegistryProvider;
this.urlRandomizer = new EndpointUtils.InstanceInfoBasedUrlRandomizer(instanceInfo);
localRegionApps.set(new Applications());
fetchRegistryGeneration = new AtomicLong(0);
remoteRegionsToFetch = new AtomicReference<String>(clientConfig.fetchRegistryForRemoteRegions());
remoteRegionsRef = new AtomicReference<>(remoteRegionsToFetch.get() == null ? null : remoteRegionsToFetch.get().split(","));
if (config.shouldFetchRegistry()) {
this.registryStalenessMonitor = new ThresholdLevelsMetric(this, METRIC_REGISTRY_PREFIX + "lastUpdateSec_", new long[]{15L, 30L, 60L, 120L, 240L, 480L});
} else {
this.registryStalenessMonitor = ThresholdLevelsMetric.NO_OP_METRIC;
}
if (config.shouldRegisterWithEureka()) {
this.heartbeatStalenessMonitor = new ThresholdLevelsMetric(this, METRIC_REGISTRATION_PREFIX + "lastHeartbeatSec_", new long[]{15L, 30L, 60L, 120L, 240L, 480L});
} else {
this.heartbeatStalenessMonitor = ThresholdLevelsMetric.NO_OP_METRIC;
}
logger.info("Initializing Eureka in region {}", clientConfig.getRegion());
if (!config.shouldRegisterWithEureka() && !config.shouldFetchRegistry()) {
logger.info("Client configured to neither register nor query for data.");
scheduler = null;
heartbeatExecutor = null;
cacheRefreshExecutor = null;
eurekaTransport = null;
instanceRegionChecker = new InstanceRegionChecker(new PropertyBasedAzToRegionMapper(config), clientConfig.getRegion());
// This is a bit of hack to allow for existing code using DiscoveryManager.getInstance()
// to work with DI'd DiscoveryClient
DiscoveryManager.getInstance().setDiscoveryClient(this);
DiscoveryManager.getInstance().setEurekaClientConfig(config);
initTimestampMs = System.currentTimeMillis();
logger.info("Discovery Client initialized at timestamp {} with initial instances count: {}",
initTimestampMs, this.getApplications().size());
return; // no need to setup up an network tasks and we are done
}
try {
// default size of 2 - 1 each for heartbeat and cacheRefresh
scheduler = Executors.newScheduledThreadPool(2,
new ThreadFactoryBuilder()
.setNameFormat("DiscoveryClient-%d")
.setDaemon(true)
.build());
heartbeatExecutor = new ThreadPoolExecutor(
1, clientConfig.getHeartbeatExecutorThreadPoolSize(), 0, TimeUnit.SECONDS,
new SynchronousQueue<Runnable>(),
new ThreadFactoryBuilder()
.setNameFormat("DiscoveryClient-HeartbeatExecutor-%d")
.setDaemon(true)
.build()
); // use direct handoff
cacheRefreshExecutor = new ThreadPoolExecutor(
1, clientConfig.getCacheRefreshExecutorThreadPoolSize(), 0, TimeUnit.SECONDS,
new SynchronousQueue<Runnable>(),
new ThreadFactoryBuilder()
.setNameFormat("DiscoveryClient-CacheRefreshExecutor-%d")
.setDaemon(true)
.build()
); // use direct handoff
eurekaTransport = new EurekaTransport();
scheduleServerEndpointTask(eurekaTransport, args);
AzToRegionMapper azToRegionMapper;
if (clientConfig.shouldUseDnsForFetchingServiceUrls()) {
azToRegionMapper = new DNSBasedAzToRegionMapper(clientConfig);
} else {
azToRegionMapper = new PropertyBasedAzToRegionMapper(clientConfig);
}
if (null != remoteRegionsToFetch.get()) {
azToRegionMapper.setRegionsToFetch(remoteRegionsToFetch.get().split(","));
}
instanceRegionChecker = new InstanceRegionChecker(azToRegionMapper, clientConfig.getRegion());
} catch (Throwable e) {
throw new RuntimeException("Failed to initialize DiscoveryClient!", e);
}
if (clientConfig.shouldFetchRegistry() && !fetchRegistry(false)) {
fetchRegistryFromBackup();
}
// call and execute the pre registration handler before all background tasks (inc registration) is started
if (this.preRegistrationHandler != null) {
this.preRegistrationHandler.beforeRegistration();
}
if (clientConfig.shouldRegisterWithEureka() && clientConfig.shouldEnforceRegistrationAtInit()) {
try {
if (!register() ) {
throw new IllegalStateException("Registration error at startup. Invalid server response.");
}
} catch (Throwable th) {
logger.error("Registration error at startup: {}", th.getMessage());
throw new IllegalStateException(th);
}
}
//重点部分
// finally, init the schedule tasks (e.g. cluster resolvers, heartbeat, instanceInfo replicator, fetch
initScheduledTasks();
try {
Monitors.registerObject(this);
} catch (Throwable e) {
logger.warn("Cannot register timers", e);
}
// This is a bit of hack to allow for existing code using DiscoveryManager.getInstance()
// to work with DI'd DiscoveryClient
DiscoveryManager.getInstance().setDiscoveryClient(this);
DiscoveryManager.getInstance().setEurekaClientConfig(config);
initTimestampMs = System.currentTimeMillis();
logger.info("Discovery Client initialized at timestamp {} with initial instances count: {}",
initTimestampMs, this.getApplications().size());
}
可以从上面的源码发现经过一系统属性值判断后,最终会调用initScheduledTasks()这个方法,下面是这个方法的实现源码
private void initScheduledTasks() {
if (clientConfig.shouldFetchRegistry()) {
// registry cache refresh timer
int registryFetchIntervalSeconds = clientConfig.getRegistryFetchIntervalSeconds();
int expBackOffBound = clientConfig.getCacheRefreshExecutorExponentialBackOffBound();
scheduler.schedule(
new TimedSupervisorTask(
"cacheRefresh",
scheduler,
cacheRefreshExecutor,
registryFetchIntervalSeconds,
TimeUnit.SECONDS,
expBackOffBound,
new CacheRefreshThread()
),
registryFetchIntervalSeconds, TimeUnit.SECONDS);
}
if (clientConfig.shouldRegisterWithEureka()) {
int renewalIntervalInSecs = instanceInfo.getLeaseInfo().getRenewalIntervalInSecs();
int expBackOffBound = clientConfig.getHeartbeatExecutorExponentialBackOffBound();
logger.info("Starting heartbeat executor: " + "renew interval is: {}", renewalIntervalInSecs);
// Heartbeat timer
scheduler.schedule(
new TimedSupervisorTask(
"heartbeat",
scheduler,
heartbeatExecutor,
renewalIntervalInSecs,
TimeUnit.SECONDS,
expBackOffBound,
new HeartbeatThread()
),
renewalIntervalInSecs, TimeUnit.SECONDS);
// InstanceInfo replicator
instanceInfoReplicator = new InstanceInfoReplicator(
this,
instanceInfo,
clientConfig.getInstanceInfoReplicationIntervalSeconds(),
2); // burstSize
statusChangeListener = new ApplicationInfoManager.StatusChangeListener() {
@Override
public String getId() {
return "statusChangeListener";
}
@Override
public void notify(StatusChangeEvent statusChangeEvent) {
if (InstanceStatus.DOWN == statusChangeEvent.getStatus() ||
InstanceStatus.DOWN == statusChangeEvent.getPreviousStatus()) {
// log at warn level if DOWN was involved
logger.warn("Saw local status change event {}", statusChangeEvent);
} else {
logger.info("Saw local status change event {}", statusChangeEvent);
}
instanceInfoReplicator.onDemandUpdate();
}
};
if (clientConfig.shouldOnDemandUpdateStatusChange()) {
applicationInfoManager.registerStatusChangeListener(statusChangeListener);
}
instanceInfoReplicator.start(clientConfig.getInitialInstanceInfoReplicationIntervalSeconds());
} else {
logger.info("Not registering with Eureka server per configuration");
}
}
从上面的函数中,可以看到一个与服务注册相关的判断语句if(clientConfig.shouldRegisterWithEureka()),在该分支内,创建了一个InstanceInfoReplicator类的实例,它会执行一个定时任务,而这个定时任务的具体工作可以查看该类的run()函数,具体如下:
public void run() {
try {
discoveryClient.refreshInstanceInfo();
Long dirtyTimestamp = instanceInfo.isDirtyWithTime();
if (dirtyTimestamp != null) {
discoveryClient.register();
instanceInfo.unsetIsDirty(dirtyTimestamp);
}
} catch (Throwable t) {
logger.warn("There was a problem with the instance info replicator", t);
} finally {
Future next = scheduler.schedule(this, replicationIntervalSeconds, TimeUnit.SECONDS);
scheduledPeriodicRef.set(next);
}
}
相信大家都发现了**discoveryClient.register()**这一行,真正触发调用注册的地方就在这里,继续看register()的实现内容,如下:
/**
* Register with the eureka service by making the appropriate REST call.
*/
boolean register() throws Throwable {
logger.info(PREFIX + "{}: registering service...", appPathIdentifier);
EurekaHttpResponse<Void> httpResponse;
try {
httpResponse = eurekaTransport.registrationClient.register(instanceInfo);
} catch (Exception e) {
logger.warn(PREFIX + "{} - registration failed {}", appPathIdentifier, e.getMessage(), e);
throw e;
}
if (logger.isInfoEnabled()) {
logger.info(PREFIX + "{} - registration status: {}", appPathIdentifier, httpResponse.getStatusCode());
}
return httpResponse.getStatusCode() == 204;
}
通过REST请求的方式注册eureka service,发起注册请求的时候,传入了一个com.netflix.appinfo.Instance对象,该对象就是注册时客户端给服务端的服务的元数据。
我们继续来看DiscoveryClient的initScheduleTasks函数,不难发现在其中还有两个定时任务,分别是“服务获取”和“服务续约”。
if (clientConfig.shouldFetchRegistry()) {
// registry cache refresh timer
int registryFetchIntervalSeconds = clientConfig.getRegistryFetchIntervalSeconds();
int expBackOffBound = clientConfig.getCacheRefreshExecutorExponentialBackOffBound();
scheduler.schedule(
new TimedSupervisorTask(
"cacheRefresh",
scheduler,
cacheRefreshExecutor,
registryFetchIntervalSeconds,
TimeUnit.SECONDS,
expBackOffBound,
new CacheRefreshThread()
),
registryFetchIntervalSeconds, TimeUnit.SECONDS);
}
if (clientConfig.shouldRegisterWithEureka()) {
int renewalIntervalInSecs = instanceInfo.getLeaseInfo().getRenewalIntervalInSecs();
int expBackOffBound = clientConfig.getHeartbeatExecutorExponentialBackOffBound();
logger.info("Starting heartbeat executor: " + "renew interval is: {}", renewalIntervalInSecs);
// Heartbeat timer
scheduler.schedule(
new TimedSupervisorTask(
"heartbeat",
scheduler,
heartbeatExecutor,
renewalIntervalInSecs,
TimeUnit.SECONDS,
expBackOffBound,
new HeartbeatThread()
),
renewalIntervalInSecs, TimeUnit.SECONDS);
......
}
从上面的代码实现可以看出,“服务获取”任务相对于“服务续约”和“服务注册”任务更为独立。“服务续约”与“服务注册”在同一个if逻辑中,服务注册到Eureka Server后,需要一个心跳去续约,防止被剔除,所以它们是成对出现的。
而“服务获取”的逻辑在独立的一个if判断中,其判断依据就是eureka.client.register-with-eureka=true参数,默认为true。为了定期更新客户端的服务清单,以保证客户端能够访问健康的服务实例,“服务获取”的请求不会只限于服务启动,而是一个定时执行的任务,从源码中我们可以看到任务运行中的registryFetchIntervalSeconds参数对应的eureka.client.registry-fetch-interval-seconds=30配置参数,它默认为30s。
“服务续约”的实现较为简单,直接以REST请求的方式进行续约,我们从**new HeartbeatThread()**这个入参进到函数中看到如下实现:
/**
* The heartbeat task that renews the lease in the given intervals.
*/
private class HeartbeatThread implements Runnable {
public void run() {
if (renew()) {
lastSuccessfulHeartbeatTimestamp = System.currentTimeMillis();
}
}
}
/**
* Renew with the eureka service by making the appropriate REST call
*/
boolean renew() {
EurekaHttpResponse<InstanceInfo> httpResponse;
try {
httpResponse = eurekaTransport.registrationClient.sendHeartBeat(instanceInfo.getAppName(), instanceInfo.getId(), instanceInfo, null);
logger.debug(PREFIX + "{} - Heartbeat status: {}", appPathIdentifier, httpResponse.getStatusCode());
if (httpResponse.getStatusCode() == 404) {
REREGISTER_COUNTER.increment();
logger.info(PREFIX + "{} - Re-registering apps/{}", appPathIdentifier, instanceInfo.getAppName());
long timestamp = instanceInfo.setIsDirtyWithTime();
boolean success = register();
if (success) {
instanceInfo.unsetIsDirty(timestamp);
}
return success;
}
return httpResponse.getStatusCode() == 200;
} catch (Throwable e) {
logger.error(PREFIX + "{} - was unable to send heartbeat!", appPathIdentifier, e);
return false;
}
}
而“服务获取”则复杂一些,会根据是否是第一次获取发起不同的REST请求和相应处理。同样从TimedSupervisorTask的最后一个入参**new CacheRefreshThread()**进去:
/**
* The task that fetches the registry information at specified intervals.
*
*/
class CacheRefreshThread implements Runnable {
public void run() {
refreshRegistry();
}
}
@VisibleForTesting
void refreshRegistry() {
try {
boolean isFetchingRemoteRegionRegistries = isFetchingRemoteRegionRegistries();
boolean remoteRegionsModified = false;
// This makes sure that a dynamic change to remote regions to fetch is honored.
//这确保了对远程区域进行动态更改以获取数据
String latestRemoteRegions = clientConfig.fetchRegistryForRemoteRegions();
if (null != latestRemoteRegions) {
String currentRemoteRegions = remoteRegionsToFetch.get();
if (!latestRemoteRegions.equals(currentRemoteRegions)) {
// Both remoteRegionsToFetch and AzToRegionMapper.regionsToFetch need to be in sync
synchronized (instanceRegionChecker.getAzToRegionMapper()) {
if (remoteRegionsToFetch.compareAndSet(currentRemoteRegions, latestRemoteRegions)) {
String[] remoteRegions = latestRemoteRegions.split(",");
remoteRegionsRef.set(remoteRegions);
instanceRegionChecker.getAzToRegionMapper().setRegionsToFetch(remoteRegions);
remoteRegionsModified = true;
} else {
logger.info("Remote regions to fetch modified concurrently," +
" ignoring change from {} to {}", currentRemoteRegions, latestRemoteRegions);
}
}
} else {
// Just refresh mapping to reflect any DNS/Property change
//只需刷新映射以反映任何DNS/属性更改
instanceRegionChecker.getAzToRegionMapper().refreshMapping();
}
}
//重点部分
boolean success = fetchRegistry(remoteRegionsModified);
if (success) {
registrySize = localRegionApps.get().size();
lastSuccessfulRegistryFetchTimestamp = System.currentTimeMillis();
}
if (logger.isDebugEnabled()) {
StringBuilder allAppsHashCodes = new StringBuilder();
allAppsHashCodes.append("Local region apps hashcode: ");
allAppsHashCodes.append(localRegionApps.get().getAppsHashCode());
allAppsHashCodes.append(", is fetching remote regions? ");
allAppsHashCodes.append(isFetchingRemoteRegionRegistries);
for (Map.Entry<String, Applications> entry : remoteRegionVsApps.entrySet()) {
allAppsHashCodes.append(", Remote region: ");
allAppsHashCodes.append(entry.getKey());
allAppsHashCodes.append(" , apps hashcode: ");
allAppsHashCodes.append(entry.getValue().getAppsHashCode());
}
logger.debug("Completed cache refresh task for discovery. All Apps hash code is {} ",
allAppsHashCodes);
}
} catch (Throwable e) {
logger.error("Cannot fetch registry from server", e);
}
}
/**
* Fetches the registry information. 获取注册表信息
*除非在协调eureka服务器和客户端注册表信息方面存在问题,否则此方法只尝试在第一次获取之后获取增量
*/
private boolean fetchRegistry(boolean forceFullRegistryFetch) {
Stopwatch tracer = FETCH_REGISTRY_TIMER.start();
try {
// If the delta is disabled or if it is the first time, get all applications
//如果禁用了增量,或者这是第一次,则获取所有应用程序
Applications applications = getApplications();
if (clientConfig.shouldDisableDelta()
|| (!Strings.isNullOrEmpty(clientConfig.getRegistryRefreshSingleVipAddress()))
|| forceFullRegistryFetch
|| (applications == null)
|| (applications.getRegisteredApplications().size() == 0)
|| (applications.getVersion() == -1)) //Client application does not have latest library supporting delta
{
logger.info("Disable delta property : {}", clientConfig.shouldDisableDelta());
logger.info("Single vip registry refresh property : {}", clientConfig.getRegistryRefreshSingleVipAddress());
logger.info("Force full registry fetch : {}", forceFullRegistryFetch);
logger.info("Application is null : {}", (applications == null));
logger.info("Registered Applications size is zero : {}",
(applications.getRegisteredApplications().size() == 0));
logger.info("Application version is -1: {}", (applications.getVersion() == -1));
//第一次获取
getAndStoreFullRegistry();
} else {
//非第一次获取
getAndUpdateDelta(applications);
}
applications.setAppsHashCode(applications.getReconcileHashCode());
logTotalInstances();
} catch (Throwable e) {
logger.error(PREFIX + "{} - was unable to refresh its cache! status = {}", appPathIdentifier, e.getMessage(), e);
return false;
} finally {
if (tracer != null) {
tracer.stop();
}
}
// Notify about cache refresh before updating the instance remote status
//在更新实例远程状态之前,通知缓存刷新
onCacheRefreshed();
// Update remote status based on refreshed data held in the cache
//根据缓存中保存的刷新数据更新远程状态
updateInstanceRemoteStatus();
// registry was fetched successfully, so return true
//成功获取registry,因此返回true
return true;
}
/**
* Gets the full registry information from the eureka server and stores it locally.
* 从eureka服务器获取完整的注册表信息,并将其存储在本地
*/
private void getAndStoreFullRegistry() throws Throwable {
long currentUpdateGeneration = fetchRegistryGeneration.get();
logger.info("Getting all instance registry info from the eureka server");
Applications apps = null;
EurekaHttpResponse<Applications> httpResponse = clientConfig.getRegistryRefreshSingleVipAddress() == null
? eurekaTransport.queryClient.getApplications(remoteRegionsRef.get())
: eurekaTransport.queryClient.getVip(clientConfig.getRegistryRefreshSingleVipAddress(), remoteRegionsRef.get());
if (httpResponse.getStatusCode() == Status.OK.getStatusCode()) {
apps = httpResponse.getEntity();
}
logger.info("The response status is {}", httpResponse.getStatusCode());
if (apps == null) {
logger.error("The application is null for some reason. Not storing this information");
} else if (fetchRegistryGeneration.compareAndSet(currentUpdateGeneration, currentUpdateGeneration + 1)) {
localRegionApps.set(this.filterAndShuffle(apps));
logger.debug("Got full registry with apps hashcode {}", apps.getAppsHashCode());
} else {
logger.warn("Not updating applications as another thread is updating it already");
}
}
private void getAndUpdateDelta(Applications applications) throws Throwable {
long currentUpdateGeneration = fetchRegistryGeneration.get();
Applications delta = null;
EurekaHttpResponse<Applications> httpResponse = eurekaTransport.queryClient.getDelta(remoteRegionsRef.get());
if (httpResponse.getStatusCode() == Status.OK.getStatusCode()) {
delta = httpResponse.getEntity();
}
if (delta == null) {
logger.warn("The server does not allow the delta revision to be applied because it is not safe. Hence got the full registry.");
//重新调用getAndStoreFullRegistry方法进行REST方式注册
getAndStoreFullRegistry();
} else if (fetchRegistryGeneration.compareAndSet(currentUpdateGeneration, currentUpdateGeneration + 1)) {
logger.debug("Got delta update with apps hashcode {}", delta.getAppsHashCode());
String reconcileHashCode = "";
if (fetchRegistryUpdateLock.tryLock()) {
try {
updateDelta(delta);
reconcileHashCode = getReconcileHashCode(applications);
} finally {
fetchRegistryUpdateLock.unlock();
}
} else {
logger.warn("Cannot acquire update lock, aborting getAndUpdateDelta");
}
// There is a diff in number of instances for some reason
if (!reconcileHashCode.equals(delta.getAppsHashCode()) || clientConfig.shouldLogDeltaDiff()) {
reconcileAndLogDifference(delta, reconcileHashCode); // this makes a remoteCall
}
} else {
logger.warn("Not updating application delta as another thread is updating it already");
logger.debug("Ignoring delta update with apps hashcode {}, as another thread is updating it already", delta.getAppsHashCode());
}
}
通过上面的源码分析,我们可以看到所有的交互都是通过REST请求来发起的。接下来一起来看看服务注册中心对这些请求的处理。Eureka Server对于各类REST请求的定义都位于com.netflix.eureka.resources包下。
以“服务注册”请求为例:
@POST
@Consumes({"application/json", "application/xml"})
public Response addInstance(InstanceInfo info,
@HeaderParam(PeerEurekaNode.HEADER_REPLICATION) String isReplication) {
logger.debug("Registering instance {} (replication={})", info.getId(), isReplication);
// validate that the instanceinfo contains all the necessary required fields
//验证instanceinfo包含所有必需的必需字段
if (isBlank(info.getId())) {
return Response.status(400).entity("Missing instanceId").build();
} else if (isBlank(info.getHostName())) {
return Response.status(400).entity("Missing hostname").build();
} else if (isBlank(info.getIPAddr())) {
return Response.status(400).entity("Missing ip address").build();
} else if (isBlank(info.getAppName())) {
return Response.status(400).entity("Missing appName").build();
} else if (!appName.equals(info.getAppName())) {
return Response.status(400).entity("Mismatched appName, expecting " + appName + " but was " + info.getAppName()).build();
} else if (info.getDataCenterInfo() == null) {
return Response.status(400).entity("Missing dataCenterInfo").build();
} else if (info.getDataCenterInfo().getName() == null) {
return Response.status(400).entity("Missing dataCenterInfo Name").build();
}
// handle cases where clients may be registering with bad DataCenterInfo with missing data
//处理客户端可能注册了错误的DataCenterInfo而丢失数据的情况
DataCenterInfo dataCenterInfo = info.getDataCenterInfo();
if (dataCenterInfo instanceof UniqueIdentifier) {
String dataCenterInfoId = ((UniqueIdentifier) dataCenterInfo).getId();
if (isBlank(dataCenterInfoId)) {
boolean experimental = "true".equalsIgnoreCase(serverConfig.getExperimental("registration.validation.dataCenterInfoId"));
if (experimental) {
String entity = "DataCenterInfo of type " + dataCenterInfo.getClass() + " must contain a valid id";
return Response.status(400).entity(entity).build();
} else if (dataCenterInfo instanceof AmazonInfo) {
AmazonInfo amazonInfo = (AmazonInfo) dataCenterInfo;
String effectiveId = amazonInfo.get(AmazonInfo.MetaDataKey.instanceId);
if (effectiveId == null) {
amazonInfo.getMetadata().put(AmazonInfo.MetaDataKey.instanceId.getName(), info.getId());
}
} else {
logger.warn("Registering DataCenterInfo of type {} without an appropriate id", dataCenterInfo.getClass());
}
}
}
registry.register(info, "true".equals(isReplication));
return Response.status(204).build(); // 204 to be backwards compatible
}
在对注册信息进行一堆校验后,会调用org.springframework.cloud.netflix.eureka.server.InstanceRegistry对象中的void **register(final InstanceInfo info, final boolean isReplication)**方法
@Override
public void register(final InstanceInfo info, final boolean isReplication) {
handleRegistration(info, resolveInstanceLeaseDuration(info), isReplication);
super.register(info, isReplication);
}
private void handleRegistration(InstanceInfo info, int leaseDuration,
boolean isReplication) {
log("register " + info.getAppName() + ", vip " + info.getVIPAddress()
+ ", leaseDuration " + leaseDuration + ", isReplication "
+ isReplication);
publishEvent(new EurekaInstanceRegisteredEvent(this, info, leaseDuration,
isReplication));
}
在注册函数中,先调用publishEvent函数,将该新服务注册的事件传播出去,然后调用com.netflix.eureka.registry.AbstractInstanceRegistry父类中的注册实现,将InstanceInfo中的元数据信息存储在一个ConcurrentHashMap对象中。注册中心存储了两层Map结构,第一层的key存储服务名:InstanceInfo中的属性名,第二层的key存储实例名;InstanceInfo中的instanceId属性。
在Eureka的服务治理体系中,主要分为服务端与客户端两个不同的角色,服务端为服务注册中心,而客户端为各个提供接口的微服务应用。
Eurake客户端的配置主要分为以下两个方面
在《服务治理:Spring-cloud Eureka入门实例详解》的示例中,我们演示了如何将一个SpringBoot应用纳入Eureka的服务治理体系,除了引入Eureka的依赖之外,就是在配置文件中指定注册中心,主要通过eureka.client.serviceUrl参数实现。该参数的定义如下所示:
private Map<String, String> serviceUrl = new HashMap<>();
{
//DEFAULT_ZONE="defaultZone",DEFAULT_URL="http://localhost:8761/eureka/"
this.serviceUrl.put(DEFAULT_ZONE, DEFAULT_URL);
}
它的配置值存储在HashMap类型中,并且设置一组默认值,默认值的key为defaultZone、value为http://localhost:8761/eureka/
在我们之前的示例中服务注册中心指定了1110和1111的端口,我们在服务注册类的配置里设置eureka.client.serviceUrl.defaultZone=http://localhost:1110/eureka/,http://localhost:1111/eureka
,可以配置一个注册中心,也可以配置多个实现高可用的服务注册中心集群。
另外,为了服务注册中心的安全考虑,我们也可以在配置serviceUrl时,在URL中加入相应的安全校验信息,比如
http://:@localhost:1110/eureka
下面整理了org.springframework.cloud.netflix.eureka.EurekaClientConfigBean类中定义的常用配置参数以及对应的说明和默认值,这些参数均以eureka.client为前缀。
参数名 | 说明 | 默认值 |
---|---|---|
enable | 启用Eurake | true |
registryFetchIntervalSeconds | 从Eureka服务端获取注册信息的间隔时间,单位秒 | 30 |
instanceInfoReplicationIntervalSeconds | 更新实例信息的变化到Eureka服务端的间隔时间,单位秒 | 30 |
initialInstanceInfoReplicationIntervalSeconds | 初始化实例信息到Eureka服务端的间隔时间,单位秒 | 40 |
eurekaServiceUrlPollIntervalSeconds | 轮询Eureka服务端地址更改的间隔时间,单位秒,当我们与Spring Cloud Config配合,动态刷新Eureka的serviceUrl地址时需要关注改参数 | 300 |
eurekaServerReadTimeoutSeconds | 读取EurekaServer信息的超时时间,单位秒 | 8 |
eurekaServerConnectTimeoutSeconds | 连接EurekaServer的超时时间,单位为秒 | 5 |
eurekaServerTotalConnections | 从Eureka客户端到所有Eureka服务端的连接总数 | 200 |
eurekaServerTotalConnectionsPerHost | 从Eureka客户端到每个Eureka服务端主机的连接总数 | 50 |
eurekaConnectionIdleTimeoutSeconds | Eureka服务端连接的空闲关闭时间,单位秒 | 30 |
heartbeatExecutorThreadPoolSize | 心跳连接池初始化线程数 | 2 |
heartbeatExecutorExponentialBackOffBound | 心跳超时重试延迟时间的最多乘数值 | 10 |
cacheRefreshExecutorThreadPoolSize | 缓存刷新线程池的初始化线程数 | 2 |
cacheRefreshExecutorExponentialBackOffBound | 缓存刷新重试延迟时间的最大乘数值 | 10 |
useDnsForFetchingServiceUrls | 使用DNS来获取Eureka服务端的serviceUrl | false |
registerWithEureka | 是否将自身的实例信息注册到Eureka服务端 | true |
preferSameZoneEureka | 是否偏好使用处于相同Zone的Eureka服务端 | true |
filterOnlyUpInstances | 获取实时时是否过来,仅保留UP状态的实例 | true |
fetchRegistry | 是否从Eureka服务端获取注册信息 | true |
关于服务实例类的配置信息,我们可以查看org.springframework.cloud.netflix.eureka.EurekaInstanceConfigBean的源码来获取详细内容,这些配置信息都以eureka.instance为前缀。
在org.springframework.cloud.netflix.eureka.EurekaInstanceConfigBean的配置信息中,有一部分内容都是对服务实例元数据的配置,所谓元数据就是Eureka客户端在向服务注册中心发送注册请求时,用来描述自身服务信息的对象,其中包含了一些标准化的元数据,比如服务名称、实例名称、实例IP、实力端口等用于服务治理额重要信息,以及一些用于负载均衡策略或是其他特殊用途的自定义元数据信息。
在使用Spring Cloud Eurake的时候,所有的配置信息都是通过org.springframework.cloud.netflix.eureka.EurekaInstanceConfigBean进行加载,但在真正进行服务注册的时候,还是会包装成com.netflix.appinfo.InstanceInfo对象发送给Eureka服务端。
其中InstanceInfo中的metadata参数是自定义的元数据信息,而其他成员变量则是标准化的元数据信息。
@XStreamAlias("metadata")
private volatile Map<String, String> metadata;
我们可以通过eureka.instance.=的格式对标准化元数据直接进行配置,其中propertises就是EurekaInstanceConfigBean对象中的成员变量名。而对于自定义元数据,可以通过eureka.instance.metadataMap.=的格式来进行配置,例如:
eureka.instance.metadataMap.zone=hangzhou
实例名,即InstanceInfo中的instanceId参数,它是区别同一服务中不同实例的唯一标识。在Netflix Eureka的原生实现中,实例名采用主机名作为默认值,这样的设置使得在同一主机上无法启动多个相同的服务实例。所以,在SpringCloudEureka的配置中,针对同一主机中启动多实例的情况,对实例名的默认命名做了更为合理的扩展,它采用了如下默认规则:
${spring.cloud.client.hostname}:${spring.application.name}:${spring.application.instance_id:${server.port}}
在InstanceInfo中,我们可以看到一些URL的配置信息,比如homePageUrl、statusPageUrl、healthCheckUrl,它们分别代表了应用主页的URL、状态页的URL、健康检查的URL。其中,状态页和健康检测的URL的Spring Cloud Eureka中默认使用了spring-boot-actuator模块提供的/info端点和/health端点。
大多数情况下,我们都不需要修改这个几个URL的配置,但是在需要特殊配置的时候,比如,为应用设置了context-path,这时所有spring-boot-actuator模块的监控端点都会增加一个前缀。那么在/info 和/health端点也加上类似的前缀信息
management.context-path=/hello
eureka.instance.statusPageUrlPath=${management.context-path}/info
eureka.instance.healthpageUrlPath=${management.context-path}/health
在上面的示例中,我们使用相对路径来进行配置,由于Eureka的服务注册中心默认会以HTTP的方式来访问和暴露这些端点,若客户端应用以HTTPS的方式来暴露服务和监控端点时,相对路径的配置方式就无法满足需求了,所以需要采用绝对路径的配置参数的方式:
eureka.instance.statusPageUrlPath=https://${eureka.instance.hostname}/info
eureka.instance.healthpageUrlPath=https://${eureka.instance.hostname}/health
eureka.instance.homePageUrl=https://${eureka.instance.hostname}/
默认情况下,Eureka中各个服务实例的健康检测并不是通过spring-boot-actuator模块的/health端点来实现的,而是依靠客户端心跳的方式来保持服务实例的存活。在Eureka的服务续约与剔除机制下,客户端的健康状态从注册到注册中心开始都会处于UP(准备接收通信)状态,除非心跳终止一段时间之后,服务注册中心将其剔除。默认的心跳实现方式可以有效检查客户端进程是否正常运作,但是却无法保证客户端应用能够正常的对外提供服务。
在Spring Cloud Eureka中,我们可以通过简单的配置把Eureka客户端的健康检测交给spring-boot-actuator模块的/health端点,以实现更加全面的健康状态维护。详细的配置步骤如下:
参数名 | 说明 | 默认值 |
---|---|---|
preferIpAddress | 是否优先使用IP地址作为主机名的标识 | false |
leaseRenewalIntervalInSeconds | Eureka客户端向服务端发送心跳的时间间隔,单位秒 | 30 |
leaseExpirationDurationInSeconds | Eureka服务端在收到最后一次心跳之后等待的时间上限,单位秒。超过该时间之后服务端会将该服务实例从服务清单中剔除,从而禁止服务调用请求被发送到该实例上 | 90 |
nonSecurePortEnabled | 是否启用非安全的通信端口号 | true |
securePortEnabled | 是否启用安全的通信端口号 | |
nonSecurePort | 非安全的通信端口号 | 80 |
securePort | 安全的通信端口号 | 443 |
appname | 服务名。默认取spring.application.name的配置值,如果没有则为unknow | |
hostname | 主机名,不配置的时候将根据操作系统的主机名来获取 |
参考文献:《Spring Cloud微服务实战》-- 翟永超 著