Ribbon的BestAvailableRule和RetryRule

Ribbon的版本是2.3.0.release.

1.BestAvailableRule

Ribbon的BestAvailableRule和RetryRule-LMLPHP

图1

ClientConfigEnabledRoundRobinRule如下所示，定义了一个类属性RoundRobinRule，choose方法中调用RoundRobinRule进行选择，所以这里面的是轮循算法。

List-1.1

public class ClientConfigEnabledRoundRobinRule extends AbstractLoadBalancerRule {
    RoundRobinRule roundRobinRule = new RoundRobinRule();

    public ClientConfigEnabledRoundRobinRule() {
    }

    public void initWithNiwsConfig(IClientConfig clientConfig) {
        this.roundRobinRule = new RoundRobinRule();
    }

    public void setLoadBalancer(ILoadBalancer lb) {
        super.setLoadBalancer(lb);
        this.roundRobinRule.setLoadBalancer(lb);
    }

    public Server choose(Object key) {
        if (this.roundRobinRule != null) {
            return this.roundRobinRule.choose(key);
        } else {
            throw new IllegalArgumentException("This class has not been initialized with the RoundRobinRule class");
        }
    }
}

BestAvailableRule继承了ClientConfigEnabledRoundRobinRule，内部实现如下List-1.2，遍历所有的服务提供者，选择并发量最小的那个服务。

List-1.2

private LoadBalancerStats loadBalancerStats;

public Server choose(Object key) {
    if (this.loadBalancerStats == null) {
        return super.choose(key);
    } else {
        List<Server> serverList = this.getLoadBalancer().getAllServers();
        int minimalConcurrentConnections = 2147483647;
        long currentTime = System.currentTimeMillis();
        Server chosen = null;
        Iterator var7 = serverList.iterator();

        while(var7.hasNext()) {
            Server server = (Server)var7.next();
            ServerStats serverStats = this.loadBalancerStats.getSingleServerStat(server);
            if (!serverStats.isCircuitBreakerTripped(currentTime)) {
                int concurrentConnections = serverStats.getActiveRequestsCount(currentTime);
                if (concurrentConnections < minimalConcurrentConnections) {
                    minimalConcurrentConnections = concurrentConnections;
                    chosen = server;
                }
            }
        }

        if (chosen == null) {
            return super.choose(key);
        } else {
            return chosen;
        }
    }
}

public void setLoadBalancer(ILoadBalancer lb) {
    super.setLoadBalancer(lb);
    if (lb instanceof AbstractLoadBalancer) {
        this.loadBalancerStats = ((AbstractLoadBalancer)lb).getLoadBalancerStats();
    }

}

choose方法重新了父类中的choose方法，

获取服务列表，遍历服务
通过ServerStats获取当前服务实例的并发连接数，如下List-3所示，并发连接数不是0，且当前时间与上次有效更改时间间隔在范围内，则返回当前并发连接数。
遍历所有的服务提供者后，如果得到的server是null，则调用父类的choose方法，用RoundRobin算法进行选择。

List-1.3

public int getActiveRequestsCount(long currentTime) {
    int count = this.activeRequestsCount.get();
    if (count == 0) {
        return 0;
    } else if (currentTime - this.lastActiveRequestsCountChangeTimestamp <= (long)(activeRequestsCountTimeout.get() * 1000) && count >= 0) {
        return count;
    } else {
        this.activeRequestsCount.set(0);
        return 0;
    }
}

2.RetryRule

Ribbon的BestAvailableRule和RetryRule-LMLPHP

图2

RetryRule的实现比较简单，基于RoundRobinRule，如下List-2.1所示

List-2.1

public class RetryRule extends AbstractLoadBalancerRule {
    IRule subRule = new RoundRobinRule();
    long maxRetryMillis = 500L;
    ...
    public Server choose(ILoadBalancer lb, Object key) {
        long requestTime = System.currentTimeMillis();
        long deadline = requestTime + this.maxRetryMillis;
        Server answer = null;
        answer = this.subRule.choose(key);
        if ((answer == null || !answer.isAlive()) && System.currentTimeMillis() < deadline) {
            InterruptTask task = new InterruptTask(deadline - System.currentTimeMillis());

            while(!Thread.interrupted()) {
                answer = this.subRule.choose(key);
                if (answer != null && answer.isAlive() || System.currentTimeMillis() >= deadline) {
                    break;
                }

                Thread.yield();
            }

            task.cancel();
        }

        return answer != null && answer.isAlive() ? answer : null;
    }
    ...

RetryRule中choose(Object key)调用choose(ILoadBalancer lb, Object key)，

当前时间加上maxRetryMillis得到deadline，即截止时间
用subRule获取服务server，如果服务有效则直接返回该服务
构造InterruptTask，它里面有个Timer定时任务，如List-2.2。之后循坏，只要当前线程没有被interrupt，则用subRule的RoundRobin算法选择一个服务实例，如果这个服务有效或者当前时间过了截止时间则跳出循坏
如果步骤３中得到的服务实例无效，且当前时间在截止时间之内，则调用Thread.yield()，让出线程资源给其它线程

通过源码可知，RetryRule在subRule.choose获得无效的服务实例后，仅仅是再次再次用subRule获取服务实例，并不会一直尝试下去，即尝试一次。

List-2.2

public class InterruptTask extends TimerTask {
    static Timer timer = new Timer("InterruptTimer", true);
    protected Thread target = null;

    public InterruptTask(long millis) {
        this.target = Thread.currentThread();
        timer.schedule(this, millis);
    }

    public boolean cancel() {
        try {
            return super.cancel();
        } catch (Exception var2) {
            return false;
        }
    }

    public void run() {
        if (this.target != null && this.target.isAlive()) {
            this.target.interrupt();
        }
    }
    ...

Reference

https://github.com/Netflix/ribbon/tree/v2.3.0

克虏伯