dubbo的负载均衡全部由AbstractLoadBalance的子类来实现
在一个截面上碰撞的概率高,但调用量越大分布越均匀,而且按概率使用权重后也比较均匀,有利于动态调整提供者权重。
@Override protected <T> Invoker<T> doSelect(List<Invoker<T>> invokers, URL url, Invocation invocation) { int length = invokers.size(); // Number of invokers boolean sameWeight = true; // Every invoker has the same weight? int firstWeight = getWeight(invokers.get(0), invocation); int totalWeight = firstWeight; // The sum of weights for (int i = 1; i < length; i++) { int weight = getWeight(invokers.get(i), invocation); totalWeight += weight; // Sum if (sameWeight && weight != firstWeight) { sameWeight = false; } } if (totalWeight > 0 && !sameWeight) { // If (not every invoker has the same weight & at least one invoker's weight>0), select randomly based on totalWeight. int offset = ThreadLocalRandom.current().nextInt(totalWeight); // Return a invoker based on the random value. for (int i = 0; i < length; i++) { offset -= getWeight(invokers.get(i), invocation); if (offset < 0) { return invokers.get(i); } } } // If all invokers have the same weight value or totalWeight=0, return evenly. return invokers.get(ThreadLocalRandom.current().nextInt(length)); }
存在慢的提供者累积请求的问题,比如:第二台机器很慢,但没挂,当请求调到第二台时就卡在那,久而久之,所有请求都卡在调到第二台上。
在老的版本上,dubbo会求出最大权重和最小权重,如果权重相等,那么就直接按取模的方式,每次取完后值加一;如果权重不相等,顺序根据权重分配。
在新的版本上,对这个类进行了重构。
这样看显然是不够清晰的,我们来举个例子:
假定有3台dubbo provider: 10.0.0.1:20884, weight=2 10.0.0.1:20886, weight=3 10.0.0.1:20888, weight=4 totalWeight=9; 那么第一次调用的时候: 10.0.0.1:20884, weight=2 selectedWRR -> current = 2 10.0.0.1:20886, weight=3 selectedWRR -> current = 3 10.0.0.1:20888, weight=4 selectedWRR -> current = 4 selectedInvoker-> 10.0.0.1:20888 调用 selectedWRR.sel(totalWeight); 10.0.0.1:20888, weight=4 selectedWRR -> current = -5 返回10.0.0.1:20888这个实例 那么第二次调用的时候: 10.0.0.1:20884, weight=2 selectedWRR -> current = 4 10.0.0.1:20886, weight=3 selectedWRR -> current = 6 10.0.0.1:20888, weight=4 selectedWRR -> current = -1 selectedInvoker-> 10.0.0.1:20886 调用 selectedWRR.sel(totalWeight); 10.0.0.1:20886 , weight=4 selectedWRR -> current = -3 返回10.0.0.1:20886这个实例 那么第三次调用的时候: 10.0.0.1:20884, weight=2 selectedWRR -> current = 6 10.0.0.1:20886, weight=3 selectedWRR -> current = 0 10.0.0.1:20888, weight=4 selectedWRR -> current = 3 selectedInvoker-> 10.0.0.1:20884 调用 selectedWRR.sel(totalWeight); 10.0.0.1:20884, weight=2 selectedWRR -> current = -3 返回10.0.0.1:20884这个实例 protected <T> Invoker<T> doSelect(List<Invoker<T>> invokers, URL url, Invocation invocation) { String key = invokers.get(0).getUrl().getServiceKey() + "." + invocation.getMethodName(); ConcurrentMap<String, WeightedRoundRobin> map = methodWeightMap.get(key); if (map == null) { methodWeightMap.putIfAbsent(key, new ConcurrentHashMap<String, WeightedRoundRobin>()); map = methodWeightMap.get(key); } int totalWeight = 0; long maxCurrent = Long.MIN_VALUE; long now = System.currentTimeMillis(); Invoker<T> selectedInvoker = null; WeightedRoundRobin selectedWRR = null; for (Invoker<T> invoker : invokers) { String identifyString = invoker.getUrl().toIdentityString(); WeightedRoundRobin weightedRoundRobin = map.get(identifyString); int weight = getWeight(invoker, invocation); if (weight < 0) { weight = 0; } if (weightedRoundRobin == null) { weightedRoundRobin = new WeightedRoundRobin(); weightedRoundRobin.setWeight(weight); map.putIfAbsent(identifyString, weightedRoundRobin); weightedRoundRobin = map.get(identifyString); } if (weight != weightedRoundRobin.getWeight()) { //weight changed weightedRoundRobin.setWeight(weight); } long cur = weightedRoundRobin.increaseCurrent(); weightedRoundRobin.setLastUpdate(now); if (cur > maxCurrent) { maxCurrent = cur; selectedInvoker = invoker; selectedWRR = weightedRoundRobin; } totalWeight += weight; } if (!updateLock.get() && invokers.size() != map.size()) { if (updateLock.compareAndSet(false, true)) { try { // copy -> modify -> update reference ConcurrentMap<String, WeightedRoundRobin> newMap = new ConcurrentHashMap<String, WeightedRoundRobin>(); newMap.putAll(map); Iterator<Entry<String, WeightedRoundRobin>> it = newMap.entrySet().iterator(); while (it.hasNext()) { Entry<String, WeightedRoundRobin> item = it.next(); if (now - item.getValue().getLastUpdate() > RECYCLE_PERIOD) { it.remove(); } } methodWeightMap.put(key, newMap); } finally { updateLock.set(false); } } } if (selectedInvoker != null) { selectedWRR.sel(totalWeight); return selectedInvoker; } // should not happen here return invokers.get(0); }
使慢的提供者收到更少请求,因为越慢的提供者的调用前后计数差会越大。
最小活跃数算法实现: 假定有3台dubbo provider: 10.0.0.1:20884, weight=2,active=2 10.0.0.1:20886, weight=3,active=4 10.0.0.1:20888, weight=4,active=3 active=2最小,且只有一个2,所以选择10.0.0.1:20884 假定有3台dubbo provider: 10.0.0.1:20884, weight=2,active=2 10.0.0.1:20886, weight=3,active=2 10.0.0.1:20888, weight=4,active=3 active=2最小,且有2个,所以从[10.0.0.1:20884,10.0.0.1:20886 ]中选择; 接下来的算法与随机算法类似: 假设offset=1(即random.nextInt(5)=1) 1-2=-1<0?是,所以选中 10.0.0.1:20884, weight=2 假设offset=4(即random.nextInt(5)=4) 4-2=2<0?否,这时候offset=2, 2-3<0?是,所以选中 10.0.0.1:20886, weight=3 1: public class LeastActiveLoadBalance extends AbstractLoadBalance { 2: 3: public static final String NAME = "leastactive"; 4: 5: private final Random random = new Random(); 6: 7: @Override 8: protected <T> Invoker<T> doSelect(List<Invoker<T>> invokers, URL url, Invocation invocation) { 9: int length = invokers.size(); // 总个数 10: int leastActive = -1; // 最小的活跃数 11: int leastCount = 0; // 相同最小活跃数的个数 12: int[] leastIndexes = new int[length]; // 相同最小活跃数的下标 13: int totalWeight = 0; // 总权重 14: int firstWeight = 0; // 第一个权重,用于于计算是否相同 15: boolean sameWeight = true; // 是否所有权重相同 16: // 计算获得相同最小活跃数的数组和个数 17: for (int i = 0; i < length; i++) { 18: Invoker<T> invoker = invokers.get(i); 19: int active = RpcStatus.getStatus(invoker.getUrl(), invocation.getMethodName()).getActive(); // 活跃数 20: int weight = invoker.getUrl().getMethodParameter(invocation.getMethodName(), Constants.WEIGHT_KEY, Constants.DEFAULT_WEIGHT); // 权重 21: if (leastActive == -1 || active < leastActive) { // 发现更小的活跃数,重新开始 22: leastActive = active; // 记录最小活跃数 23: leastCount = 1; // 重新统计相同最小活跃数的个数 24: leastIndexes[0] = i; // 重新记录最小活跃数下标 25: totalWeight = weight; // 重新累计总权重 26: firstWeight = weight; // 记录第一个权重 27: sameWeight = true; // 还原权重相同标识 28: } else if (active == leastActive) { // 累计相同最小的活跃数 29: leastIndexes[leastCount++] = i; // 累计相同最小活跃数下标 30: totalWeight += weight; // 累计总权重 31: // 判断所有权重是否一样 32: if (sameWeight && weight != firstWeight) { 33: sameWeight = false; 34: } 35: } 36: } 37: // assert(leastCount > 0) 38: if (leastCount == 1) { 39: // 如果只有一个最小则直接返回 40: return invokers.get(leastIndexes[0]); 41: } 42: if (!sameWeight && totalWeight > 0) { 43: // 如果权重不相同且权重大于0则按总权重数随机 44: int offsetWeight = random.nextInt(totalWeight); 45: // 并确定随机值落在哪个片断上 46: for (int i = 0; i < leastCount; i++) { 47: int leastIndex = leastIndexes[i]; 48: offsetWeight -= getWeight(invokers.get(leastIndex), invocation); 49: if (offsetWeight <= 0) { 50: return invokers.get(leastIndex); 51: } 52: } 53: } 54: // 如果权重相同或权重为0则均等随机 55: return invokers.get(leastIndexes[random.nextInt(leastCount)]); 56: } 57: 58: }
相同参数的请求总是发到同一提供者。当某一台提供者挂时,原本发往该提供者的请求,基于虚拟节点,平摊到其它提供者,不会引起剧烈变动。
1: public class ConsistentHashLoadBalance extends AbstractLoadBalance { 2: 3: /** 4: * 服务方法与一致性哈希选择器的映射 5: * 6: * KEY:serviceKey + "." + methodName 7: */ 8: private final ConcurrentMap<String, ConsistentHashSelector<?>> selectors = new ConcurrentHashMap<String, ConsistentHashSelector<?>>(); 9: 10: @SuppressWarnings("unchecked") 11: @Override 12: protected <T> Invoker<T> doSelect(List<Invoker<T>> invokers, URL url, Invocation invocation) { 13: String key = invokers.get(0).getUrl().getServiceKey() + "." + invocation.getMethodName(); 14: // 基于 invokers 集合,根据对象内存地址来计算定义哈希值 15: int identityHashCode = System.identityHashCode(invokers); 16: // 获得 ConsistentHashSelector 对象。若为空,或者定义哈希值变更(说明 invokers 集合发生变化),进行创建新的 ConsistentHashSelector 对象 17: ConsistentHashSelector<T> selector = (ConsistentHashSelector<T>) selectors.get(key); 18: if (selector == null || selector.identityHashCode != identityHashCode) { 19: selectors.put(key, new ConsistentHashSelector<T>(invokers, invocation.getMethodName(), identityHashCode)); 20: selector = (ConsistentHashSelector<T>) selectors.get(key); 21: } 22: return selector.select(invocation); 23: } 24: }
ConsistentHashSelector ,是 ConsistentHashLoadBalance 的内部类,一致性哈希选择器,基于 Ketama 算法。
/** * 虚拟节点与 Invoker 的映射关系 */ private final TreeMap<Long, Invoker<T>> virtualInvokers; /** * 每个Invoker 对应的虚拟节点数 */ private final int replicaNumber; /** * 定义哈希值 */ private final int identityHashCode; /** * 取值参数位置数组 */ private final int[] argumentIndex; 1: ConsistentHashSelector(List<Invoker<T>> invokers, String methodName, int identityHashCode) { 2: this.virtualInvokers = new TreeMap<Long, Invoker<T>>(); 3: // 设置 identityHashCode 4: this.identityHashCode = identityHashCode; 5: URL url = invokers.get(0).getUrl(); 6: // 初始化 replicaNumber 7: this.replicaNumber = url.getMethodParameter(methodName, "hash.nodes", 160); 8: // 初始化 argumentIndex 9: String[] index = Constants.COMMA_SPLIT_PATTERN.split(url.getMethodParameter(methodName, "hash.arguments", "0")); 10: argumentIndex = new int[index.length]; 11: for (int i = 0; i < index.length; i++) { 12: argumentIndex[i] = Integer.parseInt(index[i]); 13: } 14: // 初始化 virtualInvokers 15: for (Invoker<T> invoker : invokers) { 16: String address = invoker.getUrl().getAddress(); 17: // 每四个虚拟结点为一组,为什么这样?下面会说到 18: for (int i = 0; i < replicaNumber / 4; i++) { 19: // 这组虚拟结点得到惟一名称 20: byte[] digest = md5(address + i); 21: // Md5是一个16字节长度的数组,将16字节的数组每四个字节一组,分别对应一个虚拟结点,这就是为什么上面把虚拟结点四个划分一组的原因 22: for (int h = 0; h < 4; h++) { 23: // 对于每四个字节,组成一个long值数值,做为这个虚拟节点的在环中的惟一key 24: long m = hash(digest, h); 25: virtualInvokers.put(m, invoker); 26: } 27: } 28: } 29: } public Invoker<T> select(Invocation invocation) { // 基于方法参数,获得 KEY String key = toKey(invocation.getArguments()); // 计算 MD5 值 byte[] digest = md5(key); // 计算 KEY 值 return selectForKey(hash(digest, 0)); } private String toKey(Object[] args) { StringBuilder buf = new StringBuilder(); for (int i : argumentIndex) { if (i >= 0 && i < args.length) { buf.append(args[i]); } } return buf.toString(); } private Invoker<T> selectForKey(long hash) { // 得到大于当前 key 的那个子 Map ,然后从中取出第一个 key ,就是大于且离它最近的那个 key Map.Entry<Long, Invoker<T>> entry = virtualInvokers.tailMap(hash, true).firstEntry(); // 不存在,则取 virtualInvokers 第一个 if (entry == null) { entry = virtualInvokers.firstEntry(); } // 存在,则返回 return entry.getValue(); }