聊聊ribbon的ServerListSubsetFilter

序

本文主要研究一下 ribbon 的 ServerListSubsetFilter

ServerListSubsetFilter

ribbon-loadbalancer-2.3.0-sources.jar!/com/netflix/loadbalancer/ServerListSubsetFilter.java

public class ServerListSubsetFilter<T extends Server> extends ZoneAffinityServerListFilter<T> implements IClientConfigAware, Comparator<T>{private Random random = new Random();
    private volatile Set<T> currentSubset = Sets.newHashSet(); 
    private DynamicIntProperty sizeProp = new DynamicIntProperty(DefaultClientConfigImpl.DEFAULT_PROPERTY_NAME_SPACE + ".ServerListSubsetFilter.size", 20);
    private DynamicFloatProperty eliminationPercent = 
            new DynamicFloatProperty(DefaultClientConfigImpl.DEFAULT_PROPERTY_NAME_SPACE + ".ServerListSubsetFilter.forceEliminatePercent", 0.1f);
    private DynamicIntProperty eliminationFailureCountThreshold = 
            new DynamicIntProperty(DefaultClientConfigImpl.DEFAULT_PROPERTY_NAME_SPACE + ".ServerListSubsetFilter.eliminationFailureThresold", 0);
    private DynamicIntProperty eliminationConnectionCountThreshold = 
            new DynamicIntProperty(DefaultClientConfigImpl.DEFAULT_PROPERTY_NAME_SPACE + ".ServerListSubsetFilter.eliminationConnectionThresold", 0);
    
    @Override
    public void initWithNiwsConfig(IClientConfig clientConfig) {super.initWithNiwsConfig(clientConfig);
        sizeProp = new DynamicIntProperty(clientConfig.getClientName() + "." + clientConfig.getNameSpace() + ".ServerListSubsetFilter.size", 20);
        eliminationPercent = 
                new DynamicFloatProperty(clientConfig.getClientName() + "." + clientConfig.getNameSpace() + ".ServerListSubsetFilter.forceEliminatePercent", 0.1f);
        eliminationFailureCountThreshold = new DynamicIntProperty(clientConfig.getClientName() + "." + clientConfig.getNameSpace()
                + ".ServerListSubsetFilter.eliminationFailureThresold", 0);
        eliminationConnectionCountThreshold = new DynamicIntProperty(clientConfig.getClientName() + "." + clientConfig.getNameSpace()
                + ".ServerListSubsetFilter.eliminationConnectionThresold", 0);
    }
        
    /**
     * Given all the servers, keep only a stable subset of servers to use. This method
     * keeps the current list of subset in use and keep returning the same list, with exceptions
     * to relatively unhealthy servers, which are defined as the following:
     * <p>
     * <ul>
     * <li>Servers with their concurrent connection count exceeding the client configuration for 
     * {@code <clientName>.<nameSpace>.ServerListSubsetFilter.eliminationConnectionThresold} (default is 0)
     * <li>Servers with their failure count exceeding the client configuration for 
     * {@code <clientName>.<nameSpace>.ServerListSubsetFilter.eliminationFailureThresold} (default is 0)
     * <li>If the servers evicted above is less than the forced eviction percentage as defined by client configuration
     * {@code <clientName>.<nameSpace>.ServerListSubsetFilter.forceEliminatePercent} (default is 10%, or 0.1), the
     * remaining servers will be sorted by their health status and servers will worst health status will be
     * forced evicted.
     * </ul>
     * <p>
     * After the elimination, new servers will be randomly chosen from all servers pool to keep the
     * number of the subset unchanged. 
     * 
     */
    @Override
    public List<T> getFilteredListOfServers(List<T> servers) {List<T> zoneAffinityFiltered = super.getFilteredListOfServers(servers);
        Set<T> candidates = Sets.newHashSet(zoneAffinityFiltered);
        Set<T> newSubSet = Sets.newHashSet(currentSubset);
        LoadBalancerStats lbStats = getLoadBalancerStats();
        for (T server: currentSubset) {
            // this server is either down or out of service
            if (!candidates.contains(server)) {newSubSet.remove(server);
            } else {ServerStats stats = lbStats.getSingleServerStat(server);
                // remove the servers that do not meet health criteria
                if (stats.getActiveRequestsCount() > eliminationConnectionCountThreshold.get()
                        || stats.getFailureCount() > eliminationFailureCountThreshold.get()) {newSubSet.remove(server);
                    // also remove from the general pool to avoid selecting them again
                    candidates.remove(server);
                }
            }
        }
        int targetedListSize = sizeProp.get();
        int numEliminated = currentSubset.size() - newSubSet.size();
        int minElimination = (int) (targetedListSize * eliminationPercent.get());
        int numToForceEliminate = 0;
        if (targetedListSize < newSubSet.size()) {
            // size is shrinking
            numToForceEliminate = newSubSet.size() - targetedListSize;} else if (minElimination > numEliminated) {numToForceEliminate = minElimination - numEliminated;}
        
        if (numToForceEliminate > newSubSet.size()) {numToForceEliminate = newSubSet.size();
        }

        if (numToForceEliminate > 0) {List<T> sortedSubSet = Lists.newArrayList(newSubSet);           
            Collections.sort(sortedSubSet, this);
            List<T> forceEliminated = sortedSubSet.subList(0, numToForceEliminate);
            newSubSet.removeAll(forceEliminated);
            candidates.removeAll(forceEliminated);
        }
        
        // after forced elimination or elimination of unhealthy instances,
        // the size of the set may be less than the targeted size,
        // then we just randomly add servers from the big pool
        if (newSubSet.size() < targetedListSize) {int numToChoose = targetedListSize - newSubSet.size();
            candidates.removeAll(newSubSet);
            if (numToChoose > candidates.size()) {
                // Not enough healthy instances to choose, fallback to use the
                // total server pool
                candidates = Sets.newHashSet(zoneAffinityFiltered);
                candidates.removeAll(newSubSet);
            }
            List<T> chosen = randomChoose(Lists.newArrayList(candidates), numToChoose);
            for (T server: chosen) {newSubSet.add(server);
            }
        }
        currentSubset = newSubSet;       
        return Lists.newArrayList(newSubSet);            
    }

    /**
     * Randomly shuffle the beginning portion of server list (according to the number passed into the method) 
     * and return them.
     *  
     * @param servers
     * @param toChoose
     * @return
     */
    private List<T> randomChoose(List<T> servers, int toChoose) {int size = servers.size();
        if (toChoose >= size || toChoose < 0) {return servers;} 
        for (int i = 0; i < toChoose; i++) {int index = random.nextInt(size);
            T tmp = servers.get(index);
            servers.set(index, servers.get(i));
            servers.set(i, tmp);
        }
        return servers.subList(0, toChoose);        
    }

    /**
     * Function to sort the list by server health condition, with
     * unhealthy servers before healthy servers. The servers are first sorted by
     * failures count, and then concurrent connection count.
     */
    @Override
    public int compare(T server1, T server2) {LoadBalancerStats lbStats = getLoadBalancerStats();
        ServerStats stats1 = lbStats.getSingleServerStat(server1);
        ServerStats stats2 = lbStats.getSingleServerStat(server2);
        int failuresDiff = (int) (stats2.getFailureCount() - stats1.getFailureCount());
        if (failuresDiff != 0) {return failuresDiff;} else {return (stats2.getActiveRequestsCount() - stats1.getActiveRequestsCount());
        }
    }
}

• ServerListSubsetFilter 继承了 ZoneAffinityServerListFilter，实现了 IClientConfigAware、Comparator 接口
• initWithNiwsConfig 方法从 IClientConfig 读取了 ServerListSubsetFilter.size、ServerListSubsetFilter.forceEliminatePercent、ServerListSubsetFilter.eliminationFailureThresold、ServerListSubsetFilter.eliminationConnectionThresold 配置
• getFilteredListOfServers 方法会先调用父类 ZoneAffinityServerListFilter 的 getFilteredListOfServers 先过滤出 zoneAffinityFiltered 作为 candidates，然后遍历 currentSubset 根据 ServerStats 剔除 activeRequestsCount 及 failureCount 超出阈值的 server；然后根据 numToForceEliminate，以及 failureCount 排序，使用 subList 方法得出 forceEliminated，然后移除掉，最后 randomChoose 出来新的 newSubSet，然后重置 currentSubset

小结

在 server list 非常多的场景下，没有必要在连接池的保持这么多的连接，ServerListSubsetFilter 可以在这种场景下对 server list 进行精简，通过剔除相对不健康 (failureCount、activeRequestCount) 的 server 来达到此目标

doc

• ServerListSubsetFilter