golang mgo的mongo连接池设置：必须手动加上maxPoolSize

q987654

　　本司礼物系统使用了golang的 mongo库 mgo，中间踩了一些坑，总结下避免大家再踩坑
　　golang的mgo库说明里是说明了开启连接复用的，但观察实验发现，这并没有根本实现连接的控制，连接复用仅在有空闲连接时生效，高并发时无可用连接会不断创建新连接，所以最终还是需要程序员自行去限制最大连接才行。
　　废话不多说，开始上代码
GlobalMgoSession, err := mgo.Dial(host) func (m *MongoBaseDao) Get(tablename string,>interface{}) interface{} {    session := GlobalMgoSession.Clone()    defer session.Close()     collection := session.DB(globalMgoDbName).C(tablename)    err := collection.FindId(bson.ObjectIdHex(id)).One(result)     if err != nil {        logkit.Logger.Error("mongo_base method:Get " + err.Error())    }    return result}　　golang main入口启动时，我们会创建一个全局session，然后每次使用时clone session的信息和连接，用于本次请求，使用后调用session.Close() 释放连接。
// Clone works just like Copy, but also reuses the same socket as the original// session, in case it had already reserved one due to its consistency// guarantees.  This behavior ensures that writes performed in the old session// are necessarily observed when using the new session, as long as it was a// strong or monotonic session.  That said, it also means that long operations// may cause other goroutines using the original session to wait.func (s *Session) Clone() *Session {    s.m.Lock()    scopy := copySession(s, true)    s.m.Unlock()    return scopy}  // Close terminates the session.  It's a runtime error to use a session// after it has been closed.func (s *Session) Close() {    s.m.Lock()    if s.cluster_ != nil {        debugf("Closing session %p", s)        s.unsetSocket()  //释放当前线程占用的socket 置为nil        s.cluster_.Release()        s.cluster_ = nil    }    s.m.Unlock()}　　Clone的方法注释里说明会重用原始session的socket连接，但是并发请求一大，其他协程来不及释放连接，当前协程会怎么办？
func (s *Session) acquireSocket(slaveOk bool) (*mongoSocket, error) {    // Read-only lock to check for previously reserved socket.    s.m.RLock()    // If there is a slave socket reserved and its use is acceptable, take it as long    // as there isn't a master socket which would be preferred by the read preference mode.    if s.slaveSocket != nil && s.slaveOk && slaveOk && (s.masterSocket == nil || s.consistency != PrimaryPreferred && s.consistency != Monotonic) {        socket := s.slaveSocket        socket.Acquire()        s.m.RUnlock()        logkit.Logger.Info("sgp_test 1 acquireSocket slave is ok!")        return socket, nil    }    if s.masterSocket != nil {        socket := s.masterSocket        socket.Acquire()        s.m.RUnlock()        logkit.Logger.Info("sgp_test 1  acquireSocket master is ok!")        return socket, nil    }     s.m.RUnlock()     // No go.  We may have to request a new socket and change the session,    // so try again but with an exclusive lock now.    s.m.Lock()    defer s.m.Unlock()    if s.slaveSocket != nil && s.slaveOk && slaveOk && (s.masterSocket == nil || s.consistency != PrimaryPreferred && s.consistency != Monotonic) {        s.slaveSocket.Acquire()        logkit.Logger.Info("sgp_test 2  acquireSocket slave is ok!")        return s.slaveSocket, nil    }    if s.masterSocket != nil {        s.masterSocket.Acquire()        logkit.Logger.Info("sgp_test 2  acquireSocket master is ok!")        return s.masterSocket, nil    }     // Still not good.  We need a new socket.    sock, err := s.cluster().AcquireSocket(s.consistency, slaveOk && s.slaveOk, s.syncTimeout, s.sockTimeout, s.queryConfig.op.serverTags, s.poolLimit) ......    logkit.Logger.Info("sgp_test 3   acquireSocket cluster AcquireSocket is ok!")    return sock, nil }　　在源码中加debug，结果日志说明一切：
Mar 25 09:46:40 dev02.pandatv.com bikini[12607]:  [info] sgp_test 1  acquireSocket master is ok!Mar 25 09:46:40 dev02.pandatv.com bikini[12607]:  [info] sgp_test 1  acquireSocket master is ok!Mar 25 09:46:41 dev02.pandatv.com bikini[12607]:  [info] sgp_test 1 acquireSocket slave is ok!Mar 25 09:46:41 dev02.pandatv.com bikini[12607]:  [info] sgp_test 3   acquireSocket cluster AcquireSocket is ok!Mar 25 09:46:41 dev02.pandatv.com bikini[12607]:  [info] sgp_test 3   acquireSocket cluster AcquireSocket is ok!Mar 25 09:46:41 dev02.pandatv.com bikini[12607]:  [info] sgp_test 3   acquireSocket cluster AcquireSocket is ok!　　不断的创建连接  AcquireSocket
　　$ netstat -nat|grep -i 27017|wc -l
　　400
　　如果每个session 不调用close，会达到恐怖的4096,并堵死其他请求，所以clone或copy session时一定要defer close掉
　　启用maxPoolLimit 参数则会限制总连接大小，连接到限制则当前协程会sleep等待  直到可以创建连接，高并发时锁有问题，会导致多创建几个连接
src/gopkg.in/mgo.v2/cluster.go     s, abended, err := server.AcquireSocket(poolLimit, socketTimeout)        if err == errPoolLimit {            if !warnedLimit {                warnedLimit = true                logkit.Logger.Error("sgp_test WARNING: Per-server connection limit reached. " + err.Error())                log("WARNING: Per-server connection limit reached.")            }            time.Sleep(100 * time.Millisecond)            continue        } session.go:// SetPoolLimit sets the maximum number of sockets in use in a single server  // before this session will block waiting for a socket to be available.  // The default limit is 4096.  //  // This limit must be set to cover more than any expected workload of the  // application. It is a bad practice and an unsupported use case to use the  // database driver to define the concurrency limit of an application. Prevent  // such concurrency "at the door" instead, by properly restricting the amount  // of used resources and number of goroutines before they are created.  func (s *Session) SetPoolLimit(limit int) {      s.m.Lock()      s.poolLimit = limit      s.m.Unlock()  }　　连接池设置方法：
　　1、配置中 增加
　　[host]:[port]?maxPoolSize=10
　　2、代码中 ：
　　dao.GlobalMgoSession.SetPoolLimit(10)
　　再做压测：
　　$ netstat -nat|grep -i 27017|wc -l
　　15
　　结论：
　　每次clone session之后，操作结束时如果调用 session.Close 则会unset Socket  ，socket refer数减少，如果不设置上限，每个协程请求到来发现无空闲连接就会创建socket连接，直到达到最大值4096，而mongo的连接数上限一般也就是1万，也就是一个端口你只能启动一两个进程保证连接不被撑爆，过多的连接数客户端效率不高，server端更会耗费内存和CPU，所以需要启用自定义连接池 , 启用连接池也需要注意如果有pooMaxLimit个协程执行过长或者死循环不释放socket连接，也会悲剧。
　　mgo底层socket连接池只在maxPooMaxLimit 范围内实现复用，需要自行优化。