一般海量处理服务,都会对服务做个最大连接数限制,超过该限制之后,拒绝服务,避免发生雪崩,压坏服务。
使用golang来编写httpserver时,如何进行呢?官方已经有实现好的包。
使用示例:
import(
"golang.org/x/net/netutil"
)
l, err := net.Listen("tcp", "127.0.0.1:0") if err != nil { fmt.Fatalf("Listen: %v", err) } defer l.Close() l = LimitListener(l, max) http.Serve(l, http.HandlerFunc()) //bla bla bla.................
源码如下(url : https://github.com/golang/net/blob/master/netutil/listen.go),基本思路就是为连接数计数,通过make chan来建立一个最大连接数的channel, 每次accept就+1,close时候就-1. 当到达最大连接数时,就等待空闲连接出来之后再accept。
// Copyright 2013 The Go Authors. All rights reserved. // Use of this source code is governed by a BSD-style // license that can be found in the LICENSE file. // Package netutil provides network utility functions, complementing the more // common ones in the net package. package netutil // import "golang.org/x/net/netutil" import ( "net" "sync" ) // LimitListener returns a Listener that accepts at most n simultaneous // connections from the provided Listener. func LimitListener(l net.Listener, n int) net.Listener { return &limitListener{l, make(chan struct{}, n)} } type limitListener struct { net.Listener sem chan struct{} } func (l *limitListener) acquire() { l.sem <- struct{}{} } func (l *limitListener) release() { <-l.sem } func (l *limitListener) Accept() (net.Conn, error) {
//若connect chan已满,则会阻塞在此处 l.acquire() c, err := l.Listener.Accept() if err != nil { l.release() return nil, err } return &limitListenerConn{Conn: c, release: l.release}, nil } type limitListenerConn struct { net.Conn releaseOnce sync.Once release func() } func (l *limitListenerConn) Close() error { err := l.Conn.Close() l.releaseOnce.Do(l.release) return err }
有疑问加站长微信联系(非本文作者)