在TCP网络通信中,经常会出现客户端和服务器之间的非正常断开,需要实时检测查询链接状态。常用的解决方法就是在程序中加入心跳机制。
从网络上找到,主要有以下方法实现TCP异常断开的检测:
SO_KEEPALIVE,SIO_KEEPALIVE_VALS 和Heart-Beat线程。
(1)SO_KEEPALIVE 机制
一、SO_KEEPALIVE无法控制,它会每时每刻都发;
二、SO_KEEPALIVE设置空闲2小时才发送一个“保持存活探测分节”,不能保证实时检测。对于判断网络断开时间太长,对于需要及时响应的程序不太适应。
(2)SIO_KEEPALIVE_VALS 机制
(3)Heart-Beat线程
UNIX网络编程不推荐使用SO_KEEPALIVE来做心跳检测,还是在业务层以心跳包做检测比较好,也方便控制,楼上给的代码是UNIX下的,WIN下不适用。Windows下此处的”非正常断开”指TCP连接不是以优雅的方式断开,如网线故障等物理链路的原因,还有突然主机断电等原因。
有两种方法可以检测:1.TCP连接双方定时发握手消息 2.利用TCP协议栈中的KeepAlive探测
第二种方法简单可靠,只需对TCP连接两个Socket设定KeepAlive探测。
第一种方法是自己给SOCKET设置个timeout,如果超时则认为对方客户端断掉,需重新连接.
//处理通信
void do_client_work(int client_sock,struct sockaddr_in client)
{
unsigned char cmd;
int val;
char *VideoPtr;
send_client_info(client_sock,"hello",5);
TCP_VIDEO_AGAIN_FLAG = 1;
printf("\nrecv start ...\n");
while(1) {
fd_set rds;
struct timeval step;
int ret;
FD_ZERO(&rds);
FD_SET(client_sock,&rds);
step.tv_sec = 0;
step.tv_usec = 1000;
ret = select(client_sock+1,&rds,NULL,NULL,&step);
if(ret < 0) {
perror("select");
exit(-1);
}
if(ret == 0) {
...//执行等待超时命令
} else if(FD_ISSET(client_sock,&rds)) {
val = recv_client_info(client_sock,&cmd,1);
if(val < 0) {
break;
} else {
...//执行读写客户端数据操作
if(TcpOutFlag==1) {
TcpOutFlag=0;
break;
}
}
}
}
}
//TCP接收数据
int recv_client_info(int client_sock,char* info,int length)
{
int num;
struct timeval tv;
tv.tv_sec = 15;
tv.tv_usec = 0;
if (setsockopt(client_sock, SOL_SOCKET, SO_RCVTIMEO, &tv, sizeof(tv))>=0) {
num = recv(client_sock,info,length,MSG_NOSIGNAL);
if(num < 0) {
perror("recv");
return 0;
} else if(num == 0) {
printf("recv over\n");
return -2;
} else {
pthread_mutex_lock(&mutex);
TCP_Timeout = 0;
pthread_mutex_unlock(&mutex);
return 0;
}
} else {
printf("TCP CONNECT超时...\n");
return -1;
}
}
第二种方法在Linux,Window2000下的实现Windows 2000平台下
头文件
#include <mstcpip.h>
//定义结构及宏
struct TCP_KEEPALIVE {
u_longonoff;
u_longkeepalivetime;
u_longkeepaliveinterval;
} ;
tcp_keepalive live,liveout;
live.keepaliveinterval=500;
live.keepalivetime=3000;
live.onoff=TRUE;
int iRet = setsockopt(Socket,SOL_SOCKET,SO_KEEPALIVE,(char *)Opt,sizeof(int));
if(iRet == 0){
DWORD dw;
if(WSAIoctl(Socket,SIO_KEEPALIVE_VALS,&live,sizeof(live),&liveout,sizeof(liveout),&dw,NULL,NULL)== SOCKET_ERROR){
//Delete Client
return;
}
}
ACE下代码
int Opt = 1;
//在测试过程中,发现检测的次数是5次,即下面的设置中,从最近一次消息开始计算的10秒后,每次间隔5秒,连续发送5次,即
//35秒发现网络断了
tcp_keepalive live,liveout;
live.keepaliveinterval=5000; //每次检测的间隔 (单位毫秒)
live.keepalivetime=10000; //第一次开始发送的时间(单位毫秒)
live.onoff=TRUE;
int iRet = stream.set_option(SOL_SOCKET,SO_KEEPALIVE,&Opt,sizeof(int));
if(iRet == 0){
DWORD dw;
//此处显示了在ACE下获取套接字的方法,即句柄的(SOCKET)化就是句柄
if(WSAIoctl((SOCKET)h,SIO_KEEPALIVE_VALS,&live,sizeof(live),&liveout,sizeof(liveout),&dw,NULL,NULL)== SOCKET_ERROR){
//Delete Client
return;
}
}
Linux平台下
#include "/usr/include/linux/tcp.h"
#include "/usr/include/linux/socket.h"
KeepAlive实现,单位秒
//下面代码要求有ACE,如果没有包含ACE,则请把用到的ACE函数改成linux相应的接口
int keepAlive = 1;//设定KeepAlive
int keepIdle = 5;//开始首次KeepAlive探测前的TCP空闭时间
int keepInterval = 5;//两次KeepAlive探测间的时间间隔
int keepCount = 3;//判定断开前的KeepAlive探测次数
if(setsockopt(s,SOL_SOCKET,SO_KEEPALIVE,(void*)&keepAlive,sizeof(keepAlive)) == -1)
{
ACE_DEBUG ((LM_INFO,
ACE_TEXT ("(%P|%t) setsockopt SO_KEEPALIVE error!n")));
}
if(setsockopt(s,SOL_TCP,TCP_KEEPIDLE,(void *)&keepIdle,sizeof(keepIdle)) == -1)
{
ACE_DEBUG ((LM_INFO,
ACE_TEXT ("(%P|%t) setsockopt TCP_KEEPIDLE error!n")));
}
if(setsockopt(s,SOL_TCP,TCP_KEEPINTVL,(void *)&keepInterval,sizeof(keepInterval)) == -1)
{
ACE_DEBUG ((LM_INFO,
ACE_TEXT ("(%P|%t) setsockopt TCP_KEEPINTVL error!n")));
}
if(setsockopt(s,SOL_TCP,TCP_KEEPCNT,(void *)&keepCount,sizeof(keepCount)) == -1)
{
ACE_DEBUG ((LM_INFO,
ACE_TEXT ("(%P|%t)setsockopt TCP_KEEPCNT error!n")));
}
假设一种情况:一个socket TCP监听进程监听数据的时候如何接收关闭文件的socket连接?
不用并发
void setKeepAlive( int iSockfd , socklen_t iIdleTime , socklen_t iInterval , socklen_t iCount )
{
setsockopt( iSockfd , SOL_SOCKET , SO_KEEPALIVE , (const char*)&iSockAttrOn , sizeof(iSockAttrOn) );
setsockopt( iSockfd , SOL_TCP , TCP_KEEPIDLE , (const char*)&iIdleTime , sizeof(iIdleTime) );
setsockopt( iSockfd , SOL_TCP , TCP_KEEPINTVL , (const char*)&iInterval,sizeof(iInterval) );
setsockopt( iSockfd , SOL_TCP , TCP_KEEPCNT , (const char*)&iCount,sizeof(iCount) );
}