TCP KeepAlive的几个附加选项

TCP_KEEPALIVE选项只是一个开关，Linux中默认的Keepalive的选项如下：

$sudo sysctl -a | grep keepalive
net.ipv4.tcp_keepalive_time = 7200
net.ipv4.tcp_keepalive_probes = 9
net.ipv4.tcp_keepalive_intvl = 75

上文中的keepalive选项表示如果一个连接上7200s后没有任何数据发送，则设置了这个选项的本端向对端发送keepalive保活报文，它会有如下三种结果：

• 对端回复ACK。则本端TCP认为该连接依然存活。继续等7200s后再发送keepalive报文。
• 对端回复RESET。说明对端进程已经重启，本端的应用程序应该关闭该连接。
• 没有对端的任何回复。则本端做重试，如果重试9次（前后重试间隔为75秒）仍然不可达，则向应用程序返回错误信息，ETIMEOUT（无任何应答）或EHOST

如果应用程序向改变keepalive的默认行为，该怎么办呢？答案就是利用 TCP_KEEPIDLE、TCP_KEEPINTVL、TCP_KEEPCNT这几个TCP选项，首先看看如何使用：

int setKeepAlive(int fd, int interval)
{
    int val = 1;

    if (setsockopt(fd, SOL_SOCKET, SO_KEEPALIVE, &val, sizeof(val)) == -1) {
        printf("setsockopt SO_KEEPALIVE: %s", strerror(errno));
        return -1;
    }

    /* Send first probe after `interval' seconds. */
    val = interval;
    if (setsockopt(fd, IPPROTO_TCP, TCP_KEEPIDLE, &val, sizeof(val)) < 0) {
        printf("setsockopt TCP_KEEPIDLE: %s\n", strerror(errno));
        return -1;
    }

    /* Send next probes after the specified interval. Note that we set the
     * delay as interval / 3, as we send three probes before detecting
     * an error (see the next setsockopt call). */
    val = interval/3;
    if (val == 0) val = 1;
    if (setsockopt(fd, IPPROTO_TCP, TCP_KEEPINTVL, &val, sizeof(val)) < 0) {
        printf("setsockopt TCP_KEEPINTVL: %s\n", strerror(errno));
        return -1;
    }

    /* Consider the socket in error state after three we send three ACK
     * probes without getting a reply. */
    val = 3;
    if (setsockopt(fd, IPPROTO_TCP, TCP_KEEPCNT, &val, sizeof(val)) < 0) {
        printf("setsockopt TCP_KEEPCNT: %s\n", strerror(errno));
        return -1;
    }

    return 0;
}

• TCP_KEEPDILE 设置连接上如果没有数据发送的话，多久后发送keepalive探测分组，单位是秒
• TCP_KEEPINTVL 前后两次探测之间的时间间隔，单位是秒
• TCP_KEEPCNT 关闭一个非活跃连接之前的最大重试次数

用tcpdump抓包就可以看到设置了上面选项的那端的详细行为。

Date: 2015-07-17T19:45+0800

Author: Cobbliu

Org version 7.9.3f with Emacs version 24