最近在linux做伺服器開發的時候，發現了一個現象：伺服器在啟動的時候呼叫了 connect 函式，因為連線了一個不可用的埠，導致connect最後報出了 “Connection timed out” 的錯誤。但是這中間過了六十多秒的時間。

為何會等待這麼長的時間才超時呢？這個時間又在哪裡設定？

《UNIX網路程式設計（第一卷）——套介面 API 和 X/Open 傳輸介面 API》一書的4.3節有寫到：

　　對於TCP套介面來說，函式 connect 激發TCP的三路握手過程，且僅在連結成功建立或出錯時才返回，返回的錯誤可能有如下幾種情況：

　　1. 如果TCP客戶沒有收到SYN分節的響應，則返回ETIMEDOUT。例如在4.4BSD中，當呼叫函式 connect 時，發出一個SYN，若無響應，等待6秒之後再發一個；若仍無響應，24秒鐘之後再發一個。若總共等待了75秒鐘之後仍未響應，則返回錯誤...

從書中可以看到 connect 建立TCP連結的過程中，會發送SYN包，如果沒有收到SYN包的回包，核心會多次傳送SYN包，並且每次重試的間隔會逐漸增加，避免傳送太多的SYN包影響網路。

在CentOS上，這個重試次數是可以設定的：

$ sysctl net.ipv4 | grep tcp
net.ipv4.tcp_timestamps = 1
net.ipv4.tcp_window_scaling = 1
net.ipv4.tcp_sack = 1
net.ipv4.tcp_retrans_collapse = 1
net.ipv4.tcp_syn_retries = 5
net.ipv4.tcp_synack_retries 
 
= 5
net.ipv4.tcp_max_orphans = 262144
net.ipv4.tcp_max_tw_buckets = 262144
net.ipv4.tcp_keepalive_time = 7200
net.ipv4.tcp_keepalive_probes = 9
net.ipv4.tcp_keepalive_intvl = 75
net.ipv4.tcp_retries1 = 3
net.ipv4.tcp_retries2 = 15
net.ipv4.tcp_fin_timeout = 60
net.ipv4.tcp_syncookies = 1
net.ipv4.tcp_tw_recycle 
 
= 0
net.ipv4.tcp_abort_on_overflow = 0
net.ipv4.tcp_stdurg = 0
net.ipv4.tcp_rfc1337 = 0
net.ipv4.tcp_max_syn_backlog = 2048
net.ipv4.tcp_orphan_retries = 0
net.ipv4.tcp_fack = 1
net.ipv4.tcp_reordering = 3
net.ipv4.tcp_ecn = 2
net.ipv4.tcp_dsack = 1
net.ipv4.tcp_mem = 639168       852224  1278336
net.ipv4.tcp_wmem = 4096        16384   4194304
net.ipv4.tcp_rmem = 4096        87380   4194304
net.ipv4.tcp_app_win = 31
net.ipv4.tcp_adv_win_scale = 2
net.ipv4.tcp_tw_reuse = 0
net.ipv4.tcp_frto = 2
net.ipv4.tcp_frto_response = 0
net.ipv4.tcp_low_latency = 0
net.ipv4.tcp_no_metrics_save = 0
net.ipv4.tcp_moderate_rcvbuf = 1
net.ipv4.tcp_tso_win_divisor = 3
net.ipv4.tcp_congestion_control = cubic
net.ipv4.tcp_abc = 0
net.ipv4.tcp_mtu_probing = 0
net.ipv4.tcp_base_mss = 512
net.ipv4.tcp_workaround_signed_windows = 0
net.ipv4.tcp_challenge_ack_limit = 1000
net.ipv4.tcp_limit_output_bytes = 262144
net.ipv4.tcp_dma_copybreak = 4096
net.ipv4.tcp_slow_start_after_idle = 1
net.ipv4.tcp_available_congestion_control = cubic reno
net.ipv4.tcp_allowed_congestion_control = cubic reno
net.ipv4.tcp_max_ssthresh = 0
net.ipv4.tcp_thin_linear_timeouts = 0
net.ipv4.tcp_thin_dupack = 0
net.ipv4.tcp_min_tso_segs = 2
net.ipv4.tcp_invalid_ratelimit = 500

其中的 net.ipv4.tcp_syn_retries 選項控制著SYN的重試次數，可以通過如下命令來檢視和設定：

$ sysctl net.ipv4.tcp_syn_retries 　　　　　　#檢視
net.ipv4.tcp_syn_retries = 5
$ sudo sysctl -w net.ipv4.tcp_syn_retries=1 #設定
net.ipv4.tcp_syn_retries = 1

下面用一個簡單的程式，來驗證各種次數下的connect超時時間：

#include <iostream>
#include <sys/socket.h>
#include <sys/time.h>
#include <netinet/in.h>
#include <errno.h>
#include <string.h>
#include <arpa/inet.h>

long long GetCurrentMSecond()
{
    struct timeval tv; 
    gettimeofday(&tv, NULL);
    return tv.tv_sec * 1000 + tv.tv_usec / 1000;
}

int main()
{
    int fd = 0;
    struct sockaddr_in addr;

    fd = socket(AF_INET, SOCK_STREAM, 0); 

    socklen_t bufSize = 128 * 1024;
    int retCode = setsockopt(fd, SOL_SOCKET, SO_RCVBUF, &bufSize, sizeof(bufSize));
    addr.sin_family = AF_INET;
    addr.sin_addr.s_addr = inet_addr("192.168.207.128");
    addr.sin_port = htons(13500); //連線一個不用的埠，以保證會觸發超時
    long long llBeginTime = GetCurrentMSecond();
    if (connect(fd, (struct sockaddr*)&addr, sizeof(addr)) == -1) 
    {   
        long long llEndTime = GetCurrentMSecond();
        std::cout << "connect failed, errno: " << errno << ", error: " << strerror(errno)
                  << ", cost time: " << llEndTime - llBeginTime << std::endl;
        return 0;
    }   
    std::cout << "connect success" << std::endl;
}

通過設定不同的重試次數，探究各種重試次數下的time out時間：

$ g++ connect.cpp -o main
$ sudo sysctl -w net.ipv4.tcp_syn_retries=1
net.ipv4.tcp_syn_retries = 1
$ ./main 
connect failed, errno: 110, error: Connection timed out, cost time: 3000
$ sudo sysctl -w net.ipv4.tcp_syn_retries=2
net.ipv4.tcp_syn_retries = 2
$ ./main 
connect failed, errno: 110, error: Connection timed out, cost time: 7000

從表格中可以看到，當前設定重試次數為5的時候，超時時間是63秒，可以通過修改重試次數的方式，來改變connect的超時時間。