#!/bin/bash
# 用实测最快的边缘 IP 固定隧道出口。
#
# 之前我判断"优选 IP 没用"是**错的** —— 结论下得太急。
# 真正的价值不是降低延迟（落点仍是 lax，改不了），
# 而是**避开会握手超时的 IP**：
#
#   日志里大量  dial tcp 198.41.192.167:7844: i/o timeout
#             Retrying connection in up to 32s
#   GeoDNS 随机给的 IP 里有一部分连不上，
#   cloudflared 就连上-掉线-重连反复折腾，
#   表现就是延迟在 0.69s ~ 4.4s 之间乱跳。
#
# 实测（probe_cf_edge2.py）最快且稳定的：
#   198.41.200.238   138ms
#   198.41.192.214   143ms
set -u

echo "===== 1. 备份 ====="
cp -f /etc/hosts "/root/hosts.bak.$(date +%s)"
echo "  已备份 /etc/hosts"

echo
echo "===== 2. 先复测这两个 IP 还通不通 ====="
python3 - <<'PY'
import socket, time
for ip in ("198.41.200.238", "198.41.192.214", "198.41.200.13", "198.41.192.37"):
    s = socket.socket(); s.settimeout(3)
    t = time.perf_counter()
    try:
        s.connect((ip, 7844))
        print("    %-16s %.0f ms" % (ip, (time.perf_counter() - t) * 1000))
    except OSError as e:
        print("    %-16s 不通 (%s)" % (ip, e.__class__.__name__))
    finally:
        s.close()
PY

echo
echo "===== 3. 写 /etc/hosts（固定两个最快的）====="
sed -i '/argotunnel.com/d' /etc/hosts
{
    echo ""
    echo "# ykxmz：固定 Cloudflare 隧道边缘 IP"
    echo "# 目的不是降低延迟（落点仍是 lax），而是避开会握手超时的 IP。"
    echo "# GeoDNS 随机发的 IP 里有些连不上，会导致隧道反复重连、延迟乱跳。"
    echo "198.41.200.238 region1.v2.argotunnel.com"
    echo "198.41.192.214 region2.v2.argotunnel.com"
} >> /etc/hosts
grep argotunnel /etc/hosts | sed 's/^/  /'

echo
echo "===== 4. 确认解析走的是我们指定的 ====="
python3 - <<'PY'
import socket
for host in ("region1.v2.argotunnel.com", "region2.v2.argotunnel.com"):
    infos = socket.getaddrinfo(host, 7844, socket.AF_INET)
    ips = sorted({i[4][0] for i in infos})
    print("    %s -> %s" % (host, ips))
PY

echo
echo "===== 5. 重启隧道并观察 60 秒 ====="
systemctl restart lanremote-tunnel
: > /root/reconnect_watch.txt
for i in $(seq 1 12); do
    sleep 5
    conn=$(curl -s --max-time 3 http://127.0.0.1:20241/metrics 2>/dev/null \
           | grep '^cloudflared_tunnel_ha_connections' | awk '{print $2}')
    errs=$(curl -s --max-time 3 http://127.0.0.1:20241/metrics 2>/dev/null \
           | grep '^cloudflared_tunnel_request_errors' | awk '{print $2}')
    echo "    ${i}0s: 连接 $conn  错误 $errs"
done

echo
echo "===== 6. 这 60 秒里有几次超时告警 ====="
since=$(date -u -d '70 seconds ago' '+%Y-%m-%dT%H:%M:%S' 2>/dev/null || echo "")
if [ -n "$since" ]; then
    n=$(awk -v s="$since" '$0 >= s' /root/lanremote-relay/tunnel.log 2>/dev/null \
        | grep -c 'i/o timeout')
    echo "  超时告警 $n 次（改之前每分钟都有好几次）"
else
    echo "  （算不出时间窗，跳过）"
fi

echo
echo "===== 7. 当前落点 ====="
grep -E 'Registered tunnel connection' /root/lanremote-relay/tunnel.log 2>/dev/null | tail -3 \
    | grep -oE 'ip=[0-9.]+ location=[a-z0-9]+ protocol=[a-z0-9]+' | sed 's/^/  /'

echo
echo "===== 8. 公网延迟（10 次）====="
ok=0; total=0
for i in $(seq 1 10); do
    r=$(curl -s -o /dev/null -w '%{http_code} %{time_total}' --max-time 25 \
        -H 'User-Agent: LanRemote/1.4.0' https://relay.xxmxxs.top/v1/health 2>/dev/null)
    c=$(echo "$r" | cut -d' ' -f1)
    t=$(echo "$r" | cut -d' ' -f2)
    [ "$c" = "200" ] && { ok=$((ok+1)); total=$(echo "$total + $t" | bc); }
    sleep 1
done
avg=$(echo "scale=2; $total / $ok" | bc 2>/dev/null || echo "?")
echo "  可用 $ok/10  平均 ${avg} 秒"