#!/bin/bash
# 部署重启之后可用率掉到 76%，查清楚原因。
#
# 要排除的可能：
#   1. MSS 规则被部署脚本清掉了       -> 查规则
#   2. 隧道重建时没带上 MSS（规则生效于新连接）-> 查隧道建立时间
#   3. 链路本身在这个时段变差          -> 分组对比测量
set -u

RELAY_DIR=/root/lanremote-relay
URL=https://relay.xxmxxs.top

echo "===== 1. 调优还在不在 ====="
echo "  tcp_mtu_probing = $(cat /proc/sys/net/ipv4/tcp_mtu_probing)"
echo "  MSS 规则 $(iptables -t mangle -L -n 2>/dev/null | grep -ci mss) 条"
iptables -t mangle -L -n 2>/dev/null | grep -i mss | sed 's/^/    /'

echo
echo "===== 2. 隧道什么时候建的（要比 MSS 规则晚才对）====="
echo "  MSS 规则最早（服务启动时间）："
systemctl show lanremote-nettune.service -p ActiveEnterTimestamp --value
echo "  隧道启动时间："
systemctl show lanremote-tunnel -p ActiveEnterTimestamp --value
echo "  隧道最近的注册记录："
grep 'Registered tunnel connection' "$RELAY_DIR/tunnel.log" 2>/dev/null \
    | tail -3 | sed 's/^/    /'

echo
echo "===== 3. 服务状态 ====="
for s in lanremote-nettune.service lanremote-relay lanremote-tunnel; do
    printf "  %-28s %s\n" "$s" "$(systemctl is-active "$s")"
done
echo "  连接数 = $(curl -s --max-time 5 http://127.0.0.1:20241/metrics 2>/dev/null | grep '^cloudflared_tunnel_ha_connections' | awk '{print $2}')"

echo
echo "===== 4. 分组测量：连续 3 组，每组 20 次 ====="
for group in 1 2 3; do
    ok=0
    for i in $(seq 1 20); do
        c=$(timeout 12 curl -4 -s -o /dev/null -w '%{http_code}' \
            -H 'User-Agent: LanRemote/1.4.0' "$URL/v1/health" 2>/dev/null)
        [ "$c" = "200" ] && ok=$((ok+1))
        sleep 1
    done
    echo "  第 $group 组: $ok/20 = $((ok * 5))%"
done

echo
echo "===== 5. 路径 MTU 现状（MSS 1400 是否仍然合适）====="
for size in 1400 1440 1452 1460 1472; do
    out=$(ping -c 3 -W 3 -M do -s "$size" 1.1.1.1 2>&1)
    loss=$(echo "$out" | grep -oE '[0-9]+% packet loss' | head -1)
    rtt=$(echo "$out" | grep -oE 'min/avg/max[^=]*= [0-9./]+' | head -1)
    printf "  包 %-5s 丢包 %-14s %s\n" "$((size + 28))" "${loss:-?}" "${rtt:-}"
done

echo
echo "===== 6. 强制让隧道用上最新参数重建 ====="
systemctl restart lanremote-tunnel
sleep 40
echo "  连接数 = $(curl -s --max-time 5 http://127.0.0.1:20241/metrics 2>/dev/null | grep '^cloudflared_tunnel_ha_connections' | awk '{print $2}')"
ok=0
for i in $(seq 1 20); do
    c=$(timeout 12 curl -4 -s -o /dev/null -w '%{http_code}' \
        -H 'User-Agent: LanRemote/1.4.0' "$URL/v1/health" 2>/dev/null)
    [ "$c" = "200" ] && ok=$((ok+1))
    sleep 1
done
echo "  重建后: $ok/20 = $((ok * 5))%"