#!/bin/bash
# 立刻恢复隧道 + 确认问题。
#
# 现象：cloudflared 进程活着、指标说 ha_connections=1，
#       但日志 8 小时没动静、公网 1033。
#       -> 那是条"僵死连接"：计数还在，实际早断了。
set -u

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

echo "===== 1. 重启之前 ====="
echo -n "  ha_connections = "
curl -s --max-time 5 http://127.0.0.1:20241/metrics 2>/dev/null \
    | grep '^cloudflared_tunnel_ha_connections' | awk '{print $2}'
echo "  日志最后一条: $(tail -1 $RELAY_DIR/tunnel.log 2>/dev/null | cut -c1-30)"
echo "  现在时间:     $(date -u '+%Y-%m-%dT%H:%M:%SZ')"
echo "  公网: HTTP $(timeout 15 curl -4 -s -o /dev/null -w '%{http_code}' -H 'User-Agent: LanRemote/1.4.0' $URL/v1/health 2>/dev/null)"

echo
echo "===== 2. 重启隧道 ====="
systemctl restart lanremote-tunnel
sleep 40

echo "  服务: $(systemctl is-active lanremote-tunnel)"
echo -n "  ha_connections = "
curl -s --max-time 5 http://127.0.0.1:20241/metrics 2>/dev/null \
    | grep '^cloudflared_tunnel_ha_connections' | awk '{print $2}'
echo "  注册: $(grep -c 'Registered tunnel connection' $RELAY_DIR/tunnel.log 2>/dev/null || echo 0)"

echo
echo "===== 3. 验证公网 ====="
OK=0
for i in $(seq 1 10); do
    C=$(timeout 15 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=$((OK+1)); M="✓"; } || M="✗"
    printf "    %2d: %s %s\n" "$i" "$M" "${C:-超时}"
    sleep 2
done
echo "  公网 $OK / 10"

echo
echo "===== 4. 本地源站（对照）====="
echo "  HTTP $(timeout 8 curl -s -o /dev/null -w '%{http_code}' http://127.0.0.1:8080/v1/health)"

echo
echo "===== 5. 结论 ====="
if [ "$OK" -ge 8 ]; then
    echo "  ✓ 重启后恢复了"
    echo "  问题定性：cloudflared 的连接僵死了，但指标还报 1 ——"
    echo "            看门狗只看指标，所以 8 小时没动作。"
else
    echo "  ✗ 重启也没恢复，链路本身可能不行"
fi