#!/bin/bash

# ===================================================
# 脚本功能：
# 1. 配置 systemd-networkd 网络桥接与 VLAN
# 2. 配置 k3s agent 节点并启动服务
# 3. 可通过配置文件 bushu.conf 管理所有参数
# ===================================================
set -euo pipefail

# -----------------------
# 检查 root 权限
# -----------------------
if [[ $EUID -ne 0 ]]; then
    echo "请使用 root 用户运行此脚本"
    exit 1
fi

# -----------------------
# 引入配置文件
# -----------------------
CONFIG_FILE="./bushu.conf"
if [[ -f "$CONFIG_FILE" ]]; then
    source "$CONFIG_FILE"
else
    echo "配置文件 $CONFIG_FILE 不存在，请先创建！"
    exit 1
fi

# -----------------------
# 网络配置部分
# -----------------------
echo "=== 开始 cloud 网桥配置 ==="

# 备份现有网络配置
mkdir -p /etc/systemd/network/backup
cp /etc/systemd/network/*.network /etc/systemd/network/backup/ 2>/dev/null || true
cp /etc/systemd/network/*.netdev /etc/systemd/network/backup/ 2>/dev/null || true
echo "网络配置备份已完成"


# 创建 cloud bridge
cat > /etc/systemd/network/00-cloud.netdev << EOF
[NetDev]
Name=cloud
Kind=bridge
MACAddress=$CLOUD_MAC

[Bridge]
VLANFiltering=yes
EOF

# 配置 cloud 网络
cat > /etc/systemd/network/00-cloud.network << EOF
[Match]
Name=cloud

[Network]
Address=$CLOUD_IP
Address=$CLOUD_SECOND_IP
Gateway=$CLOUD_GATEWAY
DNS=$CLOUD_DNS
ConfigureWithoutCarrier=yes

[Link]
RequiredForOnline=no-carrier
EOF


cat > /etc/systemd/network/00-$CLOUD_IFACE.network << EOF
[Match]
Name=$CLOUD_IFACE

[Network]
Bridge=cloud
DHCP=no
# VLAN=cloud-monitor

[BridgeVLAN]
VLAN=1-4094

[BridgeVLAN]
VLAN=1
PVID=1
EgressUntagged=1
EOF


echo "cloud 网桥配置已完成"

# -----------------------
echo "=== 开始 yewu 网桥配置 ==="

# 创建 yewu bridge
cat > /etc/systemd/network/00-yewu.netdev << EOF
[NetDev]
Name=yewu
Kind=bridge
MACAddress=$YEWU_MAC

[Bridge]
VLANFiltering=yes
EOF

# 配置 yewu 网络
cat > /etc/systemd/network/00-yewu.network << EOF
[Match]
Name=yewu

[Network]
Address=$YEWU_IP
Gateway=$YEWU_GATEWAY
DNS=$YEWU_DNS
ConfigureWithoutCarrier=yes

[Link]
RequiredForOnline=no-carrier
EOF

# 用户网卡接入 yewu bridge
cat > /etc/systemd/network/00-$YEWU_IFACE.network << EOF
[Match]
Name=$YEWU_IFACE

[Network]
Bridge=yewu
DHCP=no

[BridgeVLAN]
VLAN=1-4094

[BridgeVLAN]
VLAN=1
PVID=1
EgressUntagged=1
EOF

echo "yewu 网桥配置已完成"

# 使用 networkctl reload 应用配置
echo "正在重新加载网络配置..."
networkctl reload
echo "网络配置已生效"


# ===================================================
# 新增部分：monitor VLAN 子接口配置
# ===================================================
echo "=== 开始配置 monitor VLAN 子接口 ==="

: "${MONITOR_VLAN_ID:?请在 bushu.conf 中配置 MONITOR_VLAN_ID}"
: "${MONITOR_PHY_IFACE:?请在 bushu.conf 中配置 MONITOR_PHY_IFACE}"
: "${MONITOR_ALIAS:=lsa}"

MONITOR_VLAN_IFACE="${MONITOR_PHY_IFACE}.${MONITOR_VLAN_ID}"

echo "monitor 物理网卡: ${MONITOR_PHY_IFACE}"
echo "monitor VLAN ID: ${MONITOR_VLAN_ID}"
echo "monitor VLAN 子接口: ${MONITOR_VLAN_IFACE}"
echo "monitor VLAN alias: ${MONITOR_ALIAS}"

modprobe 8021q || true

if ip link show "${MONITOR_VLAN_IFACE}" >/dev/null 2>&1; then
    echo "VLAN 子接口 ${MONITOR_VLAN_IFACE} 已存在，跳过创建"
else
    echo "正在创建 VLAN 子接口 ${MONITOR_VLAN_IFACE} ..."
    ip link add link "${MONITOR_PHY_IFACE}" name "${MONITOR_VLAN_IFACE}" type vlan id "${MONITOR_VLAN_ID}"
fi

ip link set dev "${MONITOR_VLAN_IFACE}" alias "${MONITOR_ALIAS}"
ip link set dev "${MONITOR_VLAN_IFACE}" up

echo "monitor VLAN 子接口 ${MONITOR_VLAN_IFACE} 当前已生效"

# -----------------------
# 新增部分：写入 rc.local，保证重启后自动创建
# -----------------------
echo "=== 写入 rc.local，配置 monitor VLAN 开机自动创建 ==="

if [[ ! -f /etc/rc.local ]]; then
    cat > /etc/rc.local << 'EOF'
#!/bin/bash

exit 0
EOF
    chmod +x /etc/rc.local
fi

sed -i '/# BEGIN MONITOR VLAN/,/# END MONITOR VLAN/d' /etc/rc.local

sed -i "/^exit 0/i\\
# BEGIN MONITOR VLAN\\
modprobe 8021q || true\\
while ! ip link show ${MONITOR_PHY_IFACE} >/dev/null 2>&1; do\\
    sleep 1\\
done\\
ip link show ${MONITOR_VLAN_IFACE} >/dev/null 2>&1 || ip link add link ${MONITOR_PHY_IFACE} name ${MONITOR_VLAN_IFACE} type vlan id ${MONITOR_VLAN_ID}\\
ip link set dev ${MONITOR_VLAN_IFACE} alias ${MONITOR_ALIAS}\\
ip link set dev ${MONITOR_VLAN_IFACE} up\\
# END MONITOR VLAN\\
" /etc/rc.local

chmod +x /etc/rc.local

systemctl enable rc-local 2>/dev/null || true
systemctl enable rc-local.service 2>/dev/null || true

echo "rc.local 已写入 monitor VLAN 配置"



# -----------------------
# k3s 配置部分
# -----------------------
echo "=== k3s 安装配置 ==="
echo "=== 下载 K3s 二进制到 /usr/bin/ ==="
curl -L -o /usr/bin/k3s "https://file.piicloud.com/f/acfd6bc97c7e42de8647/?dl=1"
chmod +x /usr/bin/k3s
ls -lh /usr/bin/k3s



echo "=== 检查/创建 K3s agent 镜像目录 ==="
if [ ! -d "/var/lib/rancher/agent/images" ]; then
    mkdir -p /var/lib/rancher/agent/images
    echo "目录已创建"
else
    echo "目录已存在，跳过创建"
fi

# 下载镜像
[ -f "/var/lib/rancher/agent/images/base.tar" ] || \
curl -L -o /var/lib/rancher/agent/images/base.tar "https://file.piicloud.com/f/40e4c1cc37564dd8900d/?dl=1"
[ -f "/var/lib/rancher/agent/images/expend.tar" ] || \
curl -L -o /var/lib/rancher/agent/images/expend.tar "https://file.piicloud.com/f/c6cf4a7811d344ca9bf6/?dl=1"

chown root:root /var/lib/rancher/agent/images/*.tar
chmod 644 /var/lib/rancher/agent/images/*.tar
ls -lh /var/lib/rancher/agent/images/

# 创建 systemd 服务文件
cat > /etc/systemd/system/k3s.service << EOF
[Unit]
Description=Lightweight Kubernetes
Documentation=https://k3s.io
Wants=network-online.target
After=network-online.target

[Install]
WantedBy=multi-user.target

[Service]
Type=notify
EnvironmentFile=-/etc/default/%N
EnvironmentFile=-/etc/sysconfig/%N
EnvironmentFile=-/etc/systemd/system/k3s.service.env
KillMode=process
Delegate=yes
User=root
LimitNOFILE=1048576
LimitNPROC=infinity
LimitCORE=infinity
TasksMax=infinity
TimeoutStartSec=0
Restart=always
RestartSec=5s
ExecStartPre=-/sbin/modprobe br_netfilter
ExecStartPre=-/sbin/modprobe overlay
ExecStart=/usr/bin/k3s server
EOF


# 创建 override 配置
mkdir -p /etc/systemd/system/k3s.service.d
cat > /etc/systemd/system/k3s.service.d/override.conf << EOF
[Service]
ExecStart=
ExecStart=/usr/bin/k3s agent \
  --server https://$MASTER_IP:6443 \
  --token $K3S_TOKEN \
  --node-name $NODE_NAME \
  --node-ip $NODE_IP \
  --node-external-ip $NODE_EXT_IP \
  --kubelet-arg max-pods=120
EOF

# 创建 registries.yaml
mkdir -p /etc/rancher/k3s/
cat > /etc/rancher/k3s/registries.yaml << EOF
configs:
  harbor.tdology.com:
    auth:
      password: Y_TSq6qDJBZC}538
      username: admin
mirrors:
  docker.io:
    endpoint:
      - http://localhost:30500
      - https://harbor.tdology.com
EOF

# -----------------------
# NVMe 分区与格式化
# -----------------------

# 检查磁盘是否存在
if [[ ! -b /dev/nvme0n1 ]]; then
    echo "错误：/dev/nvme0n1 不存在"
    exit 1
fi

# 如果分区不存在，则创建
if [[ ! -b /dev/nvme0n1p1 ]]; then
    echo "/dev/nvme0n1p1 不存在，开始创建分区..."

    if lsblk -nr /dev/nvme0n1 | awk '{print $6}' | grep -q '^part$'; then
        echo "检测到已有分区，但没有 nvme0n1p1，请手动检查"
        exit 1
    fi

    parted -s /dev/nvme0n1 mklabel gpt
    parted -s /dev/nvme0n1 mkpart primary 0% 100%

    partprobe /dev/nvme0n1
    sleep 2

    if [[ ! -b /dev/nvme0n1p1 ]]; then
        echo "错误：分区创建失败"
        exit 1
    fi

    echo "分区创建成功：/dev/nvme0n1p1"
else
    echo "/dev/nvme0n1p1 已存在，跳过创建分区"
fi

# 如果没有文件系统，则格式化
if ! blkid /dev/nvme0n1p1 >/dev/null 2>&1; then
    echo "/dev/nvme0n1p1 没有文件系统，开始格式化为 ext4..."
    mkfs.ext4 -F /dev/nvme0n1p1
else
    echo "/dev/nvme0n1p1 已有文件系统，跳过格式化"
fi

# -----------------------
# NVMe 挂载与绑定挂载
# -----------------------

# 确保主挂载点存在
mkdir -p /mnt/nvme

# 挂载 /mnt/nvme
if ! mountpoint -q /mnt/nvme; then
    echo "挂载 /dev/nvme0n1p1 到 /mnt/nvme ..."
    mount /dev/nvme0n1p1 /mnt/nvme
else
    echo "/mnt/nvme 已经挂载，跳过挂载"
fi

# ✅ 确保源目录存在（关键修复点）
mkdir -p /mnt/nvme/rancher
mkdir -p /mnt/nvme/longhorn

# 确保绑定挂载目标目录存在
mkdir -p /var/lib/rancher
mkdir -p /var/lib/longhorn

# 挂载 /var/lib/rancher
if ! mountpoint -q /var/lib/rancher; then
    echo "绑定挂载 /mnt/nvme/rancher 到 /var/lib/rancher ..."
    mount --bind /mnt/nvme/rancher /var/lib/rancher
else
    echo "/var/lib/rancher 已经挂载，跳过绑定挂载"
fi

# 挂载 /var/lib/longhorn
if ! mountpoint -q /var/lib/longhorn; then
    echo "绑定挂载 /mnt/nvme/longhorn 到 /var/lib/longhorn ..."
    mount --bind /mnt/nvme/longhorn /var/lib/longhorn
else
    echo "/var/lib/longhorn 已经挂载，跳过绑定挂载"
fi

# -----------------------
# 更新 /etc/fstab
# -----------------------

NVME_UUID=$(blkid -s UUID -o value /dev/nvme0n1p1)
NVME_FSTYPE=$(blkid -s TYPE -o value /dev/nvme0n1p1)

if [[ -z "$NVME_UUID" || -z "$NVME_FSTYPE" ]]; then
    echo "错误：无法获取 UUID 或文件系统类型，停止写入 fstab"
    exit 1
fi

# 主挂载（安全）
grep -qE "[[:space:]]/mnt/nvme[[:space:]]" /etc/fstab || \
echo "UUID=${NVME_UUID} /mnt/nvme ${NVME_FSTYPE} defaults,nofail,x-systemd.device-timeout=5 0 2" >> /etc/fstab

# bind mount（安全）
grep -qE "[[:space:]]/var/lib/rancher[[:space:]]" /etc/fstab || \
echo "/mnt/nvme/rancher /var/lib/rancher none bind,nofail,x-systemd.requires-mounts-for=/mnt/nvme 0 0" >> /etc/fstab

grep -qE "[[:space:]]/var/lib/longhorn[[:space:]]" /etc/fstab || \
echo "/mnt/nvme/longhorn /var/lib/longhorn none bind,nofail,x-systemd.requires-mounts-for=/mnt/nvme 0 0" >> /etc/fstab

# -----------------------
# 验证
# -----------------------

echo "测试 fstab..."
mount -a

echo "显示挂载状态"
findmnt /mnt/nvme
findmnt /var/lib/rancher
findmnt /var/lib/longhorn

# -----------------------
# 时间同步
# -----------------------
echo "=== 时间同步配置 ==="

: "${ENABLE_CHRONY:=true}"
: "${ENABLE_GPTP:=false}"

if [[ "${ENABLE_CHRONY}" == "true" && "${ENABLE_GPTP}" == "true" ]]; then
    echo "错误：ENABLE_CHRONY 和 ENABLE_GPTP 不能同时为 true"
    exit 1
fi

echo "=== 从 RTC 恢复系统时间 ==="
hwclock -s 2>/dev/null || hwclock --hctosys 2>/dev/null || true
date

if [[ "${ENABLE_CHRONY}" == "true" || "${ENABLE_GPTP}" == "true" ]]; then
    echo "安装时间同步相关组件..."
    apt update
    apt install -y chrony linuxptp
fi

if [[ "${ENABLE_CHRONY}" == "true" ]]; then
    echo "=== 启用 chrony 时间同步 ==="

    systemctl disable --now gptp.service 2>/dev/null || true
    systemctl disable --now phc2sys.service 2>/dev/null || true
    pkill ptp4l 2>/dev/null || true
    pkill phc2sys 2>/dev/null || true

    cp /etc/chrony/chrony.conf /etc/chrony/chrony.conf.bak 2>/dev/null || true

    cat > /etc/chrony/chrony.conf << EOF
# 由 bushu.sh 自动生成
EOF

    for server in "${CHRONY_SERVERS[@]}"; do
        echo "server ${server} iburst" >> /etc/chrony/chrony.conf
    done

    cat >> /etc/chrony/chrony.conf << EOF

driftfile /var/lib/chrony/chrony.drift
makestep 1.0 -1
rtcsync
logdir /var/log/chrony
EOF

    systemctl enable --now chrony
    chronyc -a makestep || true
    sleep 3

    chronyc sources -v || true
    chronyc tracking || true

elif [[ "${ENABLE_GPTP}" == "true" ]]; then
    echo "=== 启用 gPTP 时间同步 ==="

    systemctl disable --now chrony 2>/dev/null || true

    : "${GPTP_IFACE:?请在 bushu.conf 中配置 GPTP_IFACE}"

    mkdir -p /usr/local/bin
    mkdir -p /etc/linuxptp

    cat > /etc/linuxptp/gptp.cfg << EOF
# 由 bushu.sh 自动生成
# gPTP / IEEE 802.1AS 配置

[global]
gmCapable               0
priority1               248
priority2               248
logAnnounceInterval     0
logSyncInterval         -3
syncReceiptTimeout      3
neighborPropDelayThresh 800
min_neighbor_prop_delay -20000000
assume_two_step         1
path_trace_enabled      1
follow_up_info          1
transportSpecific       0x1
ptp_dst_mac             01:80:C2:00:00:0E
network_transport       L2
delay_mechanism         P2P
time_stamping           hardware
EOF

    cat > /usr/local/bin/start-gptp.sh << 'EOF'
#!/bin/bash
set -euo pipefail

GPTP_IFACE="${1:?Usage: start-gptp.sh <interface>}"

while ! ip link show "${GPTP_IFACE}" >/dev/null 2>&1; do
    echo "等待网卡 ${GPTP_IFACE} 出现..."
    sleep 1
done

ip link set dev "${GPTP_IFACE}" up

if ethtool -T "${GPTP_IFACE}" 2>/dev/null | grep -q "hardware-transmit"; then
    TS_MODE="hardware"
else
    echo "警告：${GPTP_IFACE} 不支持硬件时间戳，切换为 software timestamp"
    TS_MODE="software"
    sed -i 's/^time_stamping.*/time_stamping           software/' /etc/linuxptp/gptp.cfg
fi

echo "gPTP timestamp mode: ${TS_MODE}"

exec /usr/sbin/ptp4l \
    -i "${GPTP_IFACE}" \
    -f /etc/linuxptp/gptp.cfg \
    -m
EOF

    chmod +x /usr/local/bin/start-gptp.sh

    cat > /etc/systemd/system/gptp.service << EOF
[Unit]
Description=gPTP IEEE 802.1AS Service
After=network-online.target
Wants=network-online.target

[Service]
Type=simple
ExecStart=/usr/local/bin/start-gptp.sh ${GPTP_IFACE}
Restart=always
RestartSec=5
User=root

[Install]
WantedBy=multi-user.target
EOF

    cat > /etc/systemd/system/phc2sys.service << EOF
[Unit]
Description=Synchronize system clock from gPTP PHC
After=gptp.service
Requires=gptp.service

[Service]
Type=simple
ExecStartPre=/bin/sleep 10
ExecStart=/usr/sbin/phc2sys -s ${GPTP_IFACE} -c CLOCK_REALTIME -O 0 -m
Restart=always
RestartSec=5
User=root

[Install]
WantedBy=multi-user.target
EOF

    systemctl daemon-reload
    systemctl enable --now gptp.service
    systemctl enable --now phc2sys.service

else
    echo "=== 未启用时间同步 ==="
    systemctl disable --now chrony 2>/dev/null || true
    systemctl disable --now gptp.service 2>/dev/null || true
    systemctl disable --now phc2sys.service 2>/dev/null || true
fi

echo "=== 检查系统时间 ==="
date
timedatectl || true

YEAR=$(date +%Y)
if (( YEAR < 2024 )); then
    echo "错误：系统时间异常，当前时间：$(date)"
    echo "拒绝继续执行，避免 apt/curl/k3s TLS 失败"
    exit 1
fi


# -----------------------
# 域名解析
# -----------------------
echo "=== 配置 /etc/hosts ==="
for entry in "${HOSTS_ENTRIES[@]}"; do
    grep -qF "$entry" /etc/hosts || echo "$entry" >> /etc/hosts
done

# -----------------------
# 启动 k3s 服务
# -----------------------
systemctl daemon-reload
systemctl enable k3s
systemctl start k3s
echo "k3s 配置完成"