增加手工添加设备；machine-id重复的解决方法
This commit is contained in:
parent
edf64443b8
commit
c8612dee96
@ -67,6 +67,7 @@ func main() {
|
||||
r.Route("/api", func(r chi.Router) {
|
||||
r.Post("/discovery/search", h.Search)
|
||||
r.Get("/devices", h.ListDevices)
|
||||
r.Post("/devices", h.CreateDevice)
|
||||
r.Get("/devices/{id}", h.GetDevice)
|
||||
|
||||
// Proxy routes for device actions
|
||||
|
||||
160
docs/cloned_device_id_conflict.md
Normal file
160
docs/cloned_device_id_conflict.md
Normal file
@ -0,0 +1,160 @@
|
||||
# 克隆设备 device_id 冲突问题及解决方案
|
||||
|
||||
## 问题描述
|
||||
|
||||
部署多台 RK3588 设备时，如果它们是使用同一个镜像克隆的，可能会出现以下问题：
|
||||
|
||||
- UDP 发现只能找到部分设备（如 7 台中只能发现 2 台）
|
||||
- 设备列表中某些设备时隐时现
|
||||
- 下发配置时提示设备离线，但实际服务运行正常
|
||||
|
||||
## 原因分析
|
||||
|
||||
管理端通过 UDP 广播发现设备时，使用 `device_id` 作为唯一标识存储设备信息。
|
||||
|
||||
```go
|
||||
// 发现服务使用 device_id 作为 map 的 key
|
||||
found := make(map[string]*models.Device)
|
||||
...
|
||||
found[dev.DeviceID] = dev // 相同的 device_id 会覆盖之前的设备
|
||||
```
|
||||
|
||||
RK3588 Agent 的 `device_id` 生成优先级：
|
||||
1. `/etc/machine-id` 文件内容
|
||||
2. MAC 地址或序列号
|
||||
3. 自动生成并落盘
|
||||
|
||||
**克隆设备的 `/etc/machine-id` 完全相同**，导致所有克隆设备返回相同的 `device_id`，在发现过程中互相覆盖，最终只能看到一台。
|
||||
|
||||
## 解决方案
|
||||
|
||||
### 方案一：修复克隆设备（推荐，根治方案）
|
||||
|
||||
**⚠️ 注意：必须同时删除 dbus 的 machine-id，否则可能生成相同的 ID**
|
||||
|
||||
`systemd-machine-id-setup` 会优先复制 `/var/lib/dbus/machine-id`，克隆镜像中这个文件也是相同的，导致生成的 ID 仍然冲突。
|
||||
|
||||
**正确的修复步骤：**
|
||||
|
||||
```bash
|
||||
# 1. 同时删除两个 machine-id 文件
|
||||
sudo rm -f /etc/machine-id
|
||||
sudo rm -f /var/lib/dbus/machine-id
|
||||
|
||||
# 2. 重启（系统会自动生成新的随机 ID）
|
||||
sudo reboot
|
||||
|
||||
# 3. 验证新生成的 ID
|
||||
# 在每台设备上执行，确认输出不同
|
||||
cat /etc/machine-id
|
||||
```
|
||||
|
||||
**错误的做法（可能生成相同 ID）：**
|
||||
|
||||
```bash
|
||||
# ❌ 不要只删除 /etc/machine-id
|
||||
sudo rm -f /etc/machine-id
|
||||
sudo systemd-machine-id-setup # 会复制 dbus 的相同 ID！
|
||||
```
|
||||
|
||||
**替代方案：手动生成随机 ID**
|
||||
|
||||
```bash
|
||||
# 删除旧的
|
||||
sudo rm -f /etc/machine-id /var/lib/dbus/machine-id
|
||||
|
||||
# 手动生成新的随机 ID（每台设备分别执行）
|
||||
uuidgen | tr -d '-' | sudo tee /etc/machine-id
|
||||
|
||||
# 重启
|
||||
sudo reboot
|
||||
```
|
||||
|
||||
重启后，执行以下命令验证新的 device_id：
|
||||
|
||||
```bash
|
||||
cat /etc/machine-id
|
||||
```
|
||||
|
||||
然后重新在管理端点击"UDP 搜索"，所有设备应该都能被发现。
|
||||
|
||||
### 方案二：手动添加设备（临时方案）
|
||||
|
||||
如果暂时无法登录设备修改 machine-id，可以通过管理端的"手动添加设备"功能：
|
||||
|
||||
1. 打开管理端 Web 界面
|
||||
2. 进入设备页面
|
||||
3. 点击"手动添加设备"按钮
|
||||
4. 填写以下信息：
|
||||
|
||||
| 字段 | 必填 | 说明 |
|
||||
|------|------|------|
|
||||
| 设备 ID | 是 | 必须唯一，建议从设备执行 `cat /etc/machine-id` 获取，或使用 `uuidgen` 生成 |
|
||||
| 设备名称 | 否 | 便于识别的名称，如 cam-02 |
|
||||
| IP 地址 | 是 | 设备的实际 IP 地址 |
|
||||
| Agent 端口 | 否 | 默认 9100 |
|
||||
| Media 端口 | 否 | 默认 9000 |
|
||||
|
||||
**注意**：手动添加的设备，如果 device_id 与设备实际报告的 device_id 不一致，UDP 发现时可能会产生重复条目。
|
||||
|
||||
## 获取 device_id 的方法
|
||||
|
||||
在 RK3588 设备上执行：
|
||||
|
||||
```bash
|
||||
# 方法1：查看 machine-id（device_id 的来源）
|
||||
cat /etc/machine-id
|
||||
|
||||
# 方法2：查看 agent 信息（包含 device_id）
|
||||
curl http://127.0.0.1:9100/v1/info
|
||||
|
||||
# 方法3：生成新的唯一 ID
|
||||
uuidgen | tr -d '-'
|
||||
```
|
||||
|
||||
## 预防措施
|
||||
|
||||
制作克隆镜像前，建议清空 machine-id：
|
||||
|
||||
```bash
|
||||
# 在制作镜像前执行
|
||||
sudo rm -f /etc/machine-id
|
||||
sudo touch /etc/machine-id # 确保文件存在但为空
|
||||
```
|
||||
|
||||
这样设备首次启动时会自动生成新的 machine-id。
|
||||
|
||||
## 相关接口
|
||||
|
||||
管理端提供以下相关接口：
|
||||
|
||||
### 手动添加设备（API）
|
||||
|
||||
```bash
|
||||
POST /api/devices
|
||||
Content-Type: application/json
|
||||
|
||||
{
|
||||
"device_id": "rk3588-xxx",
|
||||
"device_name": "cam-01",
|
||||
"ip": "10.0.0.10",
|
||||
"agent_port": 9100,
|
||||
"media_port": 9000
|
||||
}
|
||||
```
|
||||
|
||||
### UDP 发现
|
||||
|
||||
```bash
|
||||
POST /api/discovery/search
|
||||
Content-Type: application/json
|
||||
|
||||
{
|
||||
"timeout_ms": 1200
|
||||
}
|
||||
```
|
||||
|
||||
## 参考
|
||||
|
||||
- [PRD_04_Manager_ConfigGUI_AgentAPI.md](../PRD_04_Manager_ConfigGUI_AgentAPI.md) - Agent API 文档
|
||||
- [API_Device_RemoteMgmt_InterfaceTable.md](../API_Device_RemoteMgmt_InterfaceTable.md) - 设备远程管理接口
|
||||
@ -6,6 +6,7 @@ import (
|
||||
"net/http"
|
||||
"net/url"
|
||||
"strings"
|
||||
"time"
|
||||
|
||||
"3588AdminBackend/internal/models"
|
||||
"3588AdminBackend/internal/service"
|
||||
@ -56,6 +57,42 @@ func (h *Handler) ListDevices(w http.ResponseWriter, r *http.Request) {
|
||||
})
|
||||
}
|
||||
|
||||
func (h *Handler) CreateDevice(w http.ResponseWriter, r *http.Request) {
|
||||
var req struct {
|
||||
DeviceID string `json:"device_id"`
|
||||
DeviceName string `json:"device_name"`
|
||||
IP string `json:"ip"`
|
||||
AgentPort int `json:"agent_port"`
|
||||
MediaPort int `json:"media_port"`
|
||||
}
|
||||
if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
|
||||
http.Error(w, "invalid request body", http.StatusBadRequest)
|
||||
return
|
||||
}
|
||||
if req.DeviceID == "" || req.IP == "" {
|
||||
http.Error(w, "device_id and ip are required", http.StatusBadRequest)
|
||||
return
|
||||
}
|
||||
if req.AgentPort == 0 {
|
||||
req.AgentPort = 9100
|
||||
}
|
||||
if req.MediaPort == 0 {
|
||||
req.MediaPort = 9000
|
||||
}
|
||||
|
||||
dev := &models.Device{
|
||||
DeviceID: req.DeviceID,
|
||||
DeviceName: req.DeviceName,
|
||||
IP: req.IP,
|
||||
AgentPort: req.AgentPort,
|
||||
MediaPort: req.MediaPort,
|
||||
Online: true,
|
||||
LastSeenMs: time.Now().UnixMilli(),
|
||||
}
|
||||
h.registry.UpdateDevice(dev)
|
||||
json.NewEncoder(w).Encode(map[string]interface{}{"ok": true})
|
||||
}
|
||||
|
||||
func (h *Handler) GetDevice(w http.ResponseWriter, r *http.Request) {
|
||||
id := chi.URLParam(r, "id")
|
||||
dev, ok := h.findDevice(id)
|
||||
@ -84,7 +121,6 @@ func (h *Handler) ProxyAgent(w http.ResponseWriter, r *http.Request) {
|
||||
http.Error(w, "device not found", http.StatusNotFound)
|
||||
return
|
||||
}
|
||||
|
||||
var agentPath string
|
||||
method := r.Method
|
||||
|
||||
@ -134,6 +170,11 @@ func (h *Handler) ProxyAgent(w http.ResponseWriter, r *http.Request) {
|
||||
return
|
||||
}
|
||||
|
||||
// Update online status on successful response
|
||||
if code < 500 && h.registry != nil {
|
||||
h.registry.TouchDevice(id)
|
||||
}
|
||||
|
||||
w.WriteHeader(code)
|
||||
w.Write(resp)
|
||||
}
|
||||
@ -146,7 +187,6 @@ func (h *Handler) ProxyAgentV1(w http.ResponseWriter, r *http.Request) {
|
||||
http.Error(w, "device not found", http.StatusNotFound)
|
||||
return
|
||||
}
|
||||
|
||||
prefix := fmt.Sprintf("/api/devices/%s", id)
|
||||
agentPath := strings.TrimPrefix(r.URL.Path, prefix)
|
||||
if !strings.HasPrefix(agentPath, "/v1/") && agentPath != "/v1" {
|
||||
@ -164,6 +204,11 @@ func (h *Handler) ProxyAgentV1(w http.ResponseWriter, r *http.Request) {
|
||||
return
|
||||
}
|
||||
|
||||
// Update online status on successful response
|
||||
if code < 500 && h.registry != nil {
|
||||
h.registry.TouchDevice(id)
|
||||
}
|
||||
|
||||
w.WriteHeader(code)
|
||||
_, _ = w.Write(resp)
|
||||
}
|
||||
|
||||
@ -53,6 +53,27 @@ func OpenAPI(w http.ResponseWriter, r *http.Request) {
|
||||
},
|
||||
},
|
||||
},
|
||||
"post": map[string]any{
|
||||
"requestBody": map[string]any{
|
||||
"required": true,
|
||||
"content": map[string]any{
|
||||
"application/json": map[string]any{
|
||||
"schema": map[string]any{
|
||||
"type": "object",
|
||||
"properties": map[string]any{
|
||||
"device_id": map[string]any{"type": "string"},
|
||||
"device_name": map[string]any{"type": "string"},
|
||||
"ip": map[string]any{"type": "string"},
|
||||
"agent_port": map[string]any{"type": "integer"},
|
||||
"media_port": map[string]any{"type": "integer"},
|
||||
},
|
||||
"required": []any{"device_id", "ip"},
|
||||
},
|
||||
},
|
||||
},
|
||||
},
|
||||
"responses": map[string]any{"200": map[string]any{"description": "ok"}},
|
||||
},
|
||||
},
|
||||
"/api/devices/{id}": map[string]any{
|
||||
"get": map[string]any{
|
||||
|
||||
@ -75,6 +75,18 @@ func (s *RegistryService) GetDevices() []*models.Device {
|
||||
return list
|
||||
}
|
||||
|
||||
// TouchDevice updates the LastSeenMs for a device to keep it online
|
||||
// when accessed via HTTP API (not just UDP discovery)
|
||||
func (s *RegistryService) TouchDevice(deviceID string) {
|
||||
s.mu.Lock()
|
||||
defer s.mu.Unlock()
|
||||
|
||||
if dev, ok := s.devices[deviceID]; ok {
|
||||
dev.LastSeenMs = time.Now().UnixMilli()
|
||||
dev.Online = true
|
||||
}
|
||||
}
|
||||
|
||||
func (s *RegistryService) startPruning() {
|
||||
ticker := time.NewTicker(2 * time.Second)
|
||||
for range ticker.C {
|
||||
|
||||
@ -121,6 +121,8 @@ func (u *UI) Routes() (chi.Router, error) {
|
||||
})
|
||||
|
||||
r.Get("/devices", u.pageDevices)
|
||||
r.Get("/devices-add", u.pageDeviceAdd)
|
||||
r.Post("/devices-add", u.actionDeviceAdd)
|
||||
r.Post("/devices/batch-action", u.actionDevicesBatchAction)
|
||||
r.Post("/discovery/search", u.actionDiscoverySearch)
|
||||
r.Get("/devices/{id}", u.pageDevice)
|
||||
@ -192,6 +194,42 @@ func (u *UI) pageDevices(w http.ResponseWriter, r *http.Request) {
|
||||
})
|
||||
}
|
||||
|
||||
func (u *UI) pageDeviceAdd(w http.ResponseWriter, r *http.Request) {
|
||||
u.render(w, r, "device_add", PageData{Title: "添加设备"})
|
||||
}
|
||||
|
||||
func (u *UI) actionDeviceAdd(w http.ResponseWriter, r *http.Request) {
|
||||
_ = r.ParseForm()
|
||||
deviceID := strings.TrimSpace(r.FormValue("device_id"))
|
||||
deviceName := strings.TrimSpace(r.FormValue("device_name"))
|
||||
ip := strings.TrimSpace(r.FormValue("ip"))
|
||||
agentPort, _ := strconv.Atoi(strings.TrimSpace(r.FormValue("agent_port")))
|
||||
mediaPort, _ := strconv.Atoi(strings.TrimSpace(r.FormValue("media_port")))
|
||||
|
||||
if deviceID == "" || ip == "" {
|
||||
u.render(w, r, "device_add", PageData{Title: "添加设备", Error: "device_id 和 ip 不能为空"})
|
||||
return
|
||||
}
|
||||
if agentPort == 0 {
|
||||
agentPort = 9100
|
||||
}
|
||||
if mediaPort == 0 {
|
||||
mediaPort = 9000
|
||||
}
|
||||
|
||||
dev := &models.Device{
|
||||
DeviceID: deviceID,
|
||||
DeviceName: deviceName,
|
||||
IP: ip,
|
||||
AgentPort: agentPort,
|
||||
MediaPort: mediaPort,
|
||||
Online: true,
|
||||
LastSeenMs: time.Now().UnixMilli(),
|
||||
}
|
||||
u.registry.UpdateDevice(dev)
|
||||
http.Redirect(w, r, "/ui/devices", http.StatusFound)
|
||||
}
|
||||
|
||||
func (u *UI) actionDiscoverySearch(w http.ResponseWriter, r *http.Request) {
|
||||
_ = r.ParseForm()
|
||||
timeoutMs, _ := strconv.Atoi(strings.TrimSpace(r.FormValue("timeout_ms")))
|
||||
|
||||
49
internal/web/ui/templates/device_add.html
Normal file
49
internal/web/ui/templates/device_add.html
Normal file
@ -0,0 +1,49 @@
|
||||
{{define "device_add"}}
|
||||
<div class="card">
|
||||
<h2>手动添加设备</h2>
|
||||
<div class="muted small">用于 UDP 发现失败时手动添加设备（如克隆设备导致 device_id 冲突）</div>
|
||||
<form method="post" action="/ui/devices-add" style="margin-top:16px">
|
||||
<div class="row" style="gap:16px">
|
||||
<div>
|
||||
<div class="muted small">设备 ID <span class="muted">*</span></div>
|
||||
<input name="device_id" placeholder="rk3588-xxx" required />
|
||||
<div class="muted small" style="margin-top:4px">必须唯一，可从 /etc/machine-id 获取</div>
|
||||
</div>
|
||||
<div>
|
||||
<div class="muted small">设备名称</div>
|
||||
<input name="device_name" placeholder="cam-01" />
|
||||
</div>
|
||||
</div>
|
||||
<div class="row" style="gap:16px;margin-top:12px">
|
||||
<div>
|
||||
<div class="muted small">IP 地址 <span class="muted">*</span></div>
|
||||
<input name="ip" placeholder="10.0.0.10" required />
|
||||
</div>
|
||||
<div>
|
||||
<div class="muted small">Agent 端口</div>
|
||||
<input name="agent_port" value="9100" />
|
||||
<div class="muted small" style="margin-top:4px">默认 9100</div>
|
||||
</div>
|
||||
<div>
|
||||
<div class="muted small">Media 端口</div>
|
||||
<input name="media_port" value="9000" />
|
||||
<div class="muted small" style="margin-top:4px">默认 9000</div>
|
||||
</div>
|
||||
</div>
|
||||
<div style="margin-top:16px">
|
||||
<button type="submit">添加设备</button>
|
||||
<a href="/ui/devices" style="margin-left:12px">返回设备列表</a>
|
||||
</div>
|
||||
</form>
|
||||
</div>
|
||||
|
||||
<div class="card">
|
||||
<h3>提示：克隆设备的 device_id 冲突问题</h3>
|
||||
<div class="muted small">
|
||||
如果多台设备是克隆的镜像，它们的 /etc/machine-id 会相同，导致 UDP 发现只能找到一台。<br/>
|
||||
<b>解决方法</b>：在每台克隆设备上执行以下命令重新生成 machine-id（<b>注意：必须同时删除 dbus 的 machine-id，否则可能生成相同的 ID</b>）：<br/>
|
||||
<pre style="margin-top:8px">sudo rm -f /etc/machine-id /var/lib/dbus/machine-id
|
||||
sudo reboot</pre>
|
||||
</div>
|
||||
</div>
|
||||
{{end}}
|
||||
@ -1,15 +1,20 @@
|
||||
{{define "devices"}}
|
||||
<div class="card">
|
||||
<h2>发现设备</h2>
|
||||
<form method="post" action="/ui/discovery/search" class="row">
|
||||
<div>
|
||||
<div class="muted small">超时（毫秒）</div>
|
||||
<input name="timeout_ms" value="1200" />
|
||||
</div>
|
||||
<div class="row" style="gap:16px">
|
||||
<form method="post" action="/ui/discovery/search" class="row">
|
||||
<div>
|
||||
<div class="muted small">超时（毫秒）</div>
|
||||
<input name="timeout_ms" value="1200" />
|
||||
</div>
|
||||
<div style="align-self:end">
|
||||
<button type="submit">搜索（UDP 广播）</button>
|
||||
</div>
|
||||
</form>
|
||||
<div style="align-self:end">
|
||||
<button type="submit">搜索（UDP 广播）</button>
|
||||
<a href="/ui/devices-add"><button type="button">手动添加设备</button></a>
|
||||
</div>
|
||||
</form>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<div class="card">
|
||||
|
||||
@ -3,6 +3,6 @@
|
||||
"discovery_port": 35688,
|
||||
"discovery_timeout_ms": 1200,
|
||||
"offline_after_ms": 10000,
|
||||
"agent_token": "CHANGE_ME",
|
||||
"agent_token": "4fe2d69fda23d0d5d04a1486d4920e68",
|
||||
"concurrency": 5
|
||||
}
|
||||
|
||||
Loading…
Reference in New Issue
Block a user