Cluster intranet（Linux Docker）

**Step One: Environment Preparation**

Please refer to [Environment Preparation](https://docs.paraverse.cc/doc/219/) to set up two servers, with at least one being a GPU server. If the master node will be used as a rendering node, follow the steps for [Deploying the Standalone Version](https://docs.paraverse.cc/doc/243/) for deployment, starting from step 8 for the rendering node configuration. If the master node is not used for rendering, proceed with the following steps for deployment.

**Step Two: System Composition**

Based on [System Components](https://docs.paraverse.cc/doc/244/),download the program package and recommend extracting it to the root directory of a drive, ensuring that the path does not contain any Chinese characters.

**Step Three: Port Availability**

Configure the system ensuring that the system ports are available and not in use. If any ports are occupied, make the necessary modifications.

**Master Node:**

|Type | Port Number | Purpose | Modification Method|
|-------- | -------- | -------- |------------|
|TCP | **8181** | Frontend page nginx forwarding and backend interface | [Modify 8181 port]( https://www.pingxingyun.com/question/detail?topicId=10078&categoryId=2 Please refer to)|
|TCP | **8182** | Backend page nginx forwarding | [Modify port 8182]( https://www.pingxingyun.com/question/detail?topicId=10114&categoryId=2 Please refer to)|
|TCP | 8184 | Management Services | [Modify 8184 Port]( https://www.pingxingyun.com/question/detail?topicId=10116&categoryId=2 "Modify port 8184")|
|TCP | 6379 | Redis | [Modify port 6379](https://www.pingxingyun.com/question/detail?topicId=10117&categoryId=2 "Modify port 6379")|

**Render Server**

|Type | Port Number | Purpose | Modification Method|
|-------- | -------- | -------- |------------|
|TCP | 8089 | Sync and Proxy Service Port | [Modify 8089 Port]( https://www.pingxingyun.com/question/detail?topicId=10118&categoryId=2 "Modify port 8089")|
|TCP | 10002 | Websocket communication port | [Modify 10002 port]( https://www.pingxingyun.com/question/detail?topicId=10119&categoryId=2 "Modify port 10002")|
|TCP | 10003 | Internal communication port of rendering server | [Modify 10003 port]( https://www.pingxingyun.com/question/detail?topicId=10121&categoryId=2 "Modify Port 10003")|
|TCP | 10004 | rendering server SSL port | [modify 10004 port]( https://www.pingxingyun.com/question/detail?topicId=10120&categoryId=2 "Modify Port 10004")|
|UDP | **10011-10040** | Audio and Video Communication Port | [Modify Audio and Video Port]( https://www.pingxingyun.com/question/detail?topicId=10122&categoryId=2 "Modify audio and video ports")|

## Center

- **1. Prepare the Linux Server**

- **2. Increase the System's Open File Limit:**
    ```html
    ulimit -n 65535
    ```
- **3.Set the System to Use UTC Time by Default:**
  ```html
  sudo ln -sf /usr/share/zoneinfo/Asia/Shanghai /etc/localtime
  ```
  Verify the time：`timedatectl` or `date`

- **4. Set a Static IP for the Virtual Machine：**
  1. Open the configuration file:  `sudo vim /etc/netplan/01-network-manager-all.yaml`
  2. Modify the configuration file as follows:：
  ```html
  network:
    ethernets:
        ens33:
            dhcp4: false
            addresses: [192.168.0.x/24]
            optional: true
            gateway4: 192.168.0.1
            nameservers:
                addresses: [114.114.114.114,8.8.8.8]
    version: 2
  ```
  3. Apply the network settings:  `sudo netplan apply` or `sudo netplan --debug apply`

- **5. Upload the Management Node Package to Linux**
    1.  Open the [Pingxingyun Development Center page](https://paraverse.cc/larkxr-resource/)in your browser and download the [Linux] compressed package.

![](https://docs.paraverse.cc/media/202505/36bfb1936a6e4137974f8b1c9ccdeda36689.png)

2.  Extract the downloaded compressed package and use WinSCP to upload the   larkxr-center-docker   package from the center management server to the virtual machine.

3. After completing the configuration, return to the   /opt/larkxr-center-ubuntu-docker   directory and start the service using the following command:
       ```html
        sh start.sh
        ```

- **6.After all services are started successfully, access IP:8182 with the default account admin and password 123456**

-  **7. Log in to the backend interface to view resource monitoring information. At this point, the number of GPU nodes is 0. Next, you need to configure and start the rendering service.**

## Render

- **1.Prepare a Linux server**

- **2. Confirm the Graphics Card Driver**
    Confirm whether the NVIDIA graphics card driver is correctly installed on the current system. If it is not installed, install the driver corresponding to your graphics card.
    ```html
    nvidia-smi
    ```
- **3.Execute the Following Commands to Modify Kernel Parameters (Optional Operation)**
   View the configuration file:
  `  sudo view /etc/sysctl.d/10-ptrace.conf`
   • Add or modify the following content: 
  ` kernel.yama.ptrace_scope=0 `
   • Save and exit, then execute the following command to apply the changes:  
   `sudo /sbin/sysctl -p  ln -sf /usr/share/zoneinfo/Asia/Shanghai /etc/localtime`

- **4.Set the System to Use UTC Time by Default:**
  ```html
  sudo ln -sf /usr/share/zoneinfo/Asia/Shanghai /etc/localtime
  ```
  Verify the time：`timedatectl`或`date`

- **5.Set a static IP for the virtual machine**
  1. Open the configuration file：`sudo vim /etc/netplan/01-network-manager-all.yaml`
  2.  Modify the configuration file as follows：
  ```html
  network:
    ethernets:
        ens33:
            dhcp4: false
            addresses: [192.168.0.x/24]
            optional: true
            gateway4: 192.168.0.1
            nameservers:
                addresses: [114.114.114.114,8.8.8.8]
    version: 2
  ```
  3. Apply the network settings：`sudo netplan apply`or`sudo netplan --debug apply`

- **6.Upload the management node package to Linux**
  1.  Open the [Pingxingyun Development Center page](https://paraverse.cc/larkxr-resource/)in your browser and download the [Linux] compressed package.

2.  Extract the downloaded compressed package and use WinSCP to upload the   larkxr-render-docker   package from the center management server to the virtual machine.

4. After completing the configuration, return to the   /opt/larkxr-render-docker   directory and start the service using the following command
     ```html
       cd /opt/larkxr-render-docker/
      sh start.sh
       ```
     To stop the service

```html
     cd /opt/larkxr-render-docker/
      sh stop.sh
      ```

- **7.Log in to the Backend Interface and View Resource Monitoring Information:At this point, the number of GPU nodes should be 1**