e想行空

ubuntu-intall-cuda

发表于 2024-11-19 更新于 2024-11-20 分类于 Linux Waline：

Ubuntu20.04如何安装CUDA

安装显卡驱动

#更新系统，安装必须用到的工具
exxk@exxk:~$ sudo apt update && sudo apt upgrade -y
exxk@exxk:~$ sudo apt install -y curl wget gnupg lsb-release
#添加 NVIDIA 驱动的 PPA
exxk@exxk:~$ sudo add-apt-repository ppa:graphics-drivers/ppa
exxk@exxk:~$ sudo apt update
#检测可用驱动： 它会列出所有设备（如显卡）和相应的推荐驱动。如果知道安装什么去驱动了，可以不用安装
exxk@exxk:~$ sudo apt install -y ubuntu-drivers-common
exxk@exxk:~$ ubuntu-drivers devices
ERROR:root:could not open aplay -l
Traceback (most recent call last):
  File "/usr/share/ubuntu-drivers-common/detect/sl-modem.py", line 35, in detect
    aplay = subprocess.Popen(
  File "/usr/lib/python3.8/subprocess.py", line 858, in __init__
    self._execute_child(args, executable, preexec_fn, close_fds,
  File "/usr/lib/python3.8/subprocess.py", line 1704, in _execute_child
    raise child_exception_type(errno_num, err_msg, err_filename)
FileNotFoundError: [Errno 2] No such file or directory: 'aplay' #系统试图使用aplay来检测音频设备，但该工具未安装，可以忽略该错误
== /sys/devices/pci0000:00/0000:00:10.0 ==
modalias : pci:v000010DEd000025E2sv000017AAsd0000382Dbc03sc00i00
vendor   : NVIDIA Corporation
driver   : nvidia-driver-470 - distro non-free
driver   : nvidia-driver-535-server - distro non-free
driver   : nvidia-driver-535 - distro non-free
driver   : nvidia-driver-550-open - third-party non-free
driver   : nvidia-driver-535-open - distro non-free
driver   : nvidia-driver-550 - third-party non-free
driver   : nvidia-driver-560 - third-party non-free recommended #推荐的驱动
driver   : nvidia-driver-535-server-open - distro non-free
driver   : nvidia-driver-545-open - third-party non-free
driver   : nvidia-driver-470-server - distro non-free
driver   : nvidia-driver-555-open - third-party non-free
driver   : nvidia-driver-545 - third-party non-free
driver   : nvidia-driver-555 - third-party non-free
driver   : nvidia-driver-560-open - third-party non-free
driver   : xserver-xorg-video-nouveau - distro free builtin
exxk@exxk:~$ sudo apt install -y nvidia-driver-560
exxk@exxk:~$ sudo reboot
exxk@exxk:~$ nvidia-smi
Tue Nov 19 07:27:03 2024
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 560.35.03              Driver Version: 560.35.03      CUDA Version: 12.6     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 3050 ...    Off |   00000000:00:10.0 Off |                  N/A |
| N/A   45C    P8              3W /   60W |      15MiB /   4096MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A       933      G   /usr/lib/xorg/Xorg                              4MiB |
+-----------------------------------------------------------------------------------------+

安装docker(可以在安装系统时勾选上docker，这一步就可以省略，建议不要省略，安装系统勾选安装的docker 版本较低)

#安装 Docker 仓库密钥和工具
exxk@exxk:~$ sudo apt install -y apt-transport-https ca-certificates curl software-properties-common
exxk@exxk:~$ curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /usr/share/keyrings/docker-archive-keyring.gpg
#添加 Docker 仓库
exxk@exxk:~$ echo "deb [arch=amd64 signed-by=/usr/share/keyrings/docker-archive-keyring.gpg] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
exxk@exxk:~$ sudo apt update
#安装 Docker
exxk@exxk:~$ sudo apt install -y docker-ce docker-ce-cli containerd.io
#将当前用户添加到 Docker 组
exxk@exxk:~$ sudo usermod -aG docker $USER
exxk@exxk:~$ sudo reboot
#查看docker版本，如果没有添加当前用户到docker组，需要加sudo
exxk@exxk:~$ docker version
Client: Docker Engine - Community
 Version:           27.3.1
 API version:       1.47
 Go version:        go1.22.7
 Git commit:        ce12230
 Built:             Fri Sep 20 11:41:03 2024
 OS/Arch:           linux/amd64
 Context:           default

Server: Docker Engine - Community
 Engine:
  Version:          27.3.1
  API version:      1.47 (minimum version 1.24)
  Go version:       go1.22.7
  Git commit:       41ca978
  Built:            Fri Sep 20 11:41:03 2024
  OS/Arch:          linux/amd64
  Experimental:     false
 containerd:
  Version:          1.7.23
  GitCommit:        57f17b0a6295a39009d861b89e3b3b87b005ca27
 runc:
  Version:          1.1.14
  GitCommit:        v1.1.14-0-g2c9f560
 docker-init:
  Version:          0.19.0
  GitCommit:        de40ad0

安装 NVIDIA Container Toolkit

# 配置生产存储库：
exxk@exxk:~$ curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg \
  && curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list | \
    sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' | \
    sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
#（可选）配置存储库以使用实验性软件包：
exxk@exxk:~$ sudo sed -i -e '/experimental/ s/^#//g' /etc/apt/sources.list.d/nvidia-container-toolkit.list   
#从存储库更新包列表：
exxk@exxk:~$ sudo apt-get update
#安装 NVIDIA Container Toolkit 软件包：
exxk@exxk:~$ sudo apt-get install -y nvidia-container-toolkit
#配置并重启 Docker
exxk@exxk:~$ sudo nvidia-ctk runtime configure --runtime=docker
exxk@exxk:~$ sudo systemctl restart docker
#运行一个简单的容器测试 GPU 是否可用
exxk@exxk:~$ docker run --rm --gpus all nvidia/cuda:11.8.0-base-ubuntu20.04 nvidia-smi
Tue Nov 19 07:58:08 2024
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 560.35.03              Driver Version: 560.35.03      CUDA Version: 12.6     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 3050 ...    Off |   00000000:00:10.0 Off |                  N/A |
| N/A   39C    P8              3W /   60W |      15MiB /   4096MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
+-----------------------------------------------------------------------------------------+

Linux-Proxmox-VE

发表于 2024-11-15 更新于 2024-11-19 分类于 Linux Waline：

PVE简介

PVE(Proxmox Virtual Environment)：开源免费，支持虚拟机和容器，灵活性高，内置 ZFS 和集群功能。
ESXi(VMware ESXi)：商业化产品，专注虚拟机，高性能且与 VMware 全栈生态深度集成。

安装

环境准备

服务器或电脑：Legion Y7000P IAH7
启动盘：用ventoy制作的启动u盘
PVE安装包：proxmox-ve_8.2-2.iso

安装步骤

复制安装包proxmox-ve_8.2-2.iso到启动盘。
启动电脑(Legion Y7000P IAH7)时按F2，调整U盘为第一启动，如果没有U盘，可以换一个U盘制作，有的U盘会不识别。
进入ventoy引导界面，依次点击proxmox-ve_8.2-2.iso->Boot in normal node，如果该模式没有进入下一个界面，可以换一个启动模式。
进入proxmox安装界面，点击Install Proxmox VE (Graphical)进行界面安装，Terminal是终端安装模式。
一直点击下一步即可，注意下面几个选项
- country选择China
- 如果电脑有多个硬盘，注意选择要安装的硬盘
- 密码需要自己设置一个
- 网络配置主要选择有线网卡
  
  Hostname：pve.iexxk.io
  
  IP Address(CIDR)：分配一个未使用的ip，后面24默认不用改，我设置的是172.16.80.244/24
  
  Gateway：路由器的网关，一般为分配ip后面一位改为xx.xx.xx.1，我设置的是172.16.80.1
  
  DNS Server: 该网络能访问的一个dns服务器，我设置的是8.8.8.8
我的因为是笔记本电脑，安装完成后，显示的是命令行界面，可以直接在电脑上输入root及密码就能进入系统，该系统就相当于是个linux
正式使用，在其他电脑访问https://172.16.80.244:8006/，输入root及密码就能进入管理平台了

额外配置，针对笔记本，ssh 进172.16.80.244，用户名和密码同管理平台，执行下面命令

#设置合上盖子不休眠
sed -i '/^#*HandleLidSwitch/s/^#*//;s/\bHandleLidSwitch=\w*/HandleLidSwitch=ignore/' /etc/systemd/logind.conf
#设置合上盖子并外接显示器，不挂起(节能)
sed -i '/^#*HandleLidSwitchExternalPower/s/^#*//;s/\bHandleLidSwitchExternalPower=\w*/HandleLidSwitchExternalPower=ignore/' /etc/systemd/logind.conf
#重启服务，生效上面的配置
systemctl restart systemd-logind

平台使用

介绍

登录进去可以看到，资源使用情况，什么都还没安装的情况：

CPU：显示20个，使用0%

内存：1.35 GiB的15.41 GiB，使用9%

存储：2.42 GiB的442.75 GiB，使用1%

服务器视图下菜单目录介绍

|--数据中心
	 |--pve 	#表示一个节点，因为只安装了一台,这里取的是Hostname里面域名的前面部分
      |--localnetwork(pve) #网络管理
      |--local(pve)  #存储VZDump备份文件, ISO镜像, 容器模板等
      |--local-lvm(pve) #磁盘映像, 容器

基础使用

配置IOS镜像：点击数据中心->pve->local->ISO镜像，在ISO镜像页面，点击上传或从URL下载，将镜像放进PVE。
- Centos7 atomic镜像URL地址： CentOS-Atomic-Host-7-Installer.iso，CentOS Atomic 是一个专为执行 Docker 容器而设的轻量操作系统，它创建自标准的 CentOS 7 组件，并追随 Red Hat 企业级 Linux Atomic 主机的组件版本。
- win10 ISO – Enterprise LTSC镜像地址：LTSC 版本去掉了许多不必要的功能（例如 Microsoft Store、Cortana、动态磁贴等），非常适合追求性能的用户，系统本身对内存的占用较低。（window安装时注意看提示，按任何键加载CD磁盘）
- Ubuntu20.04镜像地址：Server install image服务器安装映像允许您在计算机上永久安装 Ubuntu 以用作服务器。它不会安装图形用户界面。

显卡直通：VM->硬件->添加->PCI设备->原始设备->选择RTX3050->添加。

如果是添加到win10的，windows更新界面更新即可使用显卡了。

~~进入PVE->Shell~~（废弃，以下shell命令都废弃，现在直接在界面上面配置就可以了，配置之后，Kernel driver in use自动就变成了vfio-pci）

#-----------------1. 修改-------------------------------------------------
root@pve:~# nano /etc/default/grub
#GRUB_CMDLINE_LINUX_DEFAULT="quiet" #注释该配置
GRUB_CMDLINE_LINUX_DEFAULT="quiet intel_iommu=on" #新增该配置
#按^O Write Out(写入保存)和^X Exit(退出nano)
#-----------------2. 检查显卡被谁在用，以及设备地址---------------------------
root@pve:~# lspci -nnk | grep -i nvidia -A 2
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GA107BM [GeForce RTX 3050 Mobile] [10de:25e2] (rev a1)
	Subsystem: Lenovo GA107BM [GeForce RTX 3050 Mobile] [17aa:382d]
	Kernel driver in use: nouveau
	Kernel modules: nvidiafb, nouveau
01:00.1 Audio device [0403]: NVIDIA Corporation Device [10de:2291] (rev a1)
	Kernel driver in use: snd_hda_intel
	Kernel modules: snd_hda_intel
#-----------------3. 添加需要直通的设备地址-----------------------------------
root@pve:~# nano /etc/modprobe.d/vfio.conf
#设置优先级
softdep nouveau pre: vfio-pci
softdep snd_hda_intel pre: vfio-pci
#绑定直通地址
options vfio-pci ids=10de:25e2,10de:2291
#-----------------4. 更新initramfs-----------------------------------------
root@pve:~# update-initramfs -u
#-----------------5. 重启--------------------------------------------------
root@pve:~# reboot
#-----------------6. 检查显卡被谁在使用，Kernel driver in use-----------------
root@pve:~# lspci -nnk | grep -i nvidia -A 2
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GA107BM [GeForce RTX 3050 Mobile] [10de:25e2] (rev a1)
	Subsystem: Lenovo GA107BM [GeForce RTX 3050 Mobile] [17aa:382d]
	Kernel driver in use: vfio-pci
	Kernel modules: nvidiafb, nouveau
01:00.1 Audio device [0403]: NVIDIA Corporation Device [10de:2291] (rev a1)
	Kernel driver in use: vfio-pci
	Kernel modules: snd_hda_intel

tools-ossrs

发表于 2024-11-01 更新于 2024-11-05 分类于 Tools Waline：

SRS是一个开源的（MIT协议）简单高效的实时视频服务器，支持RTMP、WebRTC、HLS、HTTP-FLV、SRT、MPEG-DASH和GB28181等协议。 SRS媒体服务器和FFmpeg、OBS、VLC、 WebRTC等客户端配合使用，提供流的接收和分发的能力，是一个典型的发布（推流）和订阅（播放）服务器模型。 SRS支持互联网广泛应用的音视频协议转换，比如可以将RTMP或SRT，转成HLS或HTTP-FLV或WebRTC等协议。

k8s安装

相关设置：
Container Image: ossrs/srs:5
Command/Args:
	命令: ./objs/srs -c conf/flv/rtmp2rtc.conf #配置见文尾
#	命令: ./objs/srs -c conf/flv/realtime.flv.conf //废弃，由上面的配置替换
Environments: 
	CANDIDATE: iexxk.com #替换成域名，这样能同时兼容内外网，内网做个host映射可以直接走内网
  TZ: Asia/Shanghai
Container Ports(Host端口): 1935  默认推流端口，地面站那边修改不了，导致只能用host进行映射暴露
数据卷:
	配置目录: /usr/local/srs/conf/flv
	数据目录:/usr/local/srs/srs_dvr_data
NodePort端口映射: 
	1935 #默认推流端口
#	18080 // realtime.flv.conf配置才需要
	8080 #web、安卓拉流
	1985 #ios拉流（webrtc）
	30353 #udp端口，webrtc拉流需要

nginx代理

# ossrs代理地址
upstream ossrs_server {
  		server ossrs:8080;
}
location /view-live/ {
      rewrite ^/view-live/(.*) /$1 break;
      proxy_pass http://ossrs_server;
      proxy_set_header Host $host;
      proxy_set_header X-Real-IP $remote_addr;
      proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
  }

流程

虚拟ip处需要映射

Web: 外网ip80/443绑定节点的ingress的80/443

Ossrs: 外网ip1935绑定k8s里面的ossrs服务的Container Ports(Host端口)1935或者nodeport，不采用nodeport是因为要内网外网都要能访问ossrs，关键是内网对外网域名做了隔离，因此这里用本机的hosts映射（可以升级到dns下发）

graph LR
G[OBS推流]-->C
C[域名] -->|解析| A[移动光猫]
A -->|LAN| B(飞塔防火墙)
B -->|虚拟ip| E[K8S的Ingress]
E -->|域名匹配| F[前端web/nginx]
F -->|nginx代理|U[ossrs]
C-->|hosts映射节点ip|E

V[VLC]-->|web flv拉流|U
K[IOS]-->|ios webrtc拉流|U

测试

安装obs，用srs在线的测试工具也可以。
在来源->加号->浏览器->url里面输入要直播的网址https://www.baidu.com/s?word=在线时间
在控制按钮->设置->直播->

服务：自定义

服务器：rtmp://iexxk.com/drone/

推码流：1
点击开始直播

播放：

内网测试(ok)：

Flv播放，打开http://ip:8080/players/srs_player.html，url输入：http://ip:8080/drone/1.flv
Webrtc(WHEP)播放，打开http://ip:8080/players/whep.html，url输入：http://ip:1985/rtc/v1/whep/?app=drone&stream=1

内网域名测试(ok)：

Flv播放，打开http://iexxk.io/view-live/players/srs_player.html，url输入：http://iexxk.io/view-live/drone/1.flv
Webrtc(WHEP)播放，打开http://iexxk.io/view-live/players/whep.html，url输入：http://iexxk.io/view-webrtc/rtc/v1/whep/?app=drone&stream=1

外网测试：

Flv播放，打开https://iexxk.com/view-live/players/srs_player.html，url输入：https://iexxk.com/view-live/drone/1.flv
Webrtc(WHEP)播放，打开https://iexxk.com/view-live/players/whep.html，url输入：https://iexxk.com/view-webrtc/rtc/v1/whep/?app=drone&stream=1
Webrtc(rtc播放器)播放，打开OSSRS-RTC在线播放器，url输入： webrtc://iexxk.com/drone/1

webrtc和WHEP区别

WHEP：可以代理，但是也无法将一个域名根据环境解析为内网ip和外网ip，内网和外网是通过dns解析的，只能解析一个。在rtc里面启用保留域名，也无法解决，candidate里面会出现2个域名一个ip，优先还是走的ip

rtc_server {
    enabled on;
    listen 30353; # UDP port
    # @see https://ossrs.net/lts/zh-cn/docs/v4/doc/webrtc#config-candidate
    candidate $CANDIDATE;
    # 启用保留域名
    keep_api_domain on;
}

webrtc：直接连接，延迟更低，必须开放1985，无法修改

使用webrtc时，需要在web前端(nginx)做个代理，并要将该端口映射到外网1985上

# ossrs webrtc代理地址
upstream ossrs_webrtc {
  server ossrs:1985;
}   
   server {
        listen       1985 ssl;
        server_name  iexxk.com;

        ssl_certificate /etc/nginx/cert/iexxk.com.pem;
        ssl_certificate_key /etc/nginx/cert/iexxk.com.key;

        location /rtc/v1/play/ {
            proxy_pass http://ossrs_webrtc;
            proxy_http_version 1.1;
            proxy_set_header Upgrade $http_upgrade;
            proxy_set_header Connection "upgrade";
            proxy_set_header Host $host;
            proxy_set_header X-Real-IP $remote_addr;
            proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
            proxy_set_header X-Forwarded-Proto $scheme;
        }
    }

常见问题

在内网无法访问外网域名的时候，虽然通过host映射到内网，但是webrtc播放的时候还是解析成了外网ip

使用webrtc播放时可以在浏览器控制台能看到如下信息

Got answer:  v=0
o=SRS/5.0.213(Bee) 140129754604224 2 IN IP4 0.0.0.0
s=SRSPlaySession
t=0 0
a=ice-lite
a=group:BUNDLE 0 1
a=msid-semantic: WMS drone/1
m=audio 9 UDP/TLS/RTP/SAVPF 111
c=IN IP4 0.0.0.0
a=ice-ufrag:881501t3
a=ice-pwd:d11em20o3z9k02c97228yq5937ptj776
a=fingerprint:sha-256 35:65:58:9A:E6:78:8F:D5:E6:F1:8F:E5:74:AD:A4:1E:3A:DF:B4:CE:21:3C:C9:CB:4A:9D:25:7A:87:53:FB:47
a=setup:passive
a=mid:0
a=extmap:3 http://www.ietf.org/id/draft-holmer-rmcat-transport-wide-cc-extensions-01
a=sendonly
a=rtcp-mux
a=rtcp-rsize
a=rtpmap:111 opus/48000/2
a=rtcp-fb:111 transport-cc
a=ssrc:10186 cname:2yaj11875760968e
a=ssrc:10186 label:audio-803h25y4
a=candidate:0 1 udp 2130706431 172.16.10.44 30353 typ host generation 0
a=candidate:1 1 udp 2130706431 iexxk.com 30353 typ host generation 0
m=video 9 UDP/TLS/RTP/SAVPF 106
c=IN IP4 0.0.0.0
a=ice-ufrag:881501t3
a=ice-pwd:d11em20o3z9k02c97228yq5937ptj776
a=fingerprint:sha-256 35:65:58:9A:E6:78:8F:D5:E6:F1:8F:E5:74:AD:A4:1E:3A:DF:B4:CE:21:3C:C9:CB:4A:9D:25:7A:87:53:FB:47
a=setup:passive
a=mid:1
a=extmap:3 http://www.ietf.org/id/draft-holmer-rmcat-transport-wide-cc-extensions-01
a=sendonly
a=rtcp-mux
a=rtcp-rsize
a=rtpmap:106 H264/90000
a=rtcp-fb:106 transport-cc
a=rtcp-fb:106 nack
a=rtcp-fb:106 nack pli
a=fmtp:106 level-asymmetry-allowed=1;packetization-mode=1;profile-level-id=42e01f
a=ssrc:10187 cname:2yaj11875760968e
a=ssrc:10187 label:video-29515r30
a=candidate:0 1 udp 2130706431 172.16.10.44 30353 typ host generation 0
a=candidate:1 1 udp 2130706431 iexxk.com 30353 typ host generation 0

注意最后两行

使用内网ip的时候，输出如下信息
a=candidate:0 1 udp 2130706431 172.16.10.102 30353 typ host generation 0
a=candidate:1 1 udp 2130706431 iexxk.com 30353 typ host generation 0
使用内网域名的时候，输出如下信息,内网做了dns解析，假域名能解析到内网主机
a=candidate:0 1 udp 2130706431 172.16.10.44 30353 typ host generation 0
a=candidate:1 1 udp 2130706431 iexxk.com 30353 typ host generation 0
使用外网域名的时候，输出如下信息，17.176.159.182这个域名刚好在内网无法访问，因此内网无法播放外网地址
a=candidate:0 1 udp 2130706431 17.176.159.182 30353 typ host generation 0
a=candidate:1 1 udp 2130706431 exxk.com 30353 typ host generation 0

真域名做了host映射，但是似乎不会从host拿，应该是从dns里面拿取了

从下图可以知道，是通过1985的接口，拿到了udp的端口数据。

配置

rtmp2rtc.conf

listen              1935;
max_connections     1000;
daemon              off;
srs_log_tank        console;

http_server {
    enabled         on;
    listen          8080;
    dir             ./objs/nginx/html;
}

http_api {
    enabled         on;
    listen          1985;
}
stats {
    network         0;
}
rtc_server {
    enabled on;
    listen 30353; # UDP port
    # @see https://ossrs.net/lts/zh-cn/docs/v4/doc/webrtc#config-candidate
    candidate $CANDIDATE;
}

vhost __defaultVhost__ {
    rtc {
        enabled     on;
        # @see https://ossrs.net/lts/zh-cn/docs/v4/doc/webrtc#rtmp-to-rtc
        rtmp_to_rtc on;
        # @see https://ossrs.net/lts/zh-cn/docs/v4/doc/webrtc#rtc-to-rtmp
        rtc_to_rtmp on;
    }
    http_remux {
        enabled     on;
        mount       [vhost]/[app]/[stream].flv;
    }
   
	hls {
		enabled on;
		hls_path ./objs/nginx/html;
		hls_fragment 10;
		hls_window 20;
		hls_wait_keyframe off;
	}

	tcp_nodelay on min_latency on;

	play {
		gop_cache off;
		queue_length 10;
		mw_latency 100;
	}

	publish {
		mr off;
	}

	dvr {
		enabled on;
		dvr_apply all;
		dvr_plan session;
		dvr_path ./srs_dvr_data/[app]/[stream]/[2006]-[01]-[02]/[timestamp].[2006]-[01]-[02]_[15].[04].[05]_[999].mp4;
		dvr_duration 30;
		dvr_wait_keyframe on;
		time_jitter full;
	}
	    
}

~~realtime.flv.conf~~

listen 1935;
max_connections 1000;
daemon off;
srs_log_tank console;
http_server {
	enabled on;
	listen 18080;
	dir ./objs/nginx/html;
}
vhost __defaultVhost__ {
	http_remux {
		enabled on;
		mount [vhost]/[app]/[stream].flv;
	}

	hls {
		enabled on;
		hls_path ./objs/nginx/html;
		hls_fragment 10;
		hls_window 20;
		hls_wait_keyframe off;
	}

	tcp_nodelay on min_latency on;

	play {
		gop_cache off;
		queue_length 10;
		mw_latency 100;
	}

	publish {
		mr off;
	}

	dvr {
		enabled on;
		dvr_apply all;
		dvr_plan session;
		dvr_path ./srs_dvr_data/[app]/[stream]/[2006]-[01]-[02]/[timestamp].[2006]-[01]-[02]_[15].[04].[05]_[999].mp4;
		dvr_duration 30;
		dvr_wait_keyframe on;
		time_jitter full;
	}
}

K8s-ingress-domain

发表于 2024-10-29 更新于 2024-11-01 分类于 k8s Waline：

基础环境信息

飞塔防火墙版本：FortiGate-100F

kubernates版本： v1.26.4

k8s界面工具kuboard：v3.5.2.4

网络架构图：

graph LR
C[域名] -->|解析| A[移动光猫]
A -->|LAN| B(飞塔防火墙)
B -->|虚拟ip| E[K8S的Ingress]
E -->|域名匹配| F[应用路由]

飞塔防火墙配置步骤：

策略&对象->虚拟IP->新建

---------------------新建配置1------------------------------
虚拟IP/服务器：	k8s_ingress_https
接口：	移动出口2 (port1)
映射自：	17.176.159.182(移动光猫的外网ip)
映射的IP地址/范围：	172.16.10.44(k8s任意一个节点的ip)
外部服务端口：	TCP 443 (移动光猫外网端口，需要备案)
映射到端口：	443 (ingress对应的端口)
----------------------新建配置2------------------------------
虚拟IP/服务器：	k8s_ingress_http
接口：	移动出口2 (port1)
映射自：	17.176.159.182(移动光猫的外网ip)
映射的IP地址/范围：	172.16.10.44(k8s任意一个节点的ip)
外部服务端口：	TCP 80 (移动光猫外网端口，需要备案)
映射到端口：	80 (ingress对应的端口)

~~配置https证书（可选-废弃，后面在k8s进行管理证书，这里管理比较麻烦）~~
- 系统管理->证书->新建->证书->导入证书->证书->上传申请的免费证书
- 安全配置文件->SSL/SSH检测->新建
1
2
3
名称： k8s_ingress_ssl
启用SSL检测：保护SSL服务器
服务器证书：上一步上传的证书

策略&对象->虚拟IP->新建

名称： k8s_ingress
流入接口： 移动出口2
流出接口： lan
源地址： all
目标地址： k8s_ingress_https,k8s_ingress_http(上一步的两个虚拟ip)
计划任务： always
服务： ALL
启用NAT： 关闭
SSL检测： no-inspection(如果要配置证书，在这里进行关联k8s_ingress_ssl)
策略过期：关闭

k8s-ingress配置

新建应用路由，配置成阿里云购买的域名，并在阿里云上面做域名解析，解析到k8s的ingress所在的节点ip端口上，开启https证书，上传申请的免费证书

K8s-Endpoint

发表于 2024-08-15 更新于 2024-10-30 分类于 k8s Waline：

endpoint代理k8s之外的服务，让其能在k8s内访问，或可以通过ingress进行访问。

待补充。

Java-K8s-Quarkus

发表于 2024-07-23 分类于 k8s Waline：

快速开始

#安装 ---废弃开始----直接可以在idea里面新建quarkus的项目
➜  ~ brew install quarkusio/tap/quarkus
#检查版本
➜  ~ quarkus --version
3.12.3
#更新版本
➜  ~ brew update
➜  ~ brew upgrade quarkus
#---废弃结束----直接可以在idea里面新建quarkus的项目

记一次从vmdk镜像里面恢复数据的历程

发表于 2024-07-15 更新于 2024-07-29 分类于 Linux Waline：

问题描述

环境： ESXi 7.0+centos7虚拟机

在一次操作生产服务器时，发生了系统文件丢失，导致重启之后无法进入系统，但是重要哦的文件还在系统中，第一时间应该先对当前系统【创建快照】，创建快照之后就可以【查看vmdk文件】

解决方案一：导出vmdk文件进行打开（废弃）

首先需要导出vmdk文件，可以【开启ESXI服务器的ssh功能】，然后使用scp root@<ESXi_IP>:/vmfs/volumes/datastore_name/vm_name/disk.vmdk /local/path/拷贝出来，如果文件太大，可以【创建虚拟机】作为临时存放，内网相对快很多
打开vmdk文件，试了各种都不能打开，但是也许window系统下有工具能打开

解决方案二：创建虚拟机加载vmdk为第二硬盘

【创建虚拟机】，虚拟机需要自定义分区，主要修改卷组（VG）名称，修改卷组名称，主要是防止逻辑卷名称冲突
启动新的虚拟机，启动成功后，【添加硬盘】，硬盘文件选择需要打开的vmdk文件

进入虚拟机，执行下面命令

#检查激活状态
#有冲突就不会激活
[root@localhost ~]# lvscan
  inactive          '/dev/centos/swap' [15.75 GiB] inherit
  inactive          '/dev/centos/home' [1.93 TiB] inherit
  inactive          '/dev/centos/root' [50.00 GiB] inherit 
  ACTIVE            '/dev/centos/swap' [2.00 GiB] inherit
  ACTIVE            '/dev/centos/home' [346.99 GiB] inherit
  ACTIVE            '/dev/centos/root' [50.00 GiB] inherit
 #没有冲突会自动激活 
 [root@localhost ~]# lvscan
  ACTIVE          '/dev/centos/swap' [15.75 GiB] inherit
  ACTIVE          '/dev/centos/home' [1.93 TiB] inherit
  ACTIVE          '/dev/centos/root' [50.00 GiB] inherit 
  ACTIVE            '/dev/os/swap' [2.00 GiB] inherit
  ACTIVE            '/dev/os/home' [346.99 GiB] inherit
  ACTIVE            '/dev/os/root' [50.00 GiB] inherit
#（可选不建议）如果忘了修改卷组（VG）名称，home目录可以在虚拟机里面修改，其他目录不行
umount /dev/centos/home #卸载
lvrename /dev/centos/home /dev/centos/home_old #重命名
mount /dev/centos/home_old /home
#挂载需要打开的vmdk之后执行下面的，激活命令
lvchange -ay /dev/centos/home
#--可选不建议结束---------
#挂载目录
mount /dev/centos/home /mnt/home
#然后就可以进行拷贝备份了

VMware ESXI常用基础操作

创建快照：选择虚拟机，点击 Actions > Snapshot > Take Snapshot。

查看vmdk文件：在左侧菜单，点击Storage>Data_disk>对应的虚拟机服务

#关键文件内容如下
server.vmdk # 基础文件
server-000001.vmdk #创建快照产生，从基础文件开始记录所有新的更改
server-000002.vmdk #纪录新的改变

开启ESXI服务器的ssh功能：在左侧菜单，点击Host>Manage>Services，里面开启ssh功能。开启后用户名和密码同网页登录esxi界面一样
创建虚拟机：在左侧菜单，点击Virtual Machines>create，选择linux，centos7，在iso里面选择Storage里面的iso镜像，然后就是安装centos的步骤了。设置静态ip和ESXI服务器在同一网段中。
添加硬盘：选择虚拟机，点击 edit > add hard disk 进行添加硬盘。
合并快照磁盘：选择虚拟机，点击 Actions > Snapshot >Snapshot Manager> Consolidate。

Docker安装使用ES

发表于 2024-07-09 更新于 2024-10-30 分类于 Docker Waline：

安装

elasticsearch安装

在docker.elastic.co仓库查找springboot对应的es版本,，eg:

# springboot 2.7.18
docker pull docker.elastic.co/elasticsearch/elasticsearch:7.17.22
#如果上面的拉取太慢了，可以用下面的dockerhub的镜像
docker pull elastic/elasticsearch:7.17.22

创建配置文件，执行vim /workspace/elasticsearch.yml写入如下内容

cluster.name: "docker-cluster"
network.host: 0.0.0.0
http.port: 9200
http.cors.enabled: true
http.cors.allow-origin: "*"
xpack.security.enabled: false

安装elasticsearch7.17，执行

mkdir /workspace/v_es_data/
chmod -R 777 /workspace/v_es_data/
# -m限制使用内存1g
docker run -d -m 1g -p 9200:9200 -e "discovery.type=single-node" --name es \
-v /workspace/v_es_data:/usr/share/elasticsearch/data \
-v /workspace/elasticsearch.yml:/usr/share/elasticsearch/config/elasticsearch.yml \
docker.elastic.co/elasticsearch/elasticsearch:7.17.22

测试，执行下面命令

1
2
3

➜  ~ curl -X GET "127.0.0.1:9200/_cat/nodes?v=true&pretty"
ip         heap.percent ram.percent cpu load_1m load_5m load_15m node.role   master name
172.18.0.2           35          84   5    0.06    0.09     0.13 cdfhilmrstw *      eeeedfa1ed88

kibana安装（没采用）

可视化查看工具，类似dbeaver，非必需

执行下面命令进行安装

1
2
3

docker run -d --name kibana -p 5601:5601 \
-e "ELASTICSEARCH_HOSTS=http://127.0.0.1:9200" \
docker.elastic.co/kibana/kibana:7.17.22

访问http://localhost:5601/
然后进行各种操作，比较复杂

本地采用es-client插件

数据同步方案

应用层同步

在应用层面进行数据同步，即在应用代码中同时写入 MySQL 和 Elasticsearch。

优点：依赖少，自定义同步逻辑。

缺点：数据一致性和可靠性方面存在一些挑战，特别是在高并发的场景下。

MySQL binlog(数据库日志)

借助于数据库日志（如 MySQL binlog）来捕获数据变化，并将其同步到 Elasticsearch。这种方法通常能够实时同步数据，并且不会对数据库性能造成太大影响。

优点：实时同步数据，不会对数据库性能造成太大影响。

缺点：性能开销，binlog 会占用额外的磁盘空间。

Debezium(重量级别)

优点：功能强大，支持复杂的配置和扩展，社区活跃。

缺点：需要 Kafka，配置相对复杂。

Maxwell(轻量级别)

Maxwell是一种开源的MySQL数据库同步工具，它可以将MySQL数据库的binlog转化为JSON格式，并将其发送到消息队列中。

需要mq和kafka等中间件

优点：轻量级，配置简单，适合快速实现。

缺点：功能相对简单，不支持复杂的场景

canal

Canal是阿里巴巴开发的一款数据库同步工具，它可以实现MySQL数据库的binlog解析和日志抓取。模拟mysql的从节点实现。

springboot接入ES

在pom.xml添加依赖

<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-data-elasticsearch</artifactId>
</dependency>

maxwell

version: '3.8'

services:
  mysql:
    image: mysql:8.0
    container_name: mysql
    environment:
      MYSQL_ROOT_PASSWORD: rootpassword
      MYSQL_DATABASE: maxwell
      MYSQL_USER: maxwell
      MYSQL_PASSWORD: maxwellpassword
    ports:
      - "3306:3306"
    command: --server-id=1 --log-bin=mysql-bin --binlog-format=row --binlog-row-image=full
    volumes:
      - mysql-data:/var/lib/mysql

  maxwell:
    image: zendesk/maxwell:1.33.0
    container_name: maxwell
    environment:
      - MAXWELL_MYSQL_HOST=mysql
      - MAXWELL_MYSQL_USER=maxwell
      - MAXWELL_MYSQL_PASSWORD=maxwellpassword
      - MAXWELL_MYSQL_DATABASE=maxwell
      - MAXWELL_PRODUCER=es
      - MAXWELL_OUTPUT_DDL=true
      - MAXWELL_ES_HOST=http://119.91.96.139:9200
    depends_on:
      - mysql

volumes:
  mysql-data:

参考

4种数据同步到Elasticsearch方案

Linux-install-cuda

发表于 2024-05-10 更新于 2024-05-16 分类于 Linux Waline：

centos7如何安装CUDA

查看显卡型号lspci | grep -i vga，提示-bash: lspci: 未找到命令，执行yum install pciutils安装lspci命令
1
2
[root@exxk ~]# lspci | grep -i vga
01:00.0 VGA compatible controller: NVIDIA Corporation Device 25e2 (rev a1)
在输出结果中获取25e2关键字，在https://admin.pci-ids.ucw.cz/read/PC/10de/连接尾部拼接25e2，例如https://admin.pci-ids.ucw.cz/read/PC/10de/25e2，访问即可，页面回展示类似内容
1
2
3
Main -> PCI Devices -> Vendor 10de -> Device 10de:25e2

Name: GA107BM [GeForce RTX 3050 Mobile]
从中可以知道25e2对应的显卡是GeForce RTX 3050 Mobile
在该https://www.nvidia.cn/Download/index.aspx?lang=cn网页找到`GeForce RTX 3050 Mobile=GeForce RTX 3050 Laptop GPU该驱动，选择Linux64-bit然后下载下来 NVIDIA-Linux-x86_64-550.78.run `

拷贝到centos7上面，执行下面命令进行安装驱动

#屏蔽系统自带的nouveau显卡驱动（重启后生效---begin-----------
[root@exxk ~]# lsmod | grep nouveau
nouveau              2543616  0
drm_ttm_helper         12288  1 nouveau
ttm                    90112  2 drm_ttm_helper,nouveau
i2c_algo_bit           12288  1 nouveau
mxm_wmi                12288  1 nouveau
drm_display_helper    176128  1 nouveau
drm_kms_helper        225280  4 drm_display_helper,nouveau
drm                   643072  6 drm_kms_helper,drm_display_helper,drm_ttm_helper,ttm,nouveau
video                  69632  2 ideapad_laptop,nouveau
wmi                    36864  5 video,wmi_bmof,ideapad_laptop,mxm_wmi,nouveau
[root@exxk ~]# vi /lib/modprobe.d/dist-blacklist.conf
#将nvidiafb注释掉:
#blacklist nvidiafb 
#然后添加以下语句：
#blacklist nouveau
#options nouveau modeset=0
[root@exxk ~]# mv /boot/initramfs-$(uname -r).img /boot/initramfs-$(uname -r).img.bak
[root@exxk ~]# dracut /boot/initramfs-$(uname -r).img $(uname -r)
[root@exxk ~]# reboot
[root@exxk ~]# lsmod | grep nouveau
#重启之后发现lsmod | grep nouveau命令无任何输出代表成功
#屏蔽系统自带的nouveau显卡驱动（重启后生效---end-----------

#安装依赖环境------------begin--------------------------
#安装kernel-devel kernel-headers gcc
[root@exxk ~]# yum install kernel-devel gcc -y
[root@exxk ~]# uname -r
3.10.0-1160.118.1.el7.x86_64
[root@exxk ~]# yum info kernel-devel kernel-headers | grep 发布
发布    ：1160.118.1.el7
发布    ：1160.118.1.el7
#内核版本要和安装的版本一致
#如果不一致，检查内核是否需要升级
[root@exxk ~]# yum check-update kernel
#查看可用内核
[root@exxk ~]# cat /boot/grub2/grub.cfg |grep menuentry
#升级到和安装工具一致的内核版本
[root@exxk ~]# grub2-set-default 'CentOS Linux (3.10.0-1160.118.1.el7.x86_64) 7 (Core)'
[root@exxk ~]# reboot
#安装依赖环境------------end--------------------------

#安装显卡驱动，一致点同意即可
[root@exxk ~]# ./NVIDIA-Linux-x86_64-550.78.run
#查看驱动信息
[root@exxk ~]# nvidia-smi
Sat May 11 14:43:05 2024
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 550.78                 Driver Version: 550.78         CUDA Version: 12.4     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 3050 ...    Off |   00000000:01:00.0 Off |                  N/A |
| N/A   47C    P0             10W /   60W |       1MiB /   4096MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+

+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|  No running processes found                                                             |
+-----------------------------------------------------------------------------------------+

安装NVIDIA Container Toolkit，执行下面的命令(确保您已安装适用于您的 Linux 发行版的NVIDIA 驱动程序 请注意，您不需要在主机系统上安装 CUDA Toolkit，但需要安装 NVIDIA 驱动程序)

[root@exxk ~]# curl -s -L https://nvidia.github.io/libnvidia-container/stable/rpm/nvidia-container-toolkit.repo | sudo tee /etc/yum.repos.d/nvidia-container-toolkit.repo
[root@exxk ~]# yum-config-manager --enable nvidia-container-toolkit-experimental
yum install -y nvidia-container-toolkit
[root@exxk ~]# yum install -y nvidia-container-toolkit

配置containerd（针对Kubernetes），执行下面命令

[root@exxk ~]# nvidia-ctk runtime configure --runtime=containerd
INFO[0000] Loading config from /etc/containerd/config.toml
INFO[0000] Wrote updated config to /etc/containerd/config.toml
INFO[0000] It is recommended that containerd daemon be restarted.
[root@exxk ~]# systemctl restart containerd

安装 PaddlePaddle 镜像

1	nvidia-docker pull registry.baidubce.com/paddlepaddle/paddle:2.6.1-gpu-cuda12.0-cudnn8.9-trt8.6

新版cuda安装(废弃采用容器安装)

安装cuda 12.4.1最新版(飞浆需要12.0的版本，因此将下面的替换成12.0.1)

wget https://developer.download.nvidia.com/compute/cuda/12.4.1/local_installers/cuda_12.4.1_550.54.15_linux.run
sh cuda_12.4.1_550.54.15_linux.run
┌──────────────────────────────────────────────────────────────────────────────┐
│ CUDA Installer                                                               │
│ - [ ] Driver                                                                 │
│      [ ] 550.54.15                                                           │
│ + [X] CUDA Toolkit 12.4                                                      │
│   [X] CUDA Demo Suite 12.4                                                   │
│   [X] CUDA Documentation 12.4                                                │
│ - [ ] Kernel Objects                                                         │
│      [ ] nvidia-fs                                                           │
│   Options                                                                    │
│   Install                                                                    │
│                                                                              │
│ Up/Down: Move | Left/Right: Expand | 'Enter': Select | 'A': Advanced options │
└──────────────────────────────────────────────────────────────────────────────┘
===========
= Summary =
===========

Driver:   Not Selected
Toolkit:  Installed in /usr/local/cuda-12.4/

Please make sure that
 -   PATH includes /usr/local/cuda-12.4/bin
 -   LD_LIBRARY_PATH includes /usr/local/cuda-12.4/lib64, or, add /usr/local/cuda-12.4/lib64 to /etc/ld.so.conf and run ldconfig as root

To uninstall the CUDA Toolkit, run cuda-uninstaller in /usr/local/cuda-12.4/bin
***WARNING: Incomplete installation! This installation did not install the CUDA Driver. A driver of version at least 550.00 is required for CUDA 12.4 functionality to work.
To install the driver using this installer, run the following command, replacing <CudaInstaller> with the name of this run file:
    sudo <CudaInstaller>.run --silent --driver

Logfile is /var/log/cuda-installer.log

添加环境变量，修改/etc/profile文件，执行vi /etc/profile，增加如下内容，然后执行source /etc/profile
1
2
export PATH=${PATH}:/usr/local/cuda-12.4/bin
export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:/usr/local/cuda-12.4/lib64

验证

[root@exxk ~]# nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2024 NVIDIA Corporation
Built on Thu_Mar_28_02:18:24_PDT_2024
Cuda compilation tools, release 12.4, V12.4.131
Build cuda_12.4.r12.4/compiler.34097967_0

卸载

[root@exxk ~]# cd /usr/local/cuda-12.4/bin
[root@exxk bin]# ./cuda-uninstaller
┌──────────────────────────────────────────────────────────────────────────────┐
│ CUDA Uninstaller                                                             │
│   [X] CUDA_Toolkit_12.4                                                      │
│   [X] CUDA_Demo_Suite_12.4                                                   │
│   [X] CUDA_Documentation_12.4                                                │
│   Done                                                                       │
│                                                                              │
│ Up/Down: Move | 'Enter': Select                                              │
└──────────────────────────────────────────────────────────────────────────────┘
 Successfully uninstalled
[root@exxk local]# cd /usr/local/
[root@exxk local]# rm -rf cuda-12.4

旧版cuda废弃

下载 developer.nvidia.com页面按自己的环境选择，然后下载cuda。安装方式我选择的runfile(local)，下载完成后上传到centos

执行sh cuda_7.5.18_linux.run，按ctrl+c跳过文档阅读

[root@exxk ~]# sh cuda_7.5.18_linux.run
Logging to /tmp/cuda_install_22223.log
Using more to view the EULA.
End User License Agreement
--------------------------
--More--(0%)
Do you accept the previously read EULA? (accept/decline/quit): accept
# 非常关键，我们已经在之前安装了高版本的驱动，这个千万别装。
Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 352.39? ((y)es/(n)o/(q)uit): n
Install the CUDA 7.5 Toolkit? ((y)es/(n)o/(q)uit): y
Enter Toolkit Location [ default is /usr/local/cuda-7.5 ]:
Do you want to install a symbolic link at /usr/local/cuda? ((y)es/(n)o/(q)uit): y
Install the CUDA 7.5 Samples? ((y)es/(n)o/(q)uit): y
Enter CUDA Samples Location [ default is /root ]:
Installing the CUDA Toolkit in /usr/local/cuda-7.5 ...
Missing recommended library: libGLU.so
Missing recommended library: libX11.so
Missing recommended library: libXi.so
Missing recommended library: libXmu.so

Installing the CUDA Samples in /root ...
Copying samples to /root/NVIDIA_CUDA-7.5_Samples now...
Finished copying samples.

===========
= Summary =
===========

Driver:   Not Selected
Toolkit:  Installed in /usr/local/cuda-7.5
Samples:  Installed in /root, but missing recommended libraries

Please make sure that
 -   PATH includes /usr/local/cuda-7.5/bin
 -   LD_LIBRARY_PATH includes /usr/local/cuda-7.5/lib64, or, add /usr/local/cuda-7.5/lib64 to /etc/ld.so.conf and run ldconfig as root

To uninstall the CUDA Toolkit, run the uninstall script in /usr/local/cuda-7.5/bin
To uninstall the NVIDIA Driver, run nvidia-uninstall

Please see CUDA_Installation_Guide_Linux.pdf in /usr/local/cuda-7.5/doc/pdf for detailed information on setting up CUDA.

***WARNING: Incomplete installation! This installation did not install the CUDA Driver. A driver of version at least 352.00 is required for CUDA 7.5 functionality to work.
To install the driver using this installer, run the following command, replacing <CudaInstaller> with the name of this run file:
    sudo <CudaInstaller>.run -silent -driver

Logfile is /tmp/cuda_install_22223.log
Signal caught, cleaning up

添加环境变量，修改/etc/profile文件，执行vi /etc/profile，增加如下内容，然后执行source /etc/profile
1
2
export PATH=${PATH}:/usr/local/cuda/bin
export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:/usr/local/cuda/lib64

验证环境

[root@exxk ~]# nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2015 NVIDIA Corporation
Built on Tue_Aug_11_14:27:32_CDT_2015
Cuda compilation tools, release 7.5, V7.5.17
# 如果前面安装了CUDA的example，这里可以执行如下操作：
[root@exxk ~]# cd /root/NVIDIA_CUDA-7.5_Samples/1_Utilities/deviceQuery
[root@exxk deviceQuery]# yum install gcc gcc-c++
[root@exxk deviceQuery]# make
"/usr/local/cuda-7.5"/bin/nvcc -ccbin g++ -I../../common/inc  -m64    -gencode arch=compute_20,code=sm_20 -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode arch=compute_52,code=compute_52 -o deviceQuery.o -c deviceQuery.cpp
"/usr/local/cuda-7.5"/bin/nvcc -ccbin g++   -m64      -gencode arch=compute_20,code=sm_20 -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode arch=compute_52,code=compute_52 -o deviceQuery deviceQuery.o
mkdir -p ../../bin/x86_64/linux/release
cp deviceQuery ../../bin/x86_64/linux/release
[root@exxk deviceQuery]# ./deviceQuery
./deviceQuery Starting...

 CUDA Device Query (Runtime API) version (CUDART static linking)

Detected 1 CUDA Capable device(s)

Device 0: "NVIDIA GeForce RTX 3050 Laptop GPU"
  CUDA Driver Version / Runtime Version          12.4 / 7.5
  CUDA Capability Major/Minor version number:    8.6
  Total amount of global memory:                 3873 MBytes (4060872704 bytes)
MapSMtoCores for SM 8.6 is undefined.  Default to use 128 Cores/SM
MapSMtoCores for SM 8.6 is undefined.  Default to use 128 Cores/SM
  (16) Multiprocessors, (128) CUDA Cores/MP:     2048 CUDA Cores
  GPU Max Clock rate:                            1500 MHz (1.50 GHz)
  Memory Clock rate:                             6001 Mhz
  Memory Bus Width:                              128-bit
  L2 Cache Size:                                 1572864 bytes
  Maximum Texture Dimension Size (x,y,z)         1D=(131072), 2D=(131072, 65536), 3D=(16384, 16384, 16384)
  Maximum Layered 1D Texture Size, (num) layers  1D=(32768), 2048 layers
  Maximum Layered 2D Texture Size, (num) layers  2D=(32768, 32768), 2048 layers
  Total amount of constant memory:               65536 bytes
  Total amount of shared memory per block:       49152 bytes
  Total number of registers available per block: 65536
  Warp size:                                     32
  Maximum number of threads per multiprocessor:  1536
  Maximum number of threads per block:           1024
  Max dimension size of a thread block (x,y,z): (1024, 1024, 64)
  Max dimension size of a grid size    (x,y,z): (2147483647, 65535, 65535)
  Maximum memory pitch:                          2147483647 bytes
  Texture alignment:                             512 bytes
  Concurrent copy and kernel execution:          Yes with 2 copy engine(s)
  Run time limit on kernels:                     No
  Integrated GPU sharing Host Memory:            No
  Support host page-locked memory mapping:       Yes
  Alignment requirement for Surfaces:            Yes
  Device has ECC support:                        Disabled
  Device supports Unified Addressing (UVA):      Yes
  Device PCI Domain ID / Bus ID / location ID:   0 / 1 / 0
  Compute Mode:
     < Default (multiple host threads can use ::cudaSetDevice() with device simultaneously) >

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 12.4, CUDA Runtime Version = 7.5, NumDevs = 1, Device0 = NVIDIA GeForce RTX 3050 Laptop GPU
Result = PASS

卸载 /usr/local/cuda-7.5/bin/uninstall_cuda_7.5.pl

参考

linux服务器上查看显卡(nvidia)型号

Centos7.9安装Nvidia驱动

CentOS7安装NVIDIA显卡驱动

CentOS 7安装GPU、Cuda、Tensorflow

centos7 安装cuda toolkit

AI-PaddleDetection

发表于 2024-05-08 更新于 2024-05-13 分类于 AI Waline：

PaddleDetection是一个基于PaddlePaddle的目标检测端到端开发套件。

PP-Human 实时行人分析工具

PP-Human行人分析工具使用指南

安装使用步骤

在PyCharm欢迎页通过Get from VCS拉取github上PaddlePaddle/PaddleSeg项目。
拉取成功后，PyCharm会提示创建venv虚拟环境，点击确定即可，在PyCharm打开终端命令行前面有venv代表创建成功，如果没有，关闭终端点击Add new Interpreter（需要python3.10的环境），添加后，重新在PyCharm打开终端
1
2
例如:
(venv) ➜ PaddleSeg git:(release/2.9)
安装paddle，在PyCharm中Terminal命令行中执行pip install paddlepaddle
安装setuptools，在PyCharm中Terminal命令行中执行pip install -U pip setuptools
安装依赖，在PyCharm中Terminal命令行中执行 pip install -r requirements.txt

使用步骤

# 运行图片检测人，结果输出在output目录
python deploy/pipeline/pipeline.py --config deploy/pipeline/config/infer_cfg_pphuman.yml --image_file=demo/000000014439.jpg 
# 运行视频流检测人
python deploy/pipeline/pipeline.py --config deploy/pipeline/config/examples/infer_cfg_human_attr.yml -o visual=False --rtsp rtsp://admin:123456@172.16.1.1:554/ch01.264?dev=1

检查人并输出视频

安装ffmpeg： brew install ffmpeg
下载安装rtsp推流服务bluenviron/mediamtx，在release页面选择适合的版本，我的是inter的mac，选择mediamtx_v1.8.1_darwin_amd64.tar.gz即可
解压，然后运行./mediamtx

运行视频流检测并推流python deploy/pipeline/pipeline.py --config deploy/pipeline/config/examples/infer_cfg_human_attr.yml --rtsp rtsp://admin:123456@172.1.1.1:554/ch01.264?dev=1 --pushurl rtsp://127.0.0.1:8554

python deploy/pipeline/pipeline.py --config deploy/pipeline/config/examples/infer_cfg_human_attr.yml --rtsp rtsp://admin:123456@172.16.80.59:554/ch01.264?dev=1 --pushurl rtsp://127.0.0.1:8554 
-----------  Running Arguments -----------
ATTR:
  batch_size: 8
  enable: true
  model_dir: https://bj.bcebos.com/v1/paddledet/models/pipeline/PPLCNet_x1_0_person_attribute_945_infer.zip
DET:
  batch_size: 1
  model_dir: https://bj.bcebos.com/v1/paddledet/models/pipeline/pphuman/ppyoloe_plus_crn_t_auxhead_320_60e_pphuman.tar.gz
MOT:
  batch_size: 1
  enable: true
  model_dir: https://bj.bcebos.com/v1/paddledet/models/pipeline/pphuman/ppyoloe_plus_crn_t_auxhead_320_60e_pphuman.tar.gz
  tracker_config: deploy/pipeline/config/tracker_config.yml
attr_thresh: 0.5
crop_thresh: 0.5
visual: true
warmup_frame: 50

------------------------------------------
Multi-Object Tracking enabled
Human Attribute Recognition enabled
100%|██████████| 15897/15897 [00:05<00:00, 2768.11KB/s]
DET  model dir:  /.cache/paddle/infer_weights/ppyoloe_plus_crn_t_auxhead_320_60e_pphuman
MOT  model dir:  /.cache/paddle/infer_weights/ppyoloe_plus_crn_t_auxhead_320_60e_pphuman
ATTR  model dir:  /.cache/paddle/infer_weights/PPLCNet_x1_0_person_attribute_945_infer
-----------  Model Configuration -----------
Model Arch: StrongBaseline
Transform Order: 
--transform op: Resize
--transform op: NormalizeImage
--transform op: Permute
--------------------------------------------
-----------  Model Configuration -----------
Model Arch: PPYOLOE
Transform Order: 
--transform op: Resize
--transform op: Permute
--------------------------------------------
video fps: 25, frame_count: -2562047788015215
the result will push stream to url:rtsp://127.0.0.1:8554/ch01
ffmpeg version 7.0 Copyright (c) 2000-2024 the FFmpeg developers
  built with Apple clang version 15.0.0 (clang-1500.1.0.2.5)
  configuration: --prefix=/usr/local/Cellar/ffmpeg/7.0_1 --enable-shared --enable-pthreads --enable-version3 --cc=clang --host-cflags= --host-ldflags='-Wl,-ld_classic' --enable-ffplay --enable-gnutls --enable-gpl --enable-libaom --enable-libaribb24 --enable-libbluray --enable-libdav1d --enable-libharfbuzz --enable-libjxl --enable-libmp3lame --enable-libopus --enable-librav1e --enable-librist --enable-librubberband --enable-libsnappy --enable-libsrt --enable-libssh --enable-libsvtav1 --enable-libtesseract --enable-libtheora --enable-libvidstab --enable-libvmaf --enable-libvorbis --enable-libvpx --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxml2 --enable-libxvid --enable-lzma --enable-libfontconfig --enable-libfreetype --enable-frei0r --enable-libass --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenjpeg --enable-libopenvino --enable-libspeex --enable-libsoxr --enable-libzmq --enable-libzimg --disable-libjack --disable-indev=jack --enable-videotoolbox --enable-audiotoolbox
  libavutil      59.  8.100 / 59.  8.100
  libavcodec     61.  3.100 / 61.  3.100
  libavformat    61.  1.100 / 61.  1.100
  libavdevice    61.  1.100 / 61.  1.100
  libavfilter    10.  1.100 / 10.  1.100
  libswscale      8.  1.100 /  8.  1.100
  libswresample   5.  1.100 /  5.  1.100
  libpostproc    58.  1.100 / 58.  1.100
Thread: 0; frame id: 0
Thread: 0; trackid number: 2
Input #0, rawvideo, from 'fd:':
  Duration: N/A, start: 0.000000, bitrate: 552960 kb/s
  Stream #0:0: Video: rawvideo (BGR[24] / 0x18524742), bgr24, 1280x720, 552960 kb/s, 25 tbr, 25 tbn
Stream mapping:
  Stream #0:0 -> #0:0 (rawvideo (native) -> mpeg4 (native))
Output #0, rtsp, to 'rtsp://127.0.0.1:8554/ch01':
  Metadata:
    encoder         : Lavf61.1.100
  Stream #0:0: Video: mpeg4, yuv420p(progressive), 1280x720, q=2-31, 200 kb/s, 25 fps, 90k tbn
      Metadata:
        encoder         : Lavc61.3.100 mpeg4
      Side data:
        cpb: bitrate max/min/avg: 0/0/200000 buffer size: 0 vbv_delay: N/A
Thread: 0; frame id: 10
Thread: 0; trackid number: 2

然后使用vlc播放rtsp://127.0.0.1:8554/ch01流即可，流的地址在上一步日志里面可以看到the result will push stream to url:rtsp://127.0.0.1:8554/ch01

默认的模型，识别速度个人电脑会卡顿，可以替换为小模型，模型地址可以在PP-Human快速开始-模型下载里面进行找

修改该deploy/pipeline/config/examples/infer_cfg_human_attr.yml文件
替换
  model_dir: https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_pipeline.zip
为
  model_dir: https://bj.bcebos.com/v1/paddledet/models/pipeline/pphuman/ppyoloe_plus_crn_t_auxhead_320_60e_pphuman.tar.gz

常见错误

安装依赖，执行pip install -r requirements.txt时报AttributeError: module 'pkgutil' has no attribute 'ImpImporter'. Did you mean: 'zipimporter'?错误，详细信息如下：

(.venv) ➜  PaddleDetection git:(release/2.7) pip install -r requirements.txt
Collecting numpy<1.24 (from -r requirements.txt (line 1))
  Using cached numpy-1.23.5.tar.gz (10.7 MB)
  Installing build dependencies ... done
  Getting requirements to build wheel ... error
  error: subprocess-exited-with-error
  
  × Getting requirements to build wheel did not run successfully.
  │ exit code: 1
  ╰─> [33 lines of output]
      Traceback (most recent call last):
        File "workspace/PaddleDetection/.venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 353, in <module>
          main()
        File "workspace/PaddleDetection/.venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 335, in main
          json_out['return_val'] = hook(**hook_input['kwargs'])
                                   ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        File "workspace/PaddleDetection/.venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 112, in get_requires_for_build_wheel
          backend = _build_backend()
                    ^^^^^^^^^^^^^^^^
        File "workspace/PaddleDetection/.venv/lib/python3.12/site-packages/pip/_vendor/pyproject_hooks/_in_process/_in_process.py", line 77, in _build_backend
          obj = import_module(mod_path)
                ^^^^^^^^^^^^^^^^^^^^^^^
        File "/usr/local/Cellar/python@3.12/3.12.2_1/Frameworks/Python.framework/Versions/3.12/lib/python3.12/importlib/__init__.py", line 90, in import_module
          return _bootstrap._gcd_import(name[level:], package, level)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        File "<frozen importlib._bootstrap>", line 1387, in _gcd_import
        File "<frozen importlib._bootstrap>", line 1360, in _find_and_load
        File "<frozen importlib._bootstrap>", line 1310, in _find_and_load_unlocked
        File "<frozen importlib._bootstrap>", line 488, in _call_with_frames_removed
        File "<frozen importlib._bootstrap>", line 1387, in _gcd_import
        File "<frozen importlib._bootstrap>", line 1360, in _find_and_load
        File "<frozen importlib._bootstrap>", line 1331, in _find_and_load_unlocked
        File "<frozen importlib._bootstrap>", line 935, in _load_unlocked
        File "<frozen importlib._bootstrap_external>", line 995, in exec_module
        File "<frozen importlib._bootstrap>", line 488, in _call_with_frames_removed
        File "/private/var/folders/jd/0572r10n2rbcwn4tf1w6sy500000gn/T/pip-build-env-r6d8n8j4/overlay/lib/python3.12/site-packages/setuptools/__init__.py", line 16, in <module>
          import setuptools.version
        File "/private/var/folders/jd/0572r10n2rbcwn4tf1w6sy500000gn/T/pip-build-env-r6d8n8j4/overlay/lib/python3.12/site-packages/setuptools/version.py", line 1, in <module>
          import pkg_resources
        File "/private/var/folders/jd/0572r10n2rbcwn4tf1w6sy500000gn/T/pip-build-env-r6d8n8j4/overlay/lib/python3.12/site-packages/pkg_resources/__init__.py", line 2172, in <module>
          register_finder(pkgutil.ImpImporter, find_on_path)
                          ^^^^^^^^^^^^^^^^^^^
      AttributeError: module 'pkgutil' has no attribute 'ImpImporter'. Did you mean: 'zipimporter'?
      [end of output]
  
  note: This error originates from a subprocess, and is likely not a problem with pip.
error: subprocess-exited-with-error

× Getting requirements to build wheel did not run successfully.
│ exit code: 1
╰─> See above for output.

note: This error originates from a subprocess, and is likely not a problem with pip.

原因：在docs/tutorials/INSTALL_cn.md文档里面可以查看到只支持Python 3(3.5.1+/3.6/3.7/3.8/3.9/3.10)，64位版本。

解决：执行brew install python@3.10，然后在上面安装的第二个步骤里面，切换成3.10的python环境

出现提示安装pip install numba==0.56.4的警告，安装即可

运行视频流检测人时，报错

python deploy/pipeline/pipeline.py --config deploy/pipeline/config/examples/infer_cfg_human_attr.yml -o visual=False --rtsp rtsp://admin:123456@172.16.1.1:554/ch01.264?dev=1 
Traceback (most recent call last):
  File "workspace/PaddleDetection/deploy/pipeline/pipeline.py", line 1321, in <module>
    main()
  File "workspace/PaddleDetection/deploy/pipeline/pipeline.py", line 1303, in main
    cfg = merge_cfg(FLAGS)  # use command params to update config
  File "workspace/PaddleDetection/deploy/pipeline/cfg_utils.py", line 212, in merge_cfg
    pred_config = merge_opt(pred_config, args_dict)
  File "workspace/PaddleDetection/deploy/pipeline/cfg_utils.py", line 202, in merge_opt
    for sub_k, sub_v in value.items():
AttributeError: 'bool' object has no attribute 'items'

Process finished with exit code 1

解决：删除命令里面的-o visual=False，似乎是该参数有问题

运行视频流检测加推流的时候，报错

python deploy/pipeline/pipeline.py --config deploy/pipeline/config/examples/infer_cfg_human_attr.yml --rtsp rtsp://admin:hcy123456@172.16.80.229:554/ch01.264?dev=1 --pushurl rtsp://127.0.0.1:8554 
-----------  Running Arguments -----------
ATTR:
  batch_size: 8
  enable: true
  model_dir: https://bj.bcebos.com/v1/paddledet/models/pipeline/PPLCNet_x1_0_person_attribute_945_infer.zip
DET:
  batch_size: 1
  model_dir: https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_pipeline.zip
MOT:
  batch_size: 1
  enable: true
  model_dir: https://bj.bcebos.com/v1/paddledet/models/pipeline/mot_ppyoloe_l_36e_pipeline.zip
  tracker_config: deploy/pipeline/config/tracker_config.yml
attr_thresh: 0.5
crop_thresh: 0.5
visual: true
warmup_frame: 50

------------------------------------------
Multi-Object Tracking enabled
Human Attribute Recognition enabled
DET  model dir:  .cache/paddle/infer_weights/mot_ppyoloe_l_36e_pipeline
MOT  model dir:  .cache/paddle/infer_weights/mot_ppyoloe_l_36e_pipeline
ATTR  model dir:  .cache/paddle/infer_weights/PPLCNet_x1_0_person_attribute_945_infer
-----------  Model Configuration -----------
Model Arch: StrongBaseline
Transform Order: 
--transform op: Resize
--transform op: NormalizeImage
--transform op: Permute
--------------------------------------------
-----------  Model Configuration -----------
Model Arch: YOLO
Transform Order: 
--transform op: Resize
--transform op: Permute
--------------------------------------------
video fps: 25, frame_count: -2562047788015215
the result will push stream to url:rtsp://127.0.0.1:8554/ch01
Traceback (most recent call last):
  File "workspace/PaddleDetection/deploy/pipeline/pipeline.py", line 1321, in <module>
    main()
  File "workspace/PaddleDetection/deploy/pipeline/pipeline.py", line 1308, in main
    pipeline.run_multithreads()
  File "workspace/PaddleDetection/deploy/pipeline/pipeline.py", line 179, in run_multithreads
    self.predictor.run(self.input)
  File "workspace/PaddleDetection/deploy/pipeline/pipeline.py", line 533, in run
    self.predict_video(input, thread_idx=thread_idx)
  File "workspace/PaddleDetection/deploy/pipeline/pipeline.py", line 670, in predict_video
    pushstream.initcmd(fps, width, height)
  File "workspace/PaddleDetection/deploy/pipeline/pipe_utils.py", line 142, in initcmd
    self.pipe = sp.Popen(self.command, stdin=sp.PIPE)
  File "/usr/local/Cellar/python@3.10/3.10.14/Frameworks/Python.framework/Versions/3.10/lib/python3.10/subprocess.py", line 971, in __init__
    self._execute_child(args, executable, preexec_fn, close_fds,
  File "/usr/local/Cellar/python@3.10/3.10.14/Frameworks/Python.framework/Versions/3.10/lib/python3.10/subprocess.py", line 1863, in _execute_child
    raise child_exception_type(errno_num, err_msg, err_filename)
FileNotFoundError: [Errno 2] No such file or directory: 'ffmpeg'

Process finished with exit code 1

解决：安装ffmpeg，执行brew install ffmpeg