Website: https://blog.dgut.top/2020/08/25/docker/
Docker learning record
Recently, the visual indoor positioning based on aurco has also been completed. The camera can obtain the relative distance and angle by detecting the aruco code, and then bring it into the map of the entire robot to play a very good auxiliary positioning function. However, it is certain that the style of the aruco label cannot be changed. Just a piece of A4 paper pasted on the wall will greatly affect the appearance, so the next step should be to detect 物体检测obvious features such as a painting, a wall, and a staircase. objects to aid positioning.
To realize it, 物体检测I think it is not enough to rely on the cascade classifier of opencv, so the project must move closer to speed learning. The project team happened to have a GPU server in use. I roughly checked the configuration: the cpu is two-way Xeon E5-2640 v4 @ 2.40GHz, eight-way gtx1080ti, and 250G memory.
At present, there is a problem that the server is currently running some training programs of other members of the project team, and the programming environment we use may be different, which may cause cudaenvironment tensorflowerrors. Therefore we want to use dockeras a whole the training environment. This article is to record some dockerusage methods. As a beginner, it is necessary to record it.
Install
Introduction : Docker is an open source application container engine that allows developers to package their applications and dependencies into a portable image, and then publish it to any popular Linux or Windows machine, and can also implement virtualization . Containers use a sandbox mechanism completely, and there will be no interfaces between them.
Docker installation :
curl -fsSL https://get.docker.com | bash -s docker --mirror Aliyun
Download address: Docker Desktop for Windows - Docker Hub
Choose the stable stable version to download, the fool-like installation process, one-click to the end.
After the computer restarts, open Docker, and click the Dashboard of Docker on the task bar in the lower right corner.
After entering docker in the terminal and seeing the following information, it proves that the installation is successful:
use
Recently, a circle of AP networks was set up on the 3rd floor of the company for debugging inspection robots, but we have not been very clear about the capacity of this network. I saw this project on github today, and this project supports docker deployment, librespeed-speedtestso it is Use it to practice, and test the throughput capacity of wifi LAN by the way. docker address
- Schematic diagram of robot network
- speedtest docker project
deploy
Docker deployment is also super simple, just 2 minutes
1. Mirror download:
docker pull adolfintel/speedtest
If the network is not good, it is recommended to repeat the operation. If the text similar to the following is displayed, the download is complete:
2. Start docker
docker run -d -p 8080:80 adolfintel/speedtest:latest
- -d, run in the background (recommended)
- -p, port mapping (you can modify other ports by yourself)
At this point, you can see the containers running in the background in the previous Dashboard. This machine can open the web page http://localhost:8080, which also indicates that the service has started successfully.
3. Speed measurement
The speed measurement process is very foolish~
Other devices in the LAN open the webpage http://[PC IP]:8080 to measure the speed. If you can't access, it may be a firewall, router settings or other problems.
http://192.168.2.182:8080/
4. Set DMZ host
Allows you to expose a computer publicly on the Internet, so that all uploaded packets are redirected to the computer you specify. This is very useful when you run some applications that use non-specific incoming communication ports (incoming port). Please use with caution.
In layman's terms, the computer is connected to the company's LAN through a router. The IP of the router is different from the IP of the computer's LAN (hundred-layer nat dog head). After setting the computer to DMZ主机access the router's IP, you can access the speed test webpage.
docker command
登录服务器http://192.168.221.11/
ssh zwj@AISRV
查看docker运行列表
docker ps
运行docker
docker run -d -p 8081:80 adolfintel/speedtest:latest
停止docker
docker stop compassionate_knuth
进入docker 启动docker
cd work
cd comprehen/
./docker_start.sh
./docker_into.sh
dockerd builds pagoda management panel
Pagoda Linux panel is a server management software that improves operation and maintenance efficiency, and supports more than 100 server management functions such as one-key LAMP/LNMP/cluster/monitoring/website/FTP/database/JAVA.
A professional team of 30 people develops and maintains it. After more than 200 iterations of versions, it has full functions, less errors and is safe enough. It has been approved and installed by millions of users around the world. Operation and maintenance must be efficient and pagodas installed.
installation steps
First dockerpull a centosmirror image from
docker pull centos
run mirror
docker run -i -t -d --name baota-zwj -p 20:20 -p 21:21 -p 80:80 -p 443:443 -p 888:888 -p 8888:8888 -p 8084:8084 -p 8085:8085 --privileged=true -v /data1/zwj/baota:/www centos
-p external port number: internal port number, ports 20, 21, 80, 443, 888, 8888, 8084, 8085 are open here
-v local path: the internal path wants to mount a hard disk to docker, and this hard disk can be mapped to docker locally
Enter Pagoda docker
docker exec -it baota-zwj /bin/bash
Pagoda installation:
yum install -y wget && wget -O install.sh http://download.bt.cn/install/install_6.0.sh && sh install.sh
After the installation is successful, the login address and account password of the pagoda will be displayed. Generally: http://localhost:8888
start up
/etc/init.d/bt start
Log in
access panel
resource monitoring