vishnuexe/docker_manual.md

## docker_manual.md

      
    Raw
  

              docker_manual.md
            
          
    Docker Manual

Step 1 : Install Docker

#Docker Installation 

$ sudo apt-get update

$ curl https://get.docker.com | sh \
  && sudo systemctl --now enable docker

$ distribution=$(. /etc/os-release;echo $ID$VERSION_ID) \
  	&& 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/$distribution/libnvidia-container.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
Step 2 - Install Nvidia - Docker

#Nvidia-Docker Installation

$ sudo apt-get update

$ sudo apt-get install -y nvidia-docker2

$ sudo systemctl restart docker
Step 3 - Verify Install

#Running sample docker

$ sudo docker run --rm --gpus all nvidia/cuda:11.6.2-base-ubuntu20.04 nvidia-smi
If you see this in your terminal after you have executed the command above, you have successfully installed nvidia-docker
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.60.11    Driver Version: 525.60.11    CUDA Version: 12.0     |
|-------------------------------+----------------------+----------------------+
| 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 ...  Off  | 00000000:01:00.0  On |                  N/A |
| 30%   34C    P8    25W / 320W |   1071MiB / 10240MiB |      1%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
| No running processes found                                                  |
+-----------------------------------------------------------------------------+


Step 4 - Using a docker

Find the required docker tag - preferably from docker hub
#pulling docker image
$ sudo docker pull [docker_image]
#starting the container with gpu access in interactive mode
$ sudo docker run --gpus all -it [docker_image] bash

Example for pytorch docker
$ sudo docker pull pytorch/pytorch
$ sudo docker run --gpus all -it pytorch/pytorch bash

If you see this in your terminal after executing the above command, then you are inside the container and it works same as your normal terminal/command prompt with all the standard commands!!
root@<container-id>:/workspace# 
Now you can copy your codes or download codes from github and execute it as you would in a normal terminal/command prompt.
Example:
root@<container-id>:/workspace#git clone <git repo>
root@<container-id>:/workspace#python script.py
To exit from docker just type exit and hit enter.
root@<container-id>:/workspace#exit
user@yoursystem:~$ 
Basic Commands (to be used when outside docker)

To List all dockers images in the local machine
docker images    
Check status of containers
docker ps -a
To retain all files and packages in the docker even after exiting
docker commit <container_id> <new-name>
Delete container
docker rm <container_id>
Delete all container/images/volumes at once
docker rm -f $(docker ps -a -q)             #delete all containers
docker volume rm $(docker volume ls -q)     #delete all volumes
docker rmi $(docker images -a -q)           #delete all images
Copy from Local disk to docker
docker cp <location in local system> container-id:<location in container>
Example for moving foo.txt from local system to docker container
sudo docker cp ~/foo.txt thw8hb78bu:/workspace/
Copy from docker to Local disk
sudo docker cp container-id:<location in container> <location in local system> 
List images
docker images
To initiate Jupyter Notebook from inside a docker
Step 1: Run this in your host system
docker run --gpus all -it -p 8888:8888 image:version
Step 2: Run this inside docker
jupyter notebook --ip 0.0.0.0 --port 8888 --no-browser --allow-root
for jupyter notebook and
jupyter lab --ip 0.0.0.0 --port 8888 --no-browser --allow-root
for jupyter lab
Step 3: Access the notebook through your desktop browser on http://localhost:8888
The notebook will prompt you for a token which was generated when you create the notebook which you will find from step 2.
No results found