dwijaybane / dl-server

dl-server

steps to connect server from local machine

0. Note the dl-server ip address

1. Ensure Public Keys are available

• cd into ~/.ssh directory
• if you don't have an .ssh directory in your home folder, create it (mkdir ~/.ssh)
• if you don't have an id_rsa.pub file in your ~/.ssh folder, create it (ssh-keygen and hit Enter 3 times)

2. Add dl-server info to config file

• if you don't have a config file, create one. This example creates file using gedit editor. $ gedit config
• add these contents to your config file (replace IP address here with your dl-server IP address)

Host dl-server
     HostName 10.7.1.150
     IdentityFile ~/.ssh/id_rsa
     # StrictHostKeyChecking no  
     User <user-name>
     # LocalForward 5000 localhost:5000
     # LocalForward 8888 localhost:8888

3. ssh into dl-server from local computer

$ ssh dl-server   

Connect to dl-server

User Creation steps

$ ssh bostonai@10.7.1.150
$ sudo adduser user<#number>
$ sudo usermod -aG docker user<#number>

1. General setup

$ ssh dl-server
$ mkdir ~/docker_mount
$ cd ~/docker_mount
$ docker volume create docker_store
$ docker run hello-world
$ docker load --input /LAB_Setup/dl_lab_v3.tar

2. Once we have docker running lets setup our data for examples

$ echo "LAB=~/docker_mount" >> ~/.bashrc
$ source ~/.bashrc 

3. Make sure docker_store exist (If not create if ref:docker-setup.sh) $ docker volume inspect docker_store

4. Copy example you want to run to your docker_mount folder $ cp -R /LAB_Setup/dl-lab-files/<example> ~/docker_mount

5. Now we can download and run our dl-lab image in container

$ nvidia-docker run --rm --init -it --name container1 -p 50##:5000 -p 88##:8888 -v $LAB:/LAB -w="/LAB" -v docker_store:/docker_store streaminterrupt/dl-lab:v3

6. To Run NVIDIA Digits [In Container] $ digits-devserver #access on local browser localhost:50##

Now open new tab of terminal

$ ssh dl-server
$ docker exec -it container1 bash
$ jupyter-notebook --ip=0.0.0.0 --port=8888 --allow-root  

access on local browser by ctrl + clicking link it generates

Note Everytime you make change in docker make sure to save image

Do it in new tab while your container is still active and required changes are done $ docker commit container1 streaminterrupt/dl-lab:v3