Hello_User/gpu-jupyter
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
d7519eee7b
gpu-jupyter/Deployment-notes.md
Christoph Schranz d7519eee7b adding swarm tools and deployment notes
2020-03-11 08:26:54 +01:00

1.1 KiB
Raw Blame History

Deployment Notes

Push image with tag to Dockerhub

Based on this tutorial with the tag v1.0_cuda-10.1_ubuntu-18.04:

# on il048:
cd ~/Documents/projects/GPU-Jupyter/gpu-jupyter
git pull
bash generate_Dockerfile.sh
bash start-local -p 1234
docker image ls
docker tag [IMAGE ID] cschranz/gpu-jupyter:v1.0_cuda-10.1_ubuntu-18.04
docker push cschranz/gpu-jupyter:v1.0_cuda-10.1_ubuntu-18.04
docker save cschranz/gpu-jupyter > ../gpu-jupyter_tag-v1.0_cuda-10.1_ubuntu-18.04.tar

Then, the new tag is available on Dockerhub.

Deployment in the swarm

The GPU-Jupyter instance for deployment, that has swarm files and changed pw is in /home/iotdev/Documents/projects/dtz/src/gpu-jupyter

# on il048:
cd /home/iotdev/Documents/projects/dtz/src/gpu-jupyter
git pull
bash generate_Dockerfile.sh
bash add-to-swarm-with-defaults.sh

Then, the service will be available with data stored in data on 192.168.48.48:8848 with our password.