Google Datalab with Source Control

Go to GCP website

Open Active Cloud shell

Create datalab env

datalab create <name> --zone <zone>
sample:datalab create mydatalab --zone asia-east1-a

What service create in this time

Computing Service

Storage Service

Repository Service

Open DataLab

Develop on DataLab

Commit to GCP Repository

Open new could shell

SSH to Cloud Service (VM)

gcloud compute --project <project-id> ssh <user>@<instance> --zone zone <zoon>
sample: gcloud compute --project qwiklabs-gcp-3b87b8785bf497ac ssh google5031275_student@mydatalab --zone asia-east1-a
sudo docker ps 

# datalab是架在docker上
docker exec -it 5685ff4ca32e bash

Check in GCP Repository again

Connect with Github

Open new could shell

SSH to Cloud Service (VM)

gcloud compute --project <project-id> ssh <user>@<instance> --zone zone <zoon>
sample: gcloud compute --project qwiklabs-gcp-07c3064872c82b79 ssh google5032845_student@mydatalab --zone asia-east1-a

查看這台VM上面的Docker container ，可以發現一個用來Logging的Fluentd container，另一個是DataLab的container

sudo docker ps

連進DataLab的container

docker exec -it 96f364687a2d bash

由於避免搞壞原本的環境，所以這邊我們COPY一份Folder出來

cp /content/
cp -r datalab/ datalab2/

將原本的 .git folder 刪除

cd datalab2/notebooks
rm -rf .git

產生ssh key

ssh-keygen

將ssh public key 設定在 github 中

產生新的git設定

git init
git add .
git commit -m "<comment>"
git remote add origin <git_link>
git push -u origin master

到Github查看是否有成功

Reference

1. https://www.coursera.org/learn/google-machine-learning/gradedLti/8B9M9/lab-analyzing-data-using-datalab-and-bigquery
2. https://cloud.google.com/datalab/docs/how-to/working-with-notebooks