This script will help you to install Jupyter Server on your server.
ssh <name>@lanta.nstda.or.th
# for high-bandwidth and downloading files
ssh <name>@transfer.lanta.nstda.or.thNote: SSH into the server as the user created in the previous step first.
mkdir scripts
cd scripts
wget https://github.com/Kentakoong/init-cloudtraining/releases/download/LANTA-1.0.2/setup.sh
bash setup.shThen follow the interactive prompt to setup the Jupyter Server.
After the script finishes, check your queue to see if the script has been submitted.
myqueueIf the script has been submitted, check your current directory with ls to see the slurm file.
ls
cat slurm-xxxxxx.outThe slurm file will contain the URL to access the Jupyter Server.
To access the Jupyter Server, copy the URL from the slurm file and paste it into your browser.
ssh -L <PORT>:<HOST>:<PORT> <username>@lanta.nstda.or.th -i <path-to-private-key>Then open your browser and go to localhost:<PORT>
or copy the URL with the token from the slurm file and paste it into your browser.
To stop the Jupyter Server, go to the terminal where you ran the batch script.
Run the following command to see the job ID.
myqueueThen run the following command to stop the Jupyter Server.
scancel <job-id>ssh <username>@transfer.lanta.nstda.or.thThen run the following command to SSH into the compute node.
bash setup.sh
myqueue
ssh lanta-c-xxxNOTE: Your disk quota is limited to 100GB, if you need more space, use your house projects storage.
conda create --name <env-name> --clone <source-env>conda activate <env-name>
# To install a package
conda install <package-name>
# To install a package from a requirements file
pip install --file requirements.txtconda env remove -n <env-name>sbatch - Submit a batch script to the queue.
scancel - Cancel a job.
sbalance - Check the balance of the queue (GPU hours).
sinfo - Check nodes that are both idle and running.
myqueue - Check the queue.
myquota - Check your quota (Disk usage, projects).