Skip to content

Preparation

1. How to login to MeluXina machine

2. Use your username to connect to MeluXina

For exmaple the below example shows the user of u100490 

$ ssh u100490@login.lxp.lu -p 8822

3. Once you have logged in

  • 3.1 Once you have logged in, you will be in a default home directory 
    [u100490@login02 ~]$ pwd
/home/users/u100490
  • 3.2 After that go to project directory (Nvidia Bootcamp activites). 
    [u100490@login02 ~]$ cd /project/home/p200117
[u100490@login02 p200117]$ pwd
/project/home/p200117

4. And please create your own working folder under the project directory, for example, here it is user with u100490:

[u100490@login02 p200117]$ mkdir $USER
### or 
[u100490@login02 p200117]$ mkdir u100490

  • 4.1 Now copy climate.simg and climate.sh from project direcoty to your user directory (for exxample, here is u100490) directory: 

    [u100490@login02 p200117]$ cp /project/home/p200117/climate.simg /project/home/p200117/u100490
[u100490@login02 p200117]$ cp /project/home/p200117/climate.sh /project/home/p200117/u100490

  • 4.2 Similary, copy cfd.simg and cfd.sh from project direcoty to your user directory (for example, here is u100490) directory: 

    [u100490@login02 p200117]$ cp /project/home/p200117/cfd.simg /project/home/p200117/u100490
[u100490@login02 p200117]$ cp /project/home/p200117/cfd.sh /project/home/p200117/u100490

  • 4.3 Now go to your home (for example, here it is u100490) directory check if all the necessary files are there (.simg and .sh) 
    [u100490@login02 p200117]$ cd u100490
[u100490@login02 u100490]$ pwd
[u100490@login02 u100490]$ /project/home/p200117/u100490
[u100490@login02 u100490]$ ls -lthr
total 15G
-rw-r-x---. 1 u100490 p200117  736 Feb  8 18:59 climate.sh
-rwxr-x---. 1 u100490 p200117 7.2G Feb  8 19:19 climate.simg
-rwxr-x---. 1 u100490 p200117 6.9G Feb  8 19:21 cfd.simg
-rw-r-x---. 1 u100490 p200117  723 Feb  8 19:21 cfd.sh

5. For the dry run (9th February from 11:30-12:30), please follow the following steps:

[u100490@login02 u100490]$ salloc -A p200117 --res gpudev -q dev -N 1 -t 01:00:0
[u100490@mel2123 u100490]$ mkdir -p $PROJECT/$USER/workspace-climate
[u100490@mel2123 u100490]$ module load Singularity-CE/3.10.2-GCCcore-11.3.0

[u100490@mel2123 u100490]$ singularity run --bind $PROJECT/$USER $PROJECT/$USER/climate.simg cp -rT /workspace $PROJECT/$USER/workspace-climate
INFO:    Converting SIF file to temporary sandbox...
INFO:    Cleaning up image...
[u100490@mel2123 u100490]$ singularity run --nv --bind $PROJECT/$USER $PROJECT/$USER/climate.simg jupyter lab --notebook-dir=$PROJECT/$USER/workspace-climate/python/jupyter_notebook --port=8888 --ip=0.0.0.0 --no-browser --NotebookApp.token=""
INFO:    Converting SIF file to temporary sandbox...
WARNING: underlay of /usr/bin/nvidia-smi required more than 50 (452) bind mounts
[W 10:10:32.723 LabApp] All authentication is disabled.  Anyone who can connect to this server will be able to run code.
[I 10:10:33.043 LabApp] jupyter_tensorboard extension loaded.
[I 10:10:33.047 LabApp] JupyterLab extension loaded from /usr/local/lib/python3.8/dist-packages/jupyterlab
[I 10:10:33.047 LabApp] JupyterLab application directory is /usr/local/share/jupyter/lab
[I 10:10:33.048 LabApp] [Jupytext Server Extension] NotebookApp.contents_manager_class is (a subclass of) jupytext.TextFileContentsManager already - OK
[I 10:10:33.048 LabApp] Serving notebooks from local directory: /mnt/tier2/project/p200117/u100490/workspace-climate/python/jupyter_notebook
[I 10:10:33.048 LabApp] Jupyter Notebook 6.2.0 is running at:
[I 10:10:33.048 LabApp] http://hostname:8888/
[I 10:10:33.049 LabApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).

  • 5.1 Now open a new terminal on your local computer, and again login to MeluXina to access the port

  • 5.2 Make sure you use the name NODELIST (for example, here it is mel2123 - from squeue command you will get this number) 

    ssh -L8080:NODELIST:8888 USERNAME@login.lxp.lu -p 8822
ssh -L8080:mel2123:8888 u100490@login.lxp.lu -p 8822

  • 5.3 Keep those terminals open/alive (please do not close them)

  • 5.4 Now copy and paste localhost to your browser either to Chrome or FireFox 

    http://localhost:8080

6. For the afternoon session (9th and 10th February)

If have missed the dry run session, please go through the steps from 1-4

  • 6.1 Now it is time to edit your batch script (climate.sh) before launching your Jupyter notebook, please follow the following steps: 
    [u100490@login02 u100490]$ emacs(emacs -nw)/vim climate.sh
#!/bin/bash -l
#SBATCH --partition=gpu 
#SBATCH --ntasks=1
#SBATCH --nodes=1    
############  day one ##########
#######SBATCH --time=02:00:00         ## use this option for day one
#######SBATCH --res ai_bootcamp_day1   ## use this option for day one
################################

############  day two ##########
#SBATCH --time=03:30:00         ## use this option for day two
#SBATCH --res ai_bootcamp_day2  ## use this option for day two
################################
#SBATCH -A p200117
#SBATCH --qos default

mkdir -p $PROJECT/$USER/workspace-climate
module load Singularity-CE/3.10.2-GCCcore-11.3.0

singularity run --bind $PROJECT/$USER $PROJECT/$USER/climate.simg cp -rT /workspace $PROJECT/$USER/workspace-climate
singularity run --nv --bind $PROJECT/$USER $PROJECT/$USER/climate.simg jupyter lab --notebook-dir=$PROJECT/$USER/workspace-climate/python/jupyter_notebook --port=8888 --ip=0.0.0.0 --no-browser --NotebookApp.token=""
  • 6.2 Once you have modified your climate.sh, please launch your batch script as below: 
    [u100490@login03 u100490]$ sbatch climate.sh
Submitted batch job 276009
[u100490@login03 u100490]$ squeue 
JOBID PARTITION     NAME     USER    ACCOUNT    STATE       TIME   TIME_LIMIT  NODES NODELIST(REASON)
276009       gpu climate.  u100490    p200117  RUNNING       0:16        20:00      1 mel2077
  • 6.3 Now you have initiated your singularity container which will help you to open the Jupyter nootebook 
    [u100490@login03 u100490]$ ls -lthr
total 7.2G
-rwxr-x---. 1 u100490 p200117 7.2G Feb  3 14:53 climate.simg
-rw-r-----. 1 u100490 p200117  613 Feb  3 17:06 climate.sh
-rw-r-x---. 1 u100490 p200117  724 Feb  8 19:41 cfd.sh
-rwxr-x---. 1 u100490 p200117 6.9G Feb  8 19:42 cfd.simg
-rw-r--r--. 1 u100490 p200117 1.1K Feb  3 17:58 slurm-276009.out
  • 6.4 You can also check meantime if everything OK by executing the below command and you should get similar output: 
    [u100490@login03 u100490]$ head -30 slurm-276009.out 
INFO:    Converting SIF file to temporary sandbox...
INFO:    Cleaning up image...
INFO:    Converting SIF file to temporary sandbox...
WARNING: underlay of /usr/bin/nvidia-smi required more than 50 (452) bind mounts
[W 17:58:37.489 LabApp] All authentication is disabled.  Anyone who can connect to this server will be able to run code.
[I 17:58:37.807 LabApp] jupyter_tensorboard extension loaded.
[I 17:58:37.811 LabApp] JupyterLab extension loaded from /usr/local/lib/python3.8/dist-packages/jupyterlab
[I 17:58:37.811 LabApp] JupyterLab application directory is /usr/local/share/jupyter/lab
[I 17:58:37.813 LabApp] [Jupytext Server Extension] NotebookApp.contents_manager_class is (a subclass of) jupytext.TextFileContentsManager already - OK
[I 17:58:37.813 LabApp] Serving notebooks from local directory: /mnt/tier2/project/p200117/u100490/workspace-climate/python/jupyter_notebook
[I 17:58:37.813 LabApp] Jupyter Notebook 6.2.0 is running at:
[I 17:58:37.813 LabApp] http://hostname:8888/
[I 17:58:37.813 LabApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation).
  • 6.5 Now open a new terminal on your local computer, and again login to MeluXina to access the port
  • 6.6 Make sure you use the name NODELIST (here it is mel2077 - from squeue command you will get this number) 
    ssh -L8080:NODELIST:8888 USERNAME@login.lxp.lu -p 8822
ssh -L8080:mel2077:8888 u100490@login.lxp.lu -p 8822
  • 6.7 Keep those terminals open/alive (please do not close them)
  • 6.8 Now copy and paste localhost to your browser either to Chrome or FireFox 
    http://localhost:8080