Notebook

The Notebooks menu is used to create, data analysis, software development, or machine learning experiments using JupyterLab by running them in Pods on the cluster Deka Your GPU.

Create Notebooks

Create Notebooks is used to create new notebooks on the Deka GPU Portal Service. On the Notebooks page press the button “+ New Notebooks”.

Choose Notebook: JupyterLab

The New Notebook page appears and several columns must be filled in. In the Name column, fill in the name of the notebook that will be used. Need to know: The name must consist of lowercase alphanumeric characters or '-', start with an alphabetic character, and end with an alphanumeric character. In this guide we use JupyterLab. Adjust the name of the image created in the Registry via the Deka GPU Service Portal in the Custom Notebook section.

Add CPU/RAM and GPUs

In the CPU/RAM section in Minimum CPU and Minimum Memory GI, adjust to your needs. In the GPUs section, specify the GPU to be used and the GPU Vendor will automatically be filled with NVIDIA.

Add Workspace Volume

Workspace Volumes are used for storing data used by various components or pipelines and help in organizing and managing data during development. In Workspace Volume there are two features that you can use, namely "Add new volume” and “Attach existing volume”.

The first feature is the button “+Add new Volume” used to add new storage volumes that can be used by various components in a pipeline or notebook server.

To see the details of the new volume that will be added by pressing the icon next to the Delete icon, the new volume displays by default as follows:

a. In the type used empty volume,

b. The name used for the new volume is "-volume", you can customize the name to be used.

c. The size used is 5 Gibibyte,

d. The Storage class by default uses the default storage class, namely storage-nvme-c1, if you don't check Use Default Class there are several options including Empty Storage Class, storage-mgmt, storage-nlsas, storage-nvme-c2, and other built-in classes.

e. In access mode there are three types namely:

• ReadWriteOnce (RWO) has exclusive access, namely read or write to the notebook server.

• ReadOnlyMany (ROX) has read-only access on many server notebooks.

• ReadWriteMany (RWX) have concurrent read or write access to multiple notebook servers.

f. The mount path is the location in the container where the volume will be accessed. In this Workspace Volume, the default directory used by Jupyter Notebook is /home/jovyan.

The second feature is the button “+Attach existing volume” used to connect existing storage volumes to the notebook server or machine learning pipeline in use.

To view the details of the volume attachment you want to add, press the icon next to the Delete icon, displayed by default as follows:

a. In the type used by Kubernetes volume,

b. You can check “Readonly” to ensure the access mode on volumes connected to a notebook server or pipeline is read-only without permission to write or modify data on them.

c. The Name is used to select data for the available workspace volume. To add new volume data, follow the Create Volumes sub-chapter in this guide.

d. The mount path is the location in the container where the volume will be accessed. In this Workspace Volume, the default directory used by Jupyter Notebook is /home/jovyan.

Add Data Volume

The Data Volumes section is used to store and manage data persistently beyond the life cycle of the notebook server used, so that the data remains even if the server is stopped or deleted. In Workspace Volume there are two features that you can use, namely "Add new volume” and “Attach existing volume”. You can add more than one data volume.

The first feature is the button “+Add new Volume” used to add new storage volumes that can be used by various components in a pipeline or notebook server.

To see the details of the new volume data that will be added by pressing the icon next to the Delete icon, the new volume data displays by default as follows:

a. In the type used empty volume,

b. The name used for the new volume is “-datavol-1”, You can customize the name that will be used.

c. The size used is 5 Gibibyte,

d. The Storage class by default uses the default storage class, namely storage-nvme-c1, if you don't check Use Default Class there are several options including Empty Storage Class, storage-mgmt, storage-nlsas, storage-nvme-c2 and other built-in classes.

e. There are three types of access modes, namely:

• ReadWriteOnce (RWO) has exclusive access, namely read or write to the notebook server.

• ReadOnlyMany (ROX) has read-only access on many server notebooks.

• ReadWriteMany (RWX) have concurrent read or write access to multiple notebook servers.

f. The mount path is the location in the container where the volume will be accessed. In this Workspace Volume, the default directory used by Jupyter Notebook is /home/jovyan/-datavol-1.

Choose GPU Type

To be able to select the type of GPU used on the server notebook used, press Advanced Options.

Select the GPU used in Affinity Config. Currently there are only 5 GPUs available that can be used including:

• NVIDIA HGX-H100 SXM 80GB

• NVIDIA HGX-H100 SXM 10GB (MIG)

• NVIDIA HGX-H100 SXM 40GB (MIG)

• NVIDIA HGX-H100 SXM 20GB (MIG)

• NVIDIA EGX-L40S 48GB

Click the Launch button to create a notebook.

Open JupyterLab Notebook

After creating and running a new notebook press the Connect button to open the notebook. Jupyter Notebook automatically opens.

Stop JupyterLab Notebook

Stop Notebook Server is used to temporarily stop or pause the running notebook server. It offers several important benefits, including resource conservation, preserving the current state, and better management.

Delete Notebook

Delete Notebooks is used to delete notebooks that are no longer used. On the Notebooks page, determine the notebook you want to delete by pressing the trash icon.

Click the Delete button to continue the process of deleting Notebooks.