You can use File Storage NAS (NAS) datasets to store training, deployment, and result files for services of Platform for AI (PAI), including Deep Learning Containers (DLC), Data Science Workshop (DSW), and Elastic Algorithm Service (EAS). This topic uses a General-purpose NAS file system to show how to mount NAS dataset for PAI services.
Prerequisites
A General-purpose NAS file system is created. For more information, see Create a file system.
Step 1: Create a NAS dataset
This step is required if you want to mount NAS dataset in DLC and DSW, but is not necessary for EAS.
Create a NAS dataset and configure the following key parameters. For information about how to create the dataset and the other parameters, see Create and manage datasets.
The region of the dataset must be the same with the region of your DLC instance, DSW instance, or ESW service, and with the region where your data and code are stored.
Parameter | Description |
Select Data Storage | Choose General-purpose NAS. |
Select File System | Enter the ID of your NAS file system or select the ID from the frop-down list. You can view the NAS file system ID in the NAS console. |
File System Path | The path of the NAS file system. Set the value to |
Default Mount Path | The mount path of your DLC job or your DSW instance in the NAS file system. Set the value to |
Step 2: Use the NAS dataset
After you create a NAS dataset, you can read data from the NAS file system while using DLC, DSW, or EAS through the dataset.
Mount NAS dataset in DLC
Mount the NAS dataset when creating a DLC job. Configure the following key parameters. For information about other parameters, see Submit training jobs.
In Mount Settings under the Environment Information section, click Add. Select Custom Dataset and configure the following two parameters:
Datasets: Click
and select the created General-purpose NAS dataset.Mount Path: The system automatically sets the value to the default mount path configured for the dataset. During DLC jobs, the system can access data in the NAS file ststem based on this path.
Mount NAS dataset in DSW
Mount the NAS dataset when creating a DSW instance. Configure the following key parameters. For information about other parameters, see Create a DSW instance.
Dataset: Mount the dataset by configuring the following two parameters.
Dataset: Click
, and select the created General-purpose NAS dataset.Mount Path: The system automatically sets the value to the default mount path configured for the dataset. In the DSW development environment, you can access data in the NAS file ststem based on this path.
Working Directory: The default system disk directory is
/mnt/workspace. You can also select the mounted dataset from the drop-down list. Then, the system mount the dataset to the DSW working directory/mnt/workspace.
Mount NAS dataset in EAS
Mount the NAS dataset when deploying an EAS service. Configure the following key parameters. For information about other parameters, see Deploy a model service in the PAI console.
In the Model Service Information section, select Model Settings. Then, select Mount NAS File System and configure the following parameters:
NAS Mount Target: Select your NAS file system and a mount point for the EAS service to access the NAS file system.
NAS Source Path: The source path in the NAS instance to be mounted.
Mount Path: The destination path in the EAS service to read files from the NAS path.