Docs

Volcano Engine Function as a Service

Data processing

Using veFaaS to read and write the TOS mount directory

Mounting a TOS Bucket on a function enables persistent data storage and sharing across multiple functions. This article introduces how to use the veFaaS to read and write the mounted TOS Bucket.

Scenario introduction

Mounting a TOS Bucket on a function allows persistent data storage and sharing across multiple functions. In addition, functions can access TOS like a local file system, simplifying resource access and data processing.

Prerequisites

Torch Object Storage (TOS)

Volcano Engine Torch Object Storage (TOS) is activated.

A TOS Bucket is created in the region where functions will be deployed. For detailed instructions, see Creating buckets.

API access keys are obtained, and the access keys are required to have permission to access TOS. Refer to Access Key management and Creating user and authorize for specific operations.

warning

To better manage permissions, it is recommended to use IAM user keys with minimal authorization. Avoid using Volcano Engine root user keys or IAM user keys with excessive permissions directly.

veFaaS

The veFaaS is activated.

Procedure

Step One: Create a function and mount TOS

In the top navigation bar, select the target region.

On the function list page, click Create function and select Create [Event function].

Basic information

Enter a function name and notes.

Function code

Runtime: Select Python.

Code template: Select vefaas-python38-default.

Advanced settings

Enable TOS mount feature for storage configurations, and keep other parameters as default.

Parameter	Description
AKSK	Enter Access Key ID and Secret Access Key to authorize the veFaaS to use TOS. The access key must have reading and writing permissions for TOS.
TOS mount	Add a TOS mount point. This example adds one TOS mount point. Bucket: Select the TOS Bucket to mount. Access address: Automatically obtained after selecting the Bucket. Bucket subdirectory: Select the Bucket subdirectory the function needs to access. Must start with /. This example is /docdemo. Local directory: Configure the local directory in the function operation environment. This example is /mnt/tos. Permission: This example selects reading and writing permissions, allowing the function to read and write the local directory mounted on the Bucket.

Click Create to create the function.

Step Two: Edit code online

This article uses the following code example to demonstrate to you how to read and write the mounted TOS Bucket.

tip

The code logic is for reference only. Modify the code according to your actual business scenario as needed.

On the Code tab of the current function, click index.py file, overwrite the original content with the following code, and click Save.

Step Three: Release the function

Release the current code and configuration of the function online.

On the function details page, click Release in the upper right corner.

Select Latest for the function version. Keep other parameters as default.

Click OK. Wait until the function release is finished.

Validating results

Test the function to verify whether the response result meets expectations.

On the function details page, click Test in the upper-right corner.

On the pop-up function test page, configure the Trigger type and test request.

Trigger type: Select API Gateway Trigger.

Test request: Keep the default.

Click Test. Wait for the test results.

View the instance logs of the function to confirm whether the expected file reading and writing is completed.

Switch to the Resource management tab.

In the Instances list, select the running instance, click Logs, and enter the instance logs page.

When a function has multiple instances, it is recommended to check the logs of each instance to identify the one instance actually handling the request.

Last updated: 2026.08.19 14:06:07