Is this page helpful?

Log and Runtime Files#

NVIDIA AI Workbench creates log and runtime files that you can use to debug issues, and to understand the state of your applications. There are log files for the AI Workbench Desktop Application, Service, CLI, and for applications that you use with AI Workbench. Some log file content appears in the desktop application, so that you can monitor your projects and apps while they are running.

Use this documentation to perform the following tasks:

Access Log Files From the Desktop Application#

To access server and app log files from the AI Workbench Desktop Application (Desktop App), do the following:

  1. In the status bar, click Output to open the output pane.

  2. In the drop-down, select Service or the name of an app, such as JupyterLab to view the log file in the output pane.

Manually Access Log Files#

For local locations, all log files are on the local computer. For remote locations, the server and app log files are on the remote computer, and the desktop application and CLI log files are on your local machine. SSH into the remote computer to view the server and app log files.

To access log files manually, find them in the following locations.

Important

When you install AI Workbench on Windows, a folder named Linux\NVIDIA-Workbench (\wsl.localhost\NVIDIA-Workbench) is added to your File Explorer.

Log File Type

Location

Desktop application log files

C:\Users\<USERNAME>\AppData\Local\NVIDIA Corporation\AI Workbench\logs\main.log

Server log files

Linux\NVIDIA-Workbench\home\workbench\.nvwb\logs\workbench.log

CLI log files

Linux\NVIDIA-Workbench\home\workbench\.nvwb\logs\nvwb.log

Application log files

Linux\NVIDIA-Workbench\home\workbench\.nvwb\project-runtime-info\<PROJECT-NAME-AND-HASH>\<APP-NAME>-start.log

Log File Type

Location

Desktop application log files

/home/<USERNAME>/.config/NVIDIA/AI Workbench/logs/main.log

Server log files

/home/<USERNAME>/.nvwb/logs/workbench.log

CLI log files

/home/<USERNAME>/.nvwb/logs/nvwb.log

Application log files

/home/<USERNAME>/.nvwb/project-runtime-info/<PROJECT-NAME-AND-HASH>/<APP-NAME>-start.log

Log File Type

Location

Desktop application log files

/Users/<USERNAME>/Library/Application Support/NVIDIA/AI Workbench/logs/main.log

Server log files

/Users/<USERNAME>/.nvwb/logs/workbench.log

CLI log files

/Users/<USERNAME>/.nvwb/logs/nvwb.log

Application log files

/Users/<username>/.nvwb/project-runtime-info/<PROJECT-NAME-AND-HASH>/<APP-NAME>-start.log

Create an Error Report for Troubleshooting#

Error reports are zip files of comprehensive information about a Workbench location that can be useful for troubleshooting and debugging, especially by the support team. Error modals and messages in the Desktop App typically give you the option to create an error report for the given location that is reporting the error.

Creating an error report automatically tries to exclude sensitive or identifiable information, but there’s a broad range of ways that data can be included in logs so you should review the contents of an error report before you share it with anyone.

In some cases, you may need to create an error report outside of an error message or modal and you can do this in two ways.

Create Error Report for Local Location#

You can create an error report for your local Workbench from the system tray or status menu.

  1. Select the system-tray/status-menu

  2. Right click on the AI Workbench icon

  3. Select “Generate Error Report”

  4. Find the error report on your desktop

Create Error Report for Remote Location#

You need to use the CLI to create an error report for a remote location.

  1. Open a local terminal (on Windows open a terminal to the WSL distribution NVIDIA-Workbench)

  2. Run the command to create the report

    nvwb support create-bundle --context <location name>

  3. Find the error report at the file path that prints in the success message when the command completes.

AI Workbench Project Runtime Files#

AI Workbench creates files to track progress while building, configuring, and running each project. These files are essential for managing a project’s runtime operations.

Important

You should never commit these files to your Git repository.

The project runtime files are created in the following location based on your operating system.

Location: Linux\NVIDIA-Workbench\home\workbench\.nvwb\project-runtime-info\<PROJECT-NAME-AND-HASH>

Important

When you install AI Workbench on Windows, a folder named Linux\NVIDIA-Workbench (\wsl.localhost\NVIDIA-Workbench) is added to your File Explorer.

Location: /home/<USERNAME>/.nvwb/project-runtime-info/<PROJECT-NAME-AND-HASH>

Location: /Users/<USERNAME>/.nvwb/project-runtime-info/<PROJECT-NAME-AND-HASH>

The possible files and directories stored in this directory are the following:

File/Directory

Description

Containerfile

Generated on demand during the project build (can also be called a Dockerfile).

build-output.success

Build logs from the last successful build.

build-output.error

Build logs from the last failed build.

build-output.building

Build logs from the currently in-progress build.

cache/

Directory that acts as the build context for the container build.

edit.cache

The cache value of the last edit of the container image made by the library.

fetch-output.success

Fetch logs from the last successful fetch.

fetch-output.error

Fetch logs from the last failed fetch.

fetch-output.fetching

Fetch logs from the currently in-progress fetch.

mount/

Directory that stores mount definitions for the project’s configured mounts, as well as mount directories and unmount scripts for each mount. For more information, see Runtime Configuration (Environment Variables and Mounts).

rebuild.cache

The cache value of the last build of the container image.

runtime-flags.json

Flags used by AI Workbench.

secrets.env

The project’s secret environment variables and their values. For more information, see Environment Variables.

There are also one or more application files, depending on the applications in the project:

  • <APP_NAME>-start.log — Applicationstart logs for debugging issues. For example, jupyterlab-start.log.