Getting Started

This chapter provides pointers on using Fleet Command to add a location, a system, an application, and create a deployment. The guide also covers setting up a virtual representation of the physical system at the edge location and managing that system remotely.

Note

Fleet Command provides configurable settings for finer control over remote management, deployment security, and logging. A user with admin privileges can customize the default settings on the Settings page.

A location defines an edge location/site, such as a retail store or hospital where edge deployments will be installed.

  1. To add a location, navigate to Fleet Command > Locations. Click Add Location on the top right corner of the page.

    loc-page-add-location.png

  2. In the Add Location dialog, fill in the location information.

    add-location-ha-dialog.png

    • Location Name: Enter a location name. The name must start with a lowercase letter, followed by digits (0-9), lowercase letters (a-z), or a dash (-).

    • Description: Enter a description for the location.

    • Fleet Command Version: Choose the Fleet Command version for the location. While Fleet Command supports versions released within the last two months, we recommend selecting the latest version for all new locations.

    • Address: Enter the location’s physical address.

    • Allow High Availability: Enable high availability (HA) for this location. This option is available only during creation time and cannot be changed once the location is created. Refer to High Availability Locations for more information.

    • Enable Argo CD: Enable GitOps for this location with Argo CD.

  3. Click Add Location to save.

To remove a Location, refer to the Managing your Fleet section.

Updating Location Details

You can also edit the location description and address. Click on the ellipsis menu for a location and select Edit Location:

location-menu-ann.png


The Edit Location window opens.

edit-location-ha-dialog.png


  • You cannot change the high availability option after the location is added.

  • You can enable Argo CD after the location is added, but you cannot disable Argo CD after it is enabled.

Location Health and Availability

On the Locations page, the Health and Availability columns display the current status of a location.

  • Health : Indicates whether a location is healthy or has an issue that limits its functionality.

  • Availability : Indicates whether a location is connected to a Fleet Command service or is offline.

The green check icon indicates the location is healthy and ready to accept deployments.

location-status-column.png

Location Details

After you add a location, you can view the location details by clicking on location row.

On the location details page, the location status is displayed in the panel on the left.

By default, the Systems tab is selected and shows the systems at the location. You can click the Deployments tab to view the application deployments at the location.

location-details-status-ann.png

  • Location Health: (Same value as Health on the Locations page.) Displays whether a location is healthy or has an issue which limits its functionality.

  • Last Reported Health: Displays the last time a location reported its status and its last status. This field will only appear if the location is currently in an unhealthy state.

  • Availability: (Same value as on the Locations page.) Displays whether a system is connected to Fleet Command service or is offline.

The following table shows the possible statuses for the location health:

Value

Symbol

Description

Good
status-good.png
The location is ready to accept deployments.
Issue
status-issue.png
The location might be able to accept deployments. Expand the location for more information.
Unknown
status-unknown.png
The location could not be contacted. Applications might still be running.

The following table shows the possible statuses for the location availability:

Value

Symbol

Description

Reachable
status-reachable.png
The location is connected to the Fleet Command Cloud service.
Unreachable
status-unreachable.png
The location is not connected to the Fleet Command Cloud service.
Note

If your location experiences an issue, more information is available by clicking on the location.

For more information on the location and system health impact on Fleet Command functionality, refer to Understanding Location and System Health.

A system is a virtual representation of a physical system at the edge location. After the system is deployed at the edge location, it should be paired with a system in Fleet Command. This helps ensure that when software is deployed in Fleet Command, it is deployed to the physical system.

Follow these steps to add a system in Fleet Command.

  1. You can add a system in two ways.

    • Click Fleet Command > Locations and then click the actions button for a location and select Add System from the menu.

      location-menu-small.png

    • Select a specific location from the locations page and then click Add System from the location details page.

      add-new-system.png

  2. On the Add System window, select the type of system. Click Add System to go to the next step.

    add-system-type-x86.png

  3. Fill in the system information.

    add-system-x86.png

    • System Name: Enter a name for the new system. The name must start with a lowercase letter, followed by digits (0-9), lowercase letters (a-z), or a dash (-).

    • Description: Enter a description for the new system.

    • System Version: This is the software your system will be running. It is typically prefilled based on your location version.

    Click Confirm to go to the next step.

  4. Follow the instructions in this dialog to download the installer to flash your device. Click Finish when done.

    add-system-x86-step2.png

  5. After you add a system, the window returns to the location details page. The activation code provided by Fleet Command and the expiration time are displayed in the system messages.

    location-details-ann.png

    Use the activation code to pair a system in Fleet Command with a physical device at the edge location. Proceed to the Pairing a System with Fleet Command section to complete the pairing process.

    Some preparation at the physical edge site is needed before the system is ready to accept the activation code. Refer to the Edge Site Requirements section for more information.

    Here are the descriptions for the contents on this page:

    [1] Location messages: displays messages for this location.

    [2] Location health: displays location details, including the current health, version and available updates.

    [3] Systems summary: displays status summary for your systems that you can filter on.

    [4] System details: contains detailed information for each system. There will be one details pane per system for this location.

    [5] System messages: displays messages for this system.

    [6] Options menu to access commands for a system.

    [7] Overview, Details, and Multi-Instance GPU (MIG) tabs.

    The overview tab contains the status, config, and system type.

    The details tab contains your system’s description, GPU types, networking, storage configuration, and Kubernetes labels.

    The MIG tab contains MIG configuration information. Refer to Multi-Instance GPU for details.

    Note

    Multi-Instance GPU (MIG) support, a feature that allows secure partitioning of GPUs into separate GPU instances, can be enabled from the system details pane after deploying an application. Refer to Multi-Instance GPU for more information.


    Important

    If the measured boot failed for your edge site, you might see the following error message for the system.

    getting-started-27.png

System Configurations

There are two types of system configurations.

The first system added to the location is considered the controller-worker. The controller-worker is a Kubernetes controller node with the capability to run workloads. Additional nodes added to the location are configured as worker nodes.

two-systems-ann.png


If high availability is enabled for a location, the first three systems are set to the controller-worker role, and subsequent systems are worker nodes. Refer to High Availability Locations for more information.

four-systems-ann.png


System Commands

To access specific commands for a system, click on the options menu:

system-details-menu-ann.png


System Health and Connectivity

On the systems details tab, the health and connection fields in the overview tab display the current status.

  • Health : Displays whether a system is healthy or has an issue that limits its functionality.

  • Connection: Displays whether a system is connected to the Fleet Command Cloud service or is offline.

Additionally, the following field appears on the overview tab if the system health is unknown.

  • Last Reported Health: Displays the last time a system reported its status and its last status. This field only appears if the system is in an unhealthy state.

The following table shows the possible statuses for the system health:

Value

Symbol

Description

Online
status-online.png
The system is connected to Fleet Command Cloud service.
Good
status-good.png
The system is ready to run applications.
Issue
status-issue.png
The system might not be ready to run applications. Refer to the banner for more information.
Connecting
status-connecting.png
The system is connecting, updating, or rebooting.
Unknown
status-unknown.png
The system health is unknown.

The following table shows the possible statuses for the system connection:

Value

Symbol

Description

Online
status-online.png
The system is connected to Fleet Command Cloud service.
Offline
status-offline.png
The system is not connected to Fleet Command Cloud service.

In this example, the green check icon and online connection indicate the system is ready to run applications.

system-status-good-online-ann.png

If a system is unhealthy, the last reported health field displays the last status.

system-status-unknown-last-reported-ann.png

For more information on the location and system health impact on Fleet Command functionality, refer to Understanding Location and System Health.

System Issues and Messages

Fleet Command displays system issues and informational messages in the Location and System details panes. If your system is experiencing an issue, you will see a red error banner with details and recommendations for the next step.

system-details-errors.png

Cannot Add System

When a location is being updated, such as an OTA, you won’t be able to add a new system. You will see the following message:

cannot-add-system.png


Pairing a System with Fleet Command

To pair a system with Fleet Command, complete the following steps:

  1. Download the Fleet Command Discovery Image (ISO) from the Fleet Command Settings page to your local laptop or desktop (Ubuntu Linux, Mac, or Windows). The image will be named similarly to the following convention: fc-<version>-<arch>.iso, and administrators should always use the ISO version that matches the Location version they will be deploying. For additional information, refer to Adding a Location.

    Note

    For additional security, validate the system image with the available checksum from the Settings page, as shown below. Refer to the checksum validation for more information.

    getting-started-28.png


  2. You can load the Fleet Command Discovery Image (ISO) into your system in two ways.

Booting from Virtual Media

Follow each system provider’s instructions to load the Fleet Command Discovery Image (ISO) as a virtual media.

Booting from USB

The correct system image is loaded onto a USB flash drive. Use the following steps to create a bootable USB flash drive with the Fleet Command Discovery Image (ISO):

  1. Locate a 16GB USB flash drive.

  2. On your local laptop or desktop, burn the image:

    • Ubuntu Linux: For creating a bootable USB flash drive on Ubuntu, use the Startup Disk Creator application on your Ubuntu computer. Follow the simple instructions outlined in the Ubuntu tutorials.

    • Apple macOS: For creating a bootable USB flash drive on your Mac, follow the instructions outlined in the tutorial.

    • Microsoft Windows: For creating a bootable USB flash drive on your Windows PC/Laptop, follow the tutorial’s instructions.

    • The USB flash drive has now been loaded with a Fleet Command system image and is ready to be used to boot systems.

  3. Configure the system for the Fleet Command Stack installation:

    • Connect a monitor and a keyboard directly to the system or remotely access the system via the platform’s IPMI interface.

    • Power on the system

    • Update firmware as needed to meet company policies for systems deployed with limited access.

    • Access the BIOS settings

    • Set the system time to the local time (Fleet Command Stack will automatically set the OS time to local UTC)

  4. Follow each system provider’s instructions to enable TPM and Secure Boot.

  5. Activate the system once the system has been appropriately configured.

    Note

    You must have your Fleet Command activation code for the next set of steps. The Fleet Command activation code is only valid for 12 hours after it is created.


  6. The following steps will activate the system with Fleet Command via the activation code:

    • Plugin the provided Fleet Command Discovery USB flash drive to the edge system

    • Reboot or power on the system

    • You will be prompted to create an administrator password. This password will be used to access the edge system in the future.

    getting-started-03.png

    • Press Enter when finished.

    Note

    The Administrator password requirements are listed below. It is recommended that this password complies with any existing corporate password policy.

    • At least 14 and at max 128 characters long

    • At least one lowercase character

    • At least one upper-case character

    • At least one digit

    • At least one of the following symbols: $,#@%^=./~_-!


  7. Network and NTP will be configured automatically by the Fleet Command Stack. Please wait while the configuration is completed; it may take several minutes.

    getting-started-04.png

    Important

    If the configuration is not completed automatically, it may require a manual setup in the Advanced Configurations dialog. You will need to select the network with access to the internet.

    getting-started-05-600.png

    • You have the option to add the configuration manually in Advanced Configurations. To add the Advanced Configuration option, press CTRL+C after you enter the Administrator password.

    • You can select Configure Network Interfaces if you have multiple network interfaces to choose the right one. Once you choose the correct network interface, you can use DHCP or manually configure the static IP. You will need to select the network with access to the internet.

    getting-started-32.png

    • Once you select the correct network interface, you can use DHCP or configure the edge system with a static IP.

    • You will need to select the network with access to the internet.

    getting-started-36.png

    • Example of static IP configuration screen.

    getting-started-37.png

    • You can add the NTP servers with the Configure NTP option if you want to add any.

    getting-started-33.png

    • If you want to add any proxy to the edge system, select Configure HTTP Proxy and add a proxy address.

    getting-started-34.png

    • If you want to customize the utilization of physical storage devices attached to the system, select Configure Storage and follow the steps outlined in Advanced Storage Configuration.

    • If you want to add any additional configuration, select Edit Configuration Manually and edit the configuration.

    getting-started-35.png


  8. Enter the activation code.

    getting-started-06.png


  9. Once the activation process has been completed, you will see the location change from “offline” to “online” in the UI. The system’s IP address is also shown in Fleet Command UI now.

Note

The installation and activation are a continuous process. If you cannot complete the installation and activation simultaneously, completing the process is recommended later. If the activation code has expired, you must remove and add the system to the location again.

Once a system at the edge has been paired with a system in Fleet Command, you are now ready to deploy software to the edge using the Fleet Command interface.

Configuring Your System

Once your system is set up, you can configure for multi-instance GPU. For more information, refer to Multi-Instance GPU Configuration.

Accessing Your System

Once your system is set up, you can remotely access it using Fleet Command Remote Console. For more information, refer to Remote Console.

Fleet Command provides status indicators to inform users of the health of locations and systems deployed at edge sites. With this feature, users can quickly identify any issues related to availability or connectivity. It’s important to note that certain functionality may be temporarily unavailable depending on the location and system status. The following table outlines the functionality available under different locations and system statuses.

location-system-health.png


An application defines what can be deployed to edge locations. The primary element of an application is the Helm chart which describes how the application will be deployed.

Note

To learn how to create an application for Fleet Command, refer to the Application Developer Guide.

  1. Select Fleet Command > Applications in the Fleet Command navigation menu.

  2. Click Add Application on the top right side of the page.

    apps-add-app-button.png

  3. Fill in the details for the application.

    apps-add-app-helm.png

    • Display Name: This must be the unique name of the application.

    • Description: Create a description for the new application.

    • Source: Select Helm Application from the menu.

    • Helm Chart Name: The name of the Helm chart.

      Note

      You can create multiple applications with the same Helm Chart Name as long as the Display Name is unique.

    • Helm Chart Version: This must be the exact version of the application listed in the Helm repository.

    • Helm Repo URL (no trailing slash): The Helm repository URL where the Helm chart is stored. Helm repository URL can be NGC Catalog Helm Charts, public Helm repositories, or NGC Private Registry for maximum security.

      For example, https://helm.ngc.nvidia.com/egxdefault.

To remove an application, refer to the Managing your Fleet section.

Deployments join systems at physical edge sites with the applications hosted on Fleet Command in the cloud. Through a deployment, you select the application to deploy and the locations to deploy it to. The deployment operates based on the application’s Helm chart and given configuration.

  1. Select Fleet Command > Deployments.

  2. On the Deployments page, click Create Deployment.

    deps-create-button.png

  3. On the Create Deployment page, enter the information for the deployment.

    deps-create-helm.png

    • Deployment Name: Enter a name for the deployment.

    • Application: Select an application for the deployment.

    • Target Namespace Specify a Kubernetes namespace for the deployment. The default value is default. You cannot edit this value after you create the deployment.

    • Application Configuration: Add configuration overrides for this deployment if the application allows.

      • Using the standard template format (same as Helm Charts), you can override existing configuration from the values YAML file or set a value.

        The following image shows the YAML for setting camera IP addresses for a video analytics application:

        deploying-to-the-edge-03.png

    • Select Locations: Select the locations to deploy to.

    • Security Overrides (Optional): Use this feature to override the default security settings for a deployment. Security overrides can only be selected during the initial deployment creation and cannot be modified afterwards without recreating the deployment. If you choose to apply security overrides, a message is shown on the Locations and Deployments pages. For detailed information on the options, refer to the Security Overrides section.

  4. Select the Before Deploy checkbox and click Deploy.

    Note

    If multiple deployments/applications are assigned to a single location, you must confirm that they do not conflict. Some examples to keep in mind are service/application ports, resources (over subscribing), duplicate applications, or other application conflicts within the same namespace. Ensure you search for the application Helm chart rather than the container, model, or collections.

To perform deployment updates or removals, refer to the Managing your Fleet section.

Checking Deployment Status Across All Locations

The Deployments page summarizes the status of all deployments across all locations.

  1. Select Fleet Command > Deployments.

    fc-deployments-one-issue.png

  2. Optional: To filter the deployments that are not operating normally, click the Issues filter.

  3. Click a row in the table of deployments to view the deployment page.

Checking Deployment Status at One Location

  1. Select Fleet Command > Locations and then click the row in the table of locations.

  2. On the location details page, click the Deployments tab.

    fc-locations-deployments.png

  3. Click the row to access the deployment details page.

The following table describes the three possible deployment statuses that can appear in the Status column:

Status

Icon

Description

Deploying
dep-deploying.png
Application is deploying at the location.
Deployed
dep-deployed.png
Application is deployed successfully and pods are running at the location.
Failed
dep-failed.png
Application could not be deployed successfully at the location.

Accessing the Application

After you deploy an application, you can access the application by using the Remote Application Access feature from Fleet Command. For more information on using this feature, refer to Remote Application Access.

Previous Prerequisites
Next Managing your Fleet
© Copyright 2022-2024, NVIDIA. Last updated on May 7, 2024.