Fleet Command Stack - x86 (1.56.1)
Improvement: Updated Fleet Command Stack to Kubernetes 1.28.7. Note that prior to updating your locations, review the Kubernetes Deprecated API Migration Guide to ensure your application is up to date with the Kubernetes API.
Fleet Command Stack - x86 (1.55)
This release does not introduce any features or bug fixes. Software dependencies are updated to reduce the risk of security vulnerabilities.
Fleet Command Stack - x86 (1.54.1)
Fix: Resolved an issue that caused the NVIDIA GPU Operator to get stuck in a crash loop on start or after configuring MIG.
Fleet Command Cloud Services (ngc.nvidia.com)
Improvement: Added support for deploying applications with Argo CD. Argo CD enables you to manage your applications by performing GitOps. For more information, refer to GitOps with Argo CD.
Fleet Command Stack - x86 (1.53.0)
Improvement: Updated Fleet Command Stack to Kubernetes 1.27.8 and Helm 3.12.3. Note that prior to updating your locations, review the Kubernetes Deprecated API Migration Guide to ensure your application is up to date with the Kubernetes API.
Improvement: Updated Fleet Command Stack with NVIDIA GPU Operator 23.9.0, NVIDIA Driver 535.129.03, CUDA 12.3, and Linux kernel 5.15.0-1041.41.
Fix: Resolved an issue that prevented filtering by the helm component on the Logs page.
Fix: Resolved an issue that prevented a system from restarting deployments or continuing an OTA after disk usage returns to less than 80%.
Fleet Command Cloud Services (ngc.nvidia.com)
Improvement: The style and appearance of the web interface for Fleet Command is updated to improve usability.
Improvement: The Logs page has a new component, gpu-operator, for logs from the NVIDIA GPU Operator.
Fleet Command Stack - x86 (1.49)
Improvement: When a deployment is modified, locations with the deployment that are currently offline show the deployment state as deploying.
Fix: Resolved an issue where the installation process on an edge system briefly displayed warnings during the applying network configuration step.
Fix: Resolved an issue that prevented collecting data for alert monitoring.
Fix: Resolved an issue for high-availability locations that caused newly added systems to become stuck in the connecting state if the initially added controller-worker node was offline.
Fix: Resolved an issue that prevented the NVIDIA GPU Operator from deleting pods that were stuck after MIG was reconfigured on a node.
Fleet Command Cloud Services (ngc.nvidia.com)
Improvement: When you remove a deployment, you receive a brief message that the deployment was removed successfully.
Improvement: The Fleet Command > Deployments page is enhanced to display the deployments in a table so they are easier to access. The location details page for a location is enhanced with a Deployments tab to show all the deployments at the location. Refer to Deploying a Helm Application for more information.
Improvement: If a location update fails, you can click cancel in the red error banner to cancel the update and try again later.
Fix: Resolved an issue that prevented viewing logs when All was specified in the location filter.
Fix: Resolved an issue where the location health did not display the current health.
Fix: Resolved an issue where exporting logs did not indicate that the output limit is 10,000 log messages.
Fleet Command Stack - x86 (1.49)
Improvement: Logs for edge devices are now rotated based on size rather than time. When logs reach 400 GB, the oldest logs are deleted to provide space for new logs.
Fleet Command Stack - x86 (1.48.1)
Improvement: Added support for L40 and L40S GPUs.
Improvement: Updated Fleet Command Stack to Kubernetes 1.27.5. Note that prior to updating your locations, review the Kubernetes Deprecated API Migration Guide to ensure your application is up to date with the Kubernetes API.
Improvement: NTP servers are updated to include pool.ntp.org and time servers from CloudFlare and Google.
Improvement: Increased the OTA timeout to 120 minutes for high-availability locations.
Fleet Command Stack - x86 (1.47.0)
Improvement: Updated Fleet Command Stack with NVIDIA GPU Operator 23.6.0, NVIDIA Driver 535.53.03, and Linux kernel 5.15.0.
Improvement: Updated Fleet Command Stack to Kubernetes 1.26.8. Note that prior to updating your locations, review the Kubernetes Deprecated API Migration Guide to ensure your application is up to date with the Kubernetes API.
Improvement: Added support for querying resource usage alerts and setting alerting levels. Refer to Alerts.
Fix: Resolved an issue where deployments were stuck in a ‘Deploying’ status.
NGC CLI (3.30.0)
Fix: Resolved an issue where deleting a deployment that did not exist reported success. In this release, the CLI reports that the target deployment is not found.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where simultaneous OTA update requests to the same system resulted in an incorrect display of location health status.
Fix: Resolved an issue where the system connection status displayed “Online” for unprovisioned systems.
Fleet Command Stack - x86 (1.45.0)
Fix: Resolved an issue where newly-added worker nodes in high-availability locations were stuck in ‘Connecting’ status.
Fix: Resolved an issue where alerts were disabled after an OTA update.
Fix: Resolved an issue where the root user does not have bash-completion functionality enabled.
Fleet Command Stack - x86 (1.44.2)
Note: General fixes and improvements.
NGC CLI (3.26.0)
Fix: Resolved an issue where GPU count did not display using the NGC CLI command
ngc fleet-command usage list.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where org information was not visible in the audit logs.
Fleet Command Stack - x86 (1.43.1)
Fix: Resolved an issue where, in some cases, system log files were not created on the edge device.
Improvement: Bash completion is now enabled for
kubectlon the Fleet Command Stack remote console, enhancing the edge system’s usability and debugging capabilities.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: System status updates now occur more frequently for transitions between online and offline states, resulting in faster and more accurate updates.
Fleet Command Stack - x86 (1.41.2)
Fix: Resolved an issue where a controller-worker system in a high-availability location could not start.
NGC CLI (3.24.0)
Fix: Improved logging output when running remote console CLI with the
--debugoption.
Fleet Command Stack - x86 (1.40.0)
Fix: Resolved an issue where an edge system would previously remain offline until all other systems were online. Edge systems can now come back online independently of other systems in a location.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue with documentation links in the
Settingspage and theSecurity Overridessection on theCreate Deploymentspage.
NGC CLI (3.23.0)
Improvement: Fleet Command CLI now features enhanced location and system health information.
Fleet Command Stack - x86 (1.39.0)
Improvement: Updated Fleet Command Stack to Kubernetes 1.25.9. Note that prior to updating your locations, review the Kubernetes Deprecated API Migration Guide to ensure your application is up to date with the Kubernetes API.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue with the description field placeholder text in the Add System dialog.
NGC CLI (3.22.0)
Fix: Resolved an issue where users cannot open a remote console using the NGC CLI.
Fleet Command Stack - x86 (1.38.1)
Improvement: Updated the Fleet Command Stack to use Ubuntu 22.04.
Fix: Resolved an issue where the Fleet Command edge device could become unavailable after a cloud services update.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where users couldn’t delete controller-worker nodes without worker nodes.
Improvement: Fleet Command now features enhanced location and system health indicators, providing users a way to quickly identify any issues related to availability or connectivity.
Fleet Command Stack - x86 (1.36.0)
Improvement: Updated Fleet Command Stack to Kubernetes 1.24.12. Note that prior to updating your location, review the Kubernetes Deprecated API Migration Guide to ensure your application is up to date with the Kubernetes API.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where users could not create high-availability locations.
Fix: Resolved an issue where, in some cases, the locations table did not refresh using the sort dropdown.
NGC CLI (3.19.0)
Fix: Resolved a problem where, in some cases, users could not connect to a node using the remote console CLI, resulting in the system displaying an “Unknown” status.
Fleet Command Stack - x86 (1.34.1)
Improvement: Updated Fleet Command Stack with Containerd 1.6.18, NVIDIA Driver 525.85.12, and NVIDIA Container Toolkit 1.12.0-1.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where, in some cases, users weren’t notified when Remote Console failed to start.
Fleet Command Discovery Image - x86 (fc-1.32.0-amd64.iso)
Note: General fixes and improvements.
Fleet Command Stack - x86 (1.32.0)
Fix: Resolved an issue where following an OTA update, the high-availability cluster may experience an issue on one of its nodes, rendering it unusable for applications.
Improvement: Improved system OTA performance by reducing update payload size.
Improvement: Updated Fleet Command Stack with Calico 3.24.5. For more information, view the Calico Release Notes.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where users cannot remove the location and node of a quarantined system.
Fix: Resolved an issue where, in some cases, users’ browsers were not automatically redirected to the secure HTTPS version of the Fleet Command web application.
Fix: Resolved an issue where the remote console maximum inactivity time was displayed incorrectly.
Improvement: Fleet Command will quarantine all nodes in high-availability locations if all controller nodes in the edge cluster are quarantined. For other locations, all nodes are quarantined if the controller node is quarantined.
Fleet Command Discovery Image - x86 (fc-1.31.1-amd64.iso)
Note: General fixes and improvements.
Fleet Command Stack - x86 (1.31.1)
Fix: Resolved an issue with OTA update of high-availability clusters.
Improvement: Fleet Command will now use the GPU Operator (version 22.9.2) to configure and manage GPU resources on edge systems.
Improvement: Updated Fleet Command Stack to Kubernetes 1.24.9, Containerd 1.6.15-1 and Helm 3.10.2. Note that prior to updating your location, review the Kubernetes Deprecated API Migration Guide to ensure your application is up to date with the Kubernetes API.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where the Remote Application Access session persisted even after the application was removed from Fleet Command.
Fix: Resolved an issue where the application update message banner in the Edit Deployment pane shows an incorrectly styled documentation link.
Fix: Resolved an issue where users cannot filter on Location Name and Systems columns in the Locations table.
Fix: Users can enable high availability locations with Fleet Command versions 1.25.1 or later.
Fix: Resolved an issue where quarantined systems were still accessible via remote console or remote app access.
Fix: Resolved an issue where remote application access times out in some cases.
Fix: Resolved an intermittent performance issue with the locations page.
Fix: Resolved an issue with the Fleet Command Terms of Use page.
Improvement: Redesigned the Organization > Usage pages to improve the display of GPU monthly and daily peak usages, system information, and log and private registry usage.
Improvement: Increased the maximum timeout to twenty-four hours for remote console.
Improvement: Increased overall platform security.
Improvement: Users will now see a message banner informing them of the effects of updating an application on the deployments page.
Improvement: Fleet Command releases and announcements will display in a banner on the top of the locations page.
NGC CLI (3.13.0)
Improvement: NGC CLI users will see improved performance with the auto-renewal of expired authorization tokens.
Fleet Command Discovery Image - x86 (fc-1.29.4-amd64.iso)
Fix: Resolved an issue where in some cases, controller-worker systems in high-availability locations fail to provision.
Fleet Command Stack - x86 (1.29.4)
Fix: Resolved an issue where, in some cases, applications lose access to GPUs after an indeterminate amount of time. Note: Fleet Command version 1.28.0 is now deprecated.
Fix: Resolved an issue with provisioning using the Fleet Command installer.
Fix: Resolved an issue where users do not see a message when a system has been accessed with elevated privileges.
Improvement: Updated Fleet Command Stack to NVIDIA Driver 515.86.01 to resolve a security issue.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where new users in the organization could not use the Remote Console CLI.
Fix: Resolved an issue filtering display columns using the Logs table settings.
Improvement: Fleet Command now deploys updates with minimal interruption to its management platform. Users can expect feature updates, performance and security enhancements with little or no downtime. Use over-the-air (OTA) to upgrade your edge systems to version 1.25.1 or later to take advantage of this new feature.
Fleet Command Discovery Image - x86 (fc-1.26.0-amd64.iso)
Note: General fixes and improvements.
Fleet Command Stack - x86 (1.26.0)
Fix: Resolved an issue where in some cases, core services get stuck after an OTA update.
Fix: Resolved an issue where Remote Console would not launch in some cases.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Updated location status filter text from “Offline Mode” to “Offline.”
Fix: Resolved an issue where the location table displayed an empty entry instead of 0 (zero) if there were no messages.
Improvement: Users can set a location to allow high availability (HA) during location creation. Refer to High Availability Locations for more information.
Improvement: Users will now see the drive alias, log, and container storage paths under the Storage Configuration pane in the system Details tab.
Improvement: Users can now specify the target Kubernetes namespace to deploy applications into.
NGC CLI (3.9.0)
Improvement: Added support for Remote Application Access CLI – a new interactive command that proxies traffic from edge devices to the local host. The CLI has advanced features to support forwarding traffic to selected interfaces and ports.
Improvement: Added support for creating high availability (HA) locations using the NGC CLI command option
--allow-high-availability.
Fleet Command Discovery Image - x86 (fc-1.25.1-amd64.iso)
Note: General fixes and improvements.
Fleet Command Stack - x86 (1.25.1)
Improvement: Multiple optimizations to improve Fleet Command availability during upgrades.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where changes made in the Edit Deployment did not show when the page first reloaded.
Fix: Updated Device Type label to System Type in the system Overview tab under Locations.
Fix: Resolved an issue where the location details page displayed an incorrect format for the last updated timestamp.
Fix: Resolved an issue where an application can be deleted while referenced in a deployment.
Improvement: Users can now update the application selection on the Edit Deployment page.
Improvement: Enhanced performance and scalability of remote console and remote application services.
Improvement: Users will be redirected to the same page in Fleet Command after logging in when their authentication expires.
NGC CLI (3.8.0)
Improvement: Added support for Remote Console CLI – a new interactive command to connect to get shell access to Fleet Command nodes.
Fleet Command Discovery Image - x86 (fc-1.24.1-amd64.iso)
Fix: Resolved an issue whereby installing the Fleet Command Discovery Image from a USB flash drive makes that drive available as an alternate storage option.
Fleet Command Stack - x86 (1.24.1)
Fix: Resolved an issue on edge systems where excessive CPU was being consumed for monitoring.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where in some cases, large volumes of log messages exceeded the allocated space for an organization.
NGC CLI (3.7.0)
Improvement: Added the ability to specify a Kubernetes namespace with the option
--target-namespaceduring deployment creation. For more information, refer to the Fleet Command NGC CLI documentation.Improvement: Added the ability to update an application in a deployment using the option
--application. For more information, refer to the Fleet Command NGC CLI documentation.
Fleet Command Discovery Image - x86 (fc-1.23.0-amd64.iso)
Improvement: Added the ability to specify custom aliases for mounted drives in the installer.
Fleet Command Stack - x86 (1.23.0)
Fix: Resolved an issue where updating the remote application access timeout value in Settings did not take effect for a session already in progress.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where the Deployment Security Overrides setting is not retained.
Fix: Resolved an issue where removing a deployment does not remove its corresponding remote application service.
Fix: Resolved issue with pagination control on Safari for the Kubernetes Labels panel under Locations > Details.
Fix: Resolved an issue where users saw an error when trying to start a remote console session that has already begun.
Improvement: Performance improvements with fetching locations, systems, and deployments.
Improvement: Added a Messages column to the Locations table to display the number of messages per location.
NGC CLI (3.6.1)
Improvement: Added support to display Kubernetes labels for locations with the NGC CLI command option
--show-labels. For more information, refer to the Fleet Command NGC CLI documentation.
Fleet Command Discovery Image - x86 (fc-1.22.0-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.22.0)
Improvement: Updated Fleet Command Stack to Kubernetes 1.23.10, Containerd 1.6.6-1, and NVIDIA Driver 515.65.01. Note that prior to updating your location, review the Kubernetes Deprecated API Migration Guide to ensure your application is up to date with the Kubernetes API.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where users encountered an error when clicking on “Try Again” when the remote console did not start.
Fix: Resolved an issue where users could not continuously type quotes inside the Deployments > Application Configuration text box.
NGC CLI (3.5.1)
Note: General improvements and fixes.
Fleet Command Discovery Image - x86 (fc-1.20.1-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.20.1)
Note: General improvements and fixes.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue on the Create Deployment page where the user cannot reselect the default pagination value.
Fix: Resolved an issue with the maximum length of the Location and Deployment names now set to 52 characters.
Fix: Resolved an issue where the Deployments > Locations panel displayed offline locations.
Fix: Resolved an issue where a System could be added to the location when an OTA update is in progress.
Fix: Resolved an issue where Application Compatibility did not update with changes to MIG configuration.
Improvement: Users can view Kubernetes labels for the set of GPUs available on a node on the Locations system details tab.
Improvement: Users can view Advanced Storage information on the Locations system details tab.
NGC CLI (3.5.0)
Improvement: Users will see more detailed help options for their partially-typed NGC commands.
Improvement: The NGC CLI command for locations,
ngc fleet-command location, will also show a list of systems.Improvement: Users can use the NGC CLI
versioncommand to show available versions with information and upgrade to a new version.
Fleet Command Discovery Image - x86 (fc-1.19.2-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.19.3)
Fix: Resolved an issue where in some cases, the system reports a core services error after provisioning.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where a Location could be deleted when it’s part of an active deployment.
Fix: Resolved an issue where changes made to the user name in the My Account Settings dialog are not reflected immediately in the user profile.
Fix: Resolved an issue where clicking the Add Location button on the Locations start page did not bring up the Add Location dialog.
Fix: Resolved an issue where sorting locations by version numbers did not work.
Improvement: Added data usage monitoring for Advanced Storage Configuration.
NGC CLI (3.4.0)
Fix: Resolved an issue where the CLI did not provide informative messages for invalid responses.
Fleet Command Discovery Image - x86 (fc-1.18.0-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.18.0)
Note: General improvements and fixes.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Removed redundant hyperlink on Fleet Command top-level breadcrumb navigation.
Fix: Resolved an issue where the locations table sort order did not persist upon refresh.
Fix: Resolved an issue where the Description field in the Edit System dialog showed stale data.
Fix: Resolved an issue where the Issues quick filter showed incorrect values.
Fix: Resolved an issue where the Status message on the location details page was incorrect when updating a location.
Fix: Resolved an issue where the locations Systems column tooltip showed an incorrect value.
Improvement: Added a timer for the Remote Console to show the user how long the console is available.
NGC CLI (3.3.0)
Improvement: Display appropriate message when metrics aren’t available.
Fleet Command Discovery Image - x86 (fc-1.17.0-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.17.0)
Fix: Resolved an issue where Remote Console sessions do not automatically clear upon expiration.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where the Logs table layout would extend beyond the page boundary.
Fix: Resolved an issue where Settings > Date and Time filters did not work in the Logs table.
Improvement: Redesigned the Locations page to use a table view. Added a details page where users can view systems details and perform actions associated with a particular location or system.
NGC CLI (3.2.0)
Fix: Resolved an issue where in some cases, listing org users in CSV format results in an error.
Fix: Resolved an issue with
--teamflag validation.
Fleet Command Discovery Image - x86 (fc-1.16.0-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.16.0)
Fix: Resolved an issue where the Remote Console session persisted even after the system was removed from Fleet Command.
Improvement: Added options for users to choose which edge server disks to format and mount as separate mount points.
Improvement: Added options for users to change the disks used to store logs and container images.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where Remote Application Access services across deployments for a particular location would be displayed in a consolidated manner.
Fix: Resolved an issue where a system in the process of being provisioned could not be deleted.
Fix: Resolved an issue where in some cases, Remote Console and Remote Application Access failures were not notified to the user.
NGC CLI (3.1.0)
Note: General improvements and fixes.
Fleet Command Discovery Image - x86 (fc-1.15.0-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.15.0)
Improvement: Additional safeguards have been implemented in applying OTA updates to the systems.
Fix: Resolved an issue where in some cases, systems could not be provisioned.
Fix: Resolved an issue where internal service requests would be incorrectly sent to a user-configured HTTP proxy.
Fix: Resolved an issue where in some cases, deployments get stuck in the “deploying” state.
Deprecation Notice: Fleet Command Stack version 1.9.3 uses Kubernetes 1.21, which is end-of-life on 2022-06-28. As such, in a future release, the Fleet Command Stack version 1.9.3 will be removed from eligible versions for new locations.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where Remote Console would not launch in some cases.
Fix: Resolved an issue with multiple search filters in the Organization > Audit Reports table.
Fix: Resolved an issue on the Organization > Usage page where clearing the GPU Inventory filter would not clear the search bar.
NGC CLI (3.0.0)
Improvement: Added system and application metrics via CLI.
Improvement: Increased overall performance of executing CLI commands. Note: Please uninstall previous versions of the CLI before installing the new version.
Fleet Command Discovery Image - x86 (fc-1.14.1-amd64.iso)
Improvement: Users can now select separate disk drives for OS and data storage in the Discovery Image Installer.
Fleet Command Stack - x86 (1.14.1)
Fix: Resolved an issue where Remote Console and Remote Application Access would not use the configured HTTP proxy.
Improvement: Deployments can now use the configured HTTP proxy through injected Helm values (
_FC_http_proxyand_FC_https_proxy).Fix: Resolved an issue where Fleet Command Stack Updates would not complete in some situations.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where the graph tooltip displays previous data on the Organization > Usage page after switching organizations.
Fix: Resolved an issue where systems added to the Locations page require a refresh to display.
Fix: Resolved an issue where the Locations page could not be displayed if there was an issue with Remote Console.
Fix: Resolved an issue where long names for locations, systems, or deployments may lead to errors in subsequent operations.
NGC CLI (2.16.0)
Note: General improvements and fixes.
Fleet Command Discovery System Image - x86 (fc-1.13.0-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.13.2)
Fix: Resolved an issue where removal of core services took precedence over others in the event of system errors.
Fix: Resolved an issue where in some cases, OTA updates continue to be in pending state after completion.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where the Remote Application Access expiration timer started before access was available.
Fix: Resolved an issue where in some cases, a deployment gets stuck in the “deploying” state.
Fix: Resolved an issue where the Location drop-down on the Logs page shows extra newlines when multi-selecting.
Fix: Resolved an issue where in some cases, Remote Application Access sessions started in close succession may result in duplicate timeout values.
NGC CLI (2.15.0)
Fix: Resolved an issue where Helm charts with quotes around numeric versions in Chart.yml were incorrectly parsed.
Improvement: CLI commands will use the display name as the org name for the user if set. Otherwise, the unique org name is used.
Note: Deprecated ‘–enable-dev-mount’ and ‘–disable-dev-mount’ options for the ‘settings’ command. Use ‘–enable-deployment-security-overrides’ and ‘–disable-deployment-security-overrides’ instead.
Fleet Command Discovery System Image - x86 (fc-1.12.7-amd64.iso)
Improvement: Added options for the user to choose between Quick Install and Advanced Install in the Installer UI.
Fleet Command Stack - x86 (1.12.7)
Improvement: Updated the Fleet Command stack for x86 to Helm 3.8.2.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue in the Logs page where dropdown values were misaligned after selection.
Fix: Resolved an issue where in some cases, HTTPS-based services were not supported by Remote Application Access.
Fix: Resolved an issue where referenced UI elements in applications were not accessible through Remote Application Access.
Improvement: Added a hint to inform users on setting max duration for Remote Application Access and Remote Console in Settings.
NGC CLI (2.14.0)
Fix: Resolved an issue where Remote Console sessions can be disabled without a warning.
Fix: Resolved an issue where users with FLEET_COMMAND_VIEWER role could not access location/system info.
Fleet Command Discovery System Image - x86 (fc-1.12.4-amd64.iso)
Improvement: Optimized the Fleet Command installer UI.
Fleet Command Stack - x86 (1.12.4)
Fix: Addressed a Security vulnerability.
Improvement: Updated the Fleet Command Stack for x86 to 510.47.03 (CUDA-11.6) NVIDIA driver. Added support for A30x and A100x GPUs.
Improvement: Resolved an issue where entering an invalid activation code took a long time to be verified.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where users can redeploy applications on locations in an error state.
Fix: Resolved an issue where all external service applications were labeled as Web applications in Remote Application Access.
Fix: Resolved an issue where remote application access timeouts allowed invalid entries.
Fix: Resolved an issue where a user may see features outside of their roles in some cases.
Improvement: Remote Console automatically opens in a new browser tab.
Improvement: Increased overall platform security.
NGC CLI (2.13.2)
Fix: Resolved an issue where users encountered errors while downloading large volumes of logs.
Fleet Command Discovery System Image - x86 (fc-1.12.1-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.12.1)
Improvement: Added detailed deployment status in logs.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where the maximum GPUs value under Organization > Usage didn’t update correctly.
Fix: Resolved an issue where GPU names were not consistent.
Fix: Resolved an issue where Application Compatibility did not update with changes to MIG configuration.
Fix: Resolved an issue where Remote Management timeouts could be set to invalid entries.
Fix: Addressed a security vulnerability.
Fleet Command/NGC CLI (2.13.0)
Improvement: Added the ability to enable/disable remote application access.
Fleet Command Discovery System Image - x86 (fc-1.11.0-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.11.0)
Fix: Addressed a security vulnerability.
Fleet Command Cloud Services (ngc.nvidia.com)
Improvement: Updated onboarding messaging for improved user experience.
Improvement: Additional authentication is no longer required to access Remote Console with default privileges.
Improvement: Added support for Remote Application Access that allows users to access web-based application services running on edge systems remotely from their local machine.
Improvement: Updates to Deployment Application configuration are more immediate.
Fix: Resolved an issue with displaying GPU type details for GPU Inventory table in organization usage analytics.
Fix: Resolved an issue where duplicate application names were allowed.
Fleet Command/NGC CLI (2.12.0)
Fix: Resolved an issue where Remote Console timeouts could be updated when disabled.
Fleet Command Discovery System Image - x86 (fc-1.10.1-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.10.1)
Improvement: Added support for Remote Application Access (see Fleet Command Cloud Services above).
Improvement: Added support for GFD (GPU Feature Discovery) that allows users to automatically generate labels for the set of GPUs available on a node.
Fix: Addressed a security vulnerability.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where, when editing a deployment, users could select location versions that didn’t support security overrides selected.
Improvement: Added backend support for OTA component updates.
Fix: Resolved an issue with the date missing from the GPUs Under Management chart.
Fleet Command/NGC CLI (2.10.0)
Fix: Resolved an issue where users could remove applications that were still part of deployments.
Improvement: Updated deployment creation error messages.
Fleet Command Discovery System Image - x86 (fc-1.9.6-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.9.6)
Fix: Addressed a security vulnerability.
Improvement: Removed application deployment timeout to support slow connections, large or complex applications.
Fix: Resolved an issue where the location update status would change while the update was still in progress.
Improvement: Updated Fleet Command Stack to Kubernetes 1.22.5 and Calico 3.21.4. Note that prior to updating your location, review the Kubernetes Deprecated API Migration Guide to ensure your application is up to date with the Kubernetes API.
Improvement: Added specific messages when deployments fail due to security restrictions.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where in some cases the logging component dropdown shows erroneous entries.
Fix: Resolved an issue where the monthly usage chart was showing the last month twice.
Fix: Resolved an issue where in some cases the storage usage graph does not show.
Improvement: Increased overall security of the cloud service.
Fix: Resolved an issue where in some cases after a cloud service update some users could not log in or systems could not pair.
Improvement: Added additional deployment security overrides. Note, because additional overrides are available, the administrator setting to enable overrides has been returned back to disabled. Administrators must re-enable the use of overrides to use the previous and new overrides.
Fleet Command/NGC CLI (2.9.0)
Improvement: Added support for security overrides.
Fleet Command Discovery System Image - x86 (fc-1.9.3-amd64.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.9.3)
Improvement: Added support for security overrides.
Fix: Addressed a security vulnerability.
Improvement: Added details to deployment failed alerts.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where in some cases there are text artifacts shown in the pull-down for logging components.
Fleet Command/NGC CLI (2.8.1)
Fix: Resolved an issue where the max GPU usage was reported incorrectly.
Improvement: Added MIG controls.
Fleet Command Discovery System Image - x86 (fc-1.8.4-amd64.iso)
Improvement: Updated the file name for the Fleet Command system image to fc-<version>-<arch>.iso.
Fleet Command Stack - x86 (1.8.4)
Note: General improvements and fixes.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where selecting a location for a deployment on any page but the first brings you back to the first page.
Fleet Command/NGC CLI (2.8.0)
Fix: Resolved an issue where private registry usage information was incorrect.
Fleet Command Discovery System Image - x86 (edge_1.8.2.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.8.2)
Improvement: Optimized bandwidth used for logging and overall improved efficiency and robustness.
Improvement: Added support for NVIDIA A2 GPU.
Fix: Patched an Ubuntu OpenSSL security issue.
Fix: Resolved an issue where systems were being marked as having used elevated privileges with remote console incorrectly.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Updated cloud components to improve security.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue issue where GPU information wasn’t shown in some cases.
Fix: Resolved an issue where the system image was named incorrectly.
Fix: Updated cloud components to improve security.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where in some cases Remote Console isn’t accessible after starting.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where in some cases the Remote Console creation message continues to display after re-authentication.
Improvement: Updated the look and feel of tables.
Improvement: Remote Console updated to allow multiple Remote Consoles concurrently. This completes the roll-out of “Remote Console 2.0” that requires Fleet Command stack version 1.3 or higher.
Fix: Resolved an issue where in some cases switching teams and orgs resulted in a bad state.
Fix: Resolved an issue where max GPU usage was incorrect.
Fleet Command Discovery System Image - x86 (latest_bfe67070.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.7.0)
Note: General improvements and fixes.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where usage data was being collected. Data from 12/1/2021 through 12/6/2021 will report as 0.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue with instructions shown during log export.
Fix: Resolved an issue with the layout of MIG controls in some browsers.
Improvement: Added search and filtering for locations, deployments and applications.
Improvement: Remote console access is only marked when elevated privileges are enabled.
Fix: Resolved an issue with MIG configuration when non-MIG-capable GPUs are present.
Fleet Command Discovery System Image - x86 (latest_ed18f478.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.6.3)
Improvement: Further increased security by restricting the use of PersistentVolumes.
Fleet Command Cloud Services (ngc.nvidia.com)
Improvement: Streamlined the NGC/Fleet Command sign-in flow for new users.
Improvement: Logs default to last 1 hour instead of last week.
Fix: Resolved an issue where Firefox was displaying MIG information incorrectly.
Note: With this update to the Fleet Command cloud service, remote console requires locations to be updated to Fleet Command stack 1.3.0 or higher. Remote console 1.3.0 and higher provides direct access to the system shell for better performance and user experience and improved security by removing the need for an additional firewall port.
NGC CLI (2.7.0)
Fix: Resolved an issue when MIG configuration is not valid.
Fleet Command Discovery System Image - x86 (latest_3e1d8fb6.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.6.0)
Fix: Resolved an issue where, in some cases after reboot a system ends up in an issue state.
Fix: Resolved an issue where, in some cases a location shows in an issue state after an update.
Fleet Command Cloud Services (ngc.nvidia.com)
Improvement: The first location is automatically collapsed when accessing the Locations section.
Improvement: Added the ability to enable/disable and set MIG profiles for MIG-capable GPUs. “View System Details” has also been updated to “System Details” to reflect this additional capability.
Improvement: Decrease section loading time and addressed cases where loading introduced flickers in some cases.
Fix: Resolved an issue where in some cases logs shows “No files found” incorrectly.
Fix: Resolved an issue where “update available” was being shown when locations are offline.
Improvement: Increased overall platform security.
Fix: Resolved an issue with reporting GPU type details.
NGC CLI (2.6.0)
Fix: Resolved an issue where no error is presented when downloading logs to a location with insufficient permissions.
Fix: Resolved an issue where errors being relayed to the CLI were incorrect.
Improvement: Added MIG controls.
Fleet Command Discovery System Image - x86 (latest_f46c76c2.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.5.2)
Fix: Resolved an issue where in some cases a system will get stuck rebooting during cloud pairing.
Fix: Resolved an issue where a remote console fails to start in some cases.
Fix: Resolved an issue where in some cases an update fails without cause.
Fix: Resolved an issue where applications may not restart properly after an update.
Improvement: Added support for MIG control.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where the default gateway was not being shown for some cases in the system details page.
Fix: Resolved an issue where some logs weren’t mapped to the correct system.
Fleet Command Discovery System Image - x86 (latest_e3ba0ce2.iso)
Improvement: Added support for serial consoles.
Fleet Command Stack - x86 (1.4.2)
Note: General improvements and fixes.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where updates were shown as available when a location was offline.
Note: Other general improvements and fixes.
Fleet Command Discovery System Image - x86 (latest_5a8c20ac.iso)
Note: General improvements and fixes.
Fleet Command Stack for x86 (1.4.1)
Note: General improvements and fixes.
Fleet Command CLI (NGC CLI 2.4.0)
Fix: Resolved an issue where log download range could be negative.
Fix: Resolved an issue where log download limit wasn’t clear.
Fleet Command CLI (NGC CLI 2.3.0)
Fix: Added device access control in settings.
Fleet Command Discovery System Image - x86 (latest_d781c554.iso)
Note: This version replaces 1.0.10 that was recently rolled back.
Fix: Resolved an issue where missing network parameters can cause issues with provisioning.
Fleet Command Stack for x86 (1.3.0)
Note: This version replaces latest_fc33f6a3.iso that was recently rolled back.
Improvement: Updated the Fleet Command Stack for x86 to Containerd, Kubernetes 1.21.3, Helm 3.6.3, and the 470.57.02 (CUDA 11.4) NVIDIA driver.
Improvement: Added support for IP address changes when in DHCP mode to allow for pre-provisioning and environments without IP address reservations.
Fix: Resolved an issue where some logs were inaccessible in certain cases.
Fix: Resolved an issue where missing network parameters can cause issues with updating.
Improvement: Remote console updated to direct access to the system shell for improved performance and user experience.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where Fleet Command stack versions weren’t shown in the correct order.
Fleet Command CLI (Registry CLI 2.2.3)
Improvement: Updated the CLI help for downloading logs from locations (One, many or all locations can be targeted).
Improvement: Deprecated deployment execute command which is no longer needed (Create, update and remove are sufficient).
Fleet Command Discovery System Image - x86 (latest_a066e669.iso)
Note: The Fleet Command discovery system image latest_fc33f6a3.iso released 08/24/2021 has been rolled back to latest_a066e669.iso to support the rollback of Fleet Command stack 1.0.10. Discard any copies of latest_fc33f6a3.iso and download latest_a066e669.iso from the settings page of Fleet Command.
Fleet Command Stack - x86 (1.0.7)
Note: Fleet Command stack 1.0.10 released on 08/24/2021 has been rolled back to 1.0.7 due to issues updating from some Fleet Command stack versions and when provisioned from some versions of the Fleet Command discovery system image. For locations and systems that successfully updated to 1.0.10, you can continue to use them but you may experience issues adding new systems. Any unused locations created with version 1.0.10 should be removed and recreated with version 1.0.7.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where the filter input fields in the usage section can overflow in the UI.
Fix: Updated usage export files to “Fleet Command”.
Improvement: Edit deployments now allows users to gracefully leave vs having to cancel.
Improvement: Clarified impact of location update.
Fix: Resolved an issue where resizing the create location prevented inputs in some cases.
Fix: Added notification when Fleet Command subscription expires.
Fleet Command Discovery System Image - x86 (latest_fc33f6a3.iso)
Fix: Resolved an issue where installing Fleet Command stack on A100 failed in some cases.
Improvement: Added no DHCP mode.
Improvement: Added the ability to specify subnet mask.
Fleet Command Stack - x86 (1.0.10)
Note: General improvements and fixes.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where app config was not shown on edit deployment in some cases.
Fix: Resolved an issue where in some cases, it wasn’t possible to change the locations of a deployment.
Fix: Resolved an issue where with some screen resolutions, the device access enablement text is cropped.
Fix: Resolved an issue where the device access enablement status was not being shown properly when editing a deployment.
Fix: Resolved an issue where remote management was being shown for offline locations.
Fix: Resolved an issue where users weren’t being redirected to the sign-in screen when their session had expired.
Fix: Resolved an issue where remote console timeouts were allowing invalid entries.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where Fleet Command roles were being shown for defining teams.
Feature: Optimized sign-in workflow and added support for enterprise identity federation (Early access)
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Reset Fleet Command Terms of Use (TOU) acceptance to address cases where some users could not accept the TOU.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where permissions were incorrect for the operator role.
Fix: Resolved an issue where deleting a deployment that uses an NGC Private Registry can impact other applications deployed from the same registry.
Fix: Resolved an issue where in some cases, the UI would show blank until refreshed.
Fix: Resolved an issue where in some cases, a user can end up in a state where they appear to be logged in but aren’t.
Fix: Resolved an issue where the NGC version wasn’t showing while accessing Fleet Command.
Improvement: Added administrator setting to allow users to enable device access.
Improvement: Multiple optimizations to improve UI performance.
Fleet Command Discovery System Image - x86 (latest_a066e669.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.0.7)
Improvement: Updated systemd and patched kernel to improve security.
NGC CLI (2.2.0)
Fix: Resolved an issue where incorrect format was not being flagged.
Fix: Resolved an issue where users with invalid API keys could not configure the CLI.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where permissions were not correct for the operator role.
Fix: Resolved an issue where deleting a deployment that uses an NGC Private Registry can impact other applications deployed from the same registry.
Fix: Resolved an issue where in some cases, the UI would show blank until refreshed.
Fix: Resolved an issue where in some cases, a user can end up in a state where they appear to be logged in but aren’t.
Fix: Resolved an issue where the NGC version wasn’t showing while accessing Fleet Command.
Improvement: Added administrator setting to allow users to enable device access.
Improvement: Multiple optimizations to improve UI performance.
Fleet Command Discovery System Image - x86 (latest_0ae6dbda.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.0.5)
Improvement: Updated systemd and patched kernel to improve security.
Fleet Command CLI (Registry CLI 2.0.0)
Fix: Resolved an issue where private registry storage was not included in the usage output.
Fix: Resolved an issue where file name length could cause an error.
Fix: Resolved an issue where the help commands were not enabled for components.
Fix: Resolved an issue where location was duplicated in the usage list command.
Improvement: Added the ability to use CLI to filter by component list for logs.
Fleet Command Cloud Services (ngc.nvidia.com)
Improvement: When downloading logs, if a file of the same name exists, -1 will be appended and incremented on subsequent downloads automatically.
Improvement: Added advanced system details in system information.
Improvement: Consolidated NGC CLI into a single CLI. Commands available are based on roles assigned.
Fix: Resolved an issue where application display name and description cannot be edited.
Fix: Resolved an issue where navigation selections weren’t working.
Fix: Resolved an issue where viewers and operators couldn’t access the locations page.
Fix: Resolved an issue where logging components available were not always current.
Fix: Resolved an issue where user activity columns couldn’t be properly resized without overlapping information.
Fix: Resolved an issue where user activity filters were not working or not applicable.
Fix: Resolved an issue where the browser tab was not showing the navigation information when using the organization section.
Fix: Updated usage charts to use UTC.
Improvement: Updated the styling of the usage page.
Fix: Resolved an issue where teams weren’t shown properly when changing orgs.
Fix: Resolved an issue where user was being redirected to the catalog when accessing teams.
Fleet Command Discovery System Image - x86 (latest_05f45852.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (1.0.2)
Note: General improvements and fixes.
Fleet Command CLI (Registry CLI 2.0.0)
Improvement: Updated error message text for logging.
Fix: Resolved an issue when setting remote console settings.
Improvement: Aligned logging behavior with UI, when no location, system or deployment is provided, show all is the default.
Fix: Resolved an issue where Fleet Command roles were incorrect in some contexts.
Fleet Command Cloud Services (ngc.nvidia.com)
Improvement: Increased security of the cloud service architecture.
Fix: Resolved an issue where loading would be displayed when there were no entities.
Fix: Resolved an issue where locations wouldn’t load if remote management was disabled.
Fix: Resolved an issue where the NGC version was not showing while Fleet Command was active.
Improvement: Browser tab now shows navigation details.
Improvement: Added a details page that provides networking information and other information for a selected system.
Fleet Command CLI (Registry CLI 1.47.0)
Fix: Resolved an issue where remote console session time setting was not being validated.
Improvement: Updated the help for enabling device access to match the UI.
Fix: Updated location information to allow display of system information.
Improvement: Updated application list command to list applications in alphabetical order.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue with consecutive system updates. Note, system updates are restricted to 1.0.0 to address this issue.
Improvement: Improved security of the cloud service.
Improvement: Improved resiliency of the cloud service.
Improvement: Added user activity logging.
Improvement: Updated remote console dialog for improved user experience.
Improvement: Updated logging query drop downs to being relational.
Fix: Resolved an issue where custom dates for logging queries were not showing as custom.
Fix: Added warning when logging query exceeds 60,000 pages.
Fix: Disabled remote console, reboot and logs while systems are still connecting.
Fix: Added error handling for system update failures.
Fleet Command Discovery System Image - x86 (latest_7f969d83.iso)
Fix: Resolved an issue with stability of systems with NVIDIA MLNX networking.
Fleet Command Stack - x86 (1.0.0)
Note: Fleet Command Stack 1.0.0 includes a number of improvements for additional security. Please modify the application or change the deployment settings as applicable before updating to prevent impact to your applications.
Fix: Resolved an issue where some deployments were not showing as failed in some situations.
Fix: Resolved an issue where in some cases a deployment gets stuck in the “deploying” state.
Improvement: Added explicit HostPath mounts for /tmp and /opt for additional security.
Improvement: Disallowed application access to HostIPC and HostPID for additional security.
Improvement: Restricted application’s ability to access privileged Docker permissions.
Improvement: Limited allowed HostPath mounts except for /tmp and /opt to read-only for additional security.
Improvement: By default, disabled access to /dev resources for additional security, including existing deployments, and added a deployment setting to enable access as needed.
Fleet Command CLI (Registry CLI 1.46.0)
Improvement: Added logging to CLI.
Fleet Command Cloud Services (ngc.nvidia.com)
Improvement: Improved resiliency of cloud service updates.
Improvement: Added the ability to download the system image source and checksum for additional security verification and validation.
Fix: Resolved an issue where reboot was not issued when selected.
Fix: Resolved an issue where quarantined systems could not be removed.
Fix: Resolved an issue where the remote console minimum session time could not be set to 1 minute.
Improvement: Increased security of remote console by requiring authentication before starting a session.
Improvement: Updated “Last deployed” to “Last modified” to better represent the information presented.
Fix: Resolved an issue where only the current month’s usage was being shown for monthly usage.
Improvement: Updated role names from EGX to Fleet Command. Note, also updated for CLI.
Improvement: Added user activity reports.
Improvement: Updated layout and look of NGC.
Fleet Command Discovery System Image - x86 ()
Note: General improvements and fixes.
Fleet Command Stack - x86 (<0.4.53>)
Improvement: General fixes to improve the Fleet Command stack security.
Fleet Command CLI (Registry CLI 1.45.0)
Fix: Removed deprecated log links from info command.
Fix: Resolved an issue where deployment naming requirements were not being enforced.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where settings were not retained between updates. Note, this requires configuring the settings after the update to have them persist through future updates.
Fix: Resolved an issue where in some cases, an update triggers a remote attestation failure.
Improvement: Integrated logging into the NGC user interface.
Improvement: Updated layout of the settings section.
Fleet Command Discovery System Image - x86 (latest_3bece252.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (0.4.52)
Improvement: Added root file system modification as a quarantine trigger.
Improvement: Improved resiliency of application deployments.
Fix: Resolved an issue where in some cases, an update triggers a remote attestation failure.
Fix: Updated to Linux Kernel 5.4.0-72 to resolve an overlay vulnerability.
Fleet Command CLI (ngc.nvidia.com CLI 1.44.0)
Improvement: Added usage analytics to CLI.
Fleet Command Cloud Services (ngc.nvidia.com)
Improvement: Added usage analytics. Note, this requires using Fleet Command Stack 0.4.47 and above.
Improvement: Update style of messages and icons.
Improvement: Increased overall cloud infrastructure security.
Fix: Resolved an issue where settings were not always persisted.
Improvement: Increased resiliency of system pairing.
Improvement: Updated remote attestation failure from an alert to system quarantine.
Improvement: Increased resiliency of remote console.
Improvement: Added encryption at rest for all logs stored in NGC.
Improvement: Added run-time validation of the immutable root filesystem for additional security. Note, this only applies when the system is reimaged, it will not be implemented via over-the-air updates.
Note: Updated system role terminology from master-worker to controller-worker.
Fleet Command Discovery System Image - x86 (latest_589b20d9.iso)
Note: General improvements and fixes.
Fleet Command Stack - x86 (0.4.47)
Improvement: Update remote attestation failure from an alert to system quarantine.
Improvement: Increased resiliency of network configuration.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where a duplicate activation code could be created.
Improvement: Simplified remote console controls.
Fix: Updated applications page to be sorted alphabetically by display name.
Fix: Resolved an issue where Helm Repo URL was only referenced as Helm Repo in a warning.
Fix: Reduced the time the remote console banner is shown after a session is ended.
Fix: Resolved an issue where a system name couldn’t be re-used.
Improvement: Optimized location page loading.
Fix: Resolved an issue where a remote console session doesn’t always get created.
Fleet Command Discovery System Image - x86 (latest_fb973578.iso)
Improvement: Use network FQDN instead of generic.
Improvement: Added an option to specify hostnames and/or IPs that bypass the proxy (if specified).
Improvement: Added support for proxy authentication.
Fleet Command Stack - x86 (0.4.44)
Improvement: Updated to NVIDIA Driver 460.32.03 (CUDA 11.2).
Fix: Resolved an issue where OTA caused remote attestation failure.
Improvement: Limited HostPath mounts for additional security. Refer to the restrictions in the Fleet Command Application Developer Guide. Note, updating to Fleet Command Stack 0.4.44 may result in applications not being updated for restrictions to stop running or fail to deploy.
Fleet Command Cloud Services (ngc.nvidia.com)
Fix: Resolved an issue where it was possible to create an application with a duplicate display name.
Fix: Resolved an issue where remote console warnings were not always shown with Fleet Command Stacks prior to 0.4.41.
Fleet Command CLI (ngc.nvidia.com CLI 1.43.0)
Improvement: Fleet Command CLI commands updated from egx to fleet-command.
Fix: Resolved an issue where Helm name was referenced as application name.
Fix: Resolved an issue where CLI and UI had different character limits for application Helm chart name.
Improvement: Increased cloud service security.
Fix: Resolved an issue where the API key button was missing text for some roles.
Fix: Resolved an issue where UI elements were not aligned.
Fix: Resolved an issue where the application name was incorrect in the edit deployment screen.
Fix: Resolved an issue where some Fleet Command Stack versions weren’t selectable when creating a new location.
Improvement: Increased overall platform security via additional monitoring and architecture improvements.
Improvement: Improved overall platform resiliency and performance.
Fleet Command Discovery System Image - x86 (latest_c931313c)
Improvement: Improved security and resilience of the Fleet Command Discovery System Image.
Fleet Command Stack - x86 (0.4.41)
Improvement: Improved security and resilience of Fleet Command Stack.
Improvement: Added warning when systems are accessed with elevated privileges.
Known issues:
Remote console access warnings requires Fleet Command Stack 0.4.41. Warnings with previous stacks may not be represented in all cases.
Improvement: Updated application CLI to allow for duplicate Helm chart names.
Improvement: Updated application list CLI to display name.
Fix: Resolved an issue where it was possible to create duplicate application display names via CLI.
Fleet Command Cloud Service (ngc.nvidia.com)
Fix: Resolved an issue where systems and/or locations may show an issue state while being rebooted or going through an update.
Fix: Resolved an issue where location wasn’t showing an issue state when core services were in an error state.
Fix: Resolved an issue where deployments show the wrong application name.
Improvement: Added status indicator when location, application or deployment data is being loaded.
Fix: Resolved an issue where the API key doesn’t show.
Fleet Command Discovery System Image - x86 (latest_c862ac44.iso)
Improvement: Added protection against multiple login attempts.
Fleet Command Stack - x86 (0.4.39)
Improvement: Added protection against multiple login attempts.
Note: Versions 0.4.30 and 0.4.31 have been removed from availability to ensure users benefit from the fixes and improvements in 0.4.36.
Improvement: Added the ability to create applications with the same Helm chart name as long as the application name is unique.
Improvement: Added idle time control for remote console, supported via UI and CLI.
Fix: Resolved an issue where remote administration controls were visible to non-administrators.
Fix: Resolved an issue where the remote console showed “continue session” before the session was available to be resumed.
Fix: Resolved an issue where remote console access warning was triggered before the session had started.
Improvement: Improved start time of remote console sessions.
Improvement: Resolved an issue where there was a reboot warning at the start of a remote console session.
Fix: Resolved an issue where the system name was duplicated in some alerts.
Fix: Resolved an issue where certain sequences of actions could result in loss of remote session settings.
Fix: Resolved an issue where additional systems cannot be added to a location in some situations.
Fix: Resolved an issue where systems were incorrectly identified as cloud instances during removal.
Improvement: Added successful pod creation as a criterion for a successful deployment.
Fix: Resolved an issue where a failed deployment showed deploying instead of failed.
Improvement: Improved response time to close remote console when remote management is disabled.
Fleet Command Discovery System Image - x86 (latest_9268ee85)
Fix: Resolved an issue where a failed install does not automatically log the user out.
Note: Users should discontinue using the latest_d27de765.iso and latest_597f3424.iso to benefit from the fixes above.
Fleet Command Stack - x86 (0.4.36)
Fix: Resolved an issue where a remote console could be accessed via local network.
Fix: Resolved an issue where an unexpected port was open.
Improvement: Streamlined the installation process by removing the user prompt that prevented fully unattended installation.
Fix: Updated Fleet Command branding.
Improvement: Updated the installer UI.
Improvement: Added advanced installation settings menu.
Improvement: Added support for customer-defined networking: HTTP/HTTPS proxy and custom NTP servers.
Improvement: Updated network connectivity test from ICMP to HTTPS.
Improvement: Added Helm and Docker interfaces to remote console.
Note: Versions 0.4.30 and 0.4.31 have been removed from availability to ensure users benefit from the fixes and improvements above.
Improvement: Added user-configured columns for locations list CLI.
Improvement: Added support for deployment alerts in CLI.
Improvement: Improved readability of system details in locations and deployments CLI by adding additional indentation.
Improvement: Updated system removal CLI messaging to match new UI workflow. (Released 1/26/2021)
Improvement: Updated system version CLI messaging to match UI. (Released 1/26/2021) 2/2/2021 8:17pm PT
Improvement: Fleet Command Stack (v0.4.30) updated to Kubernetes 1.18.8, NVIDIA K8s device plug-in version 0.7.0 and moved to the signed version of the NVIDIA 450.51.06 Driver (to support secure boot). Note, Fleet Command Stack v0.4.31 is functionally equivalent to v0.4.30 but is made available to allow the user to try over-the-air updates (OTA) available in this release.
Improvement: Added measured boot with remote attestation for added security. On failure, the system will be marked as not fully secured. Note, TPM must be set to SHA256 for this security feature to function.
Improvement: Added the ability to reboot and access a terminal/console remotely on Fleet Command systems. Enable remote console from the “Settings” section to enable these capabilities.
Improvement: Added the ability to access system, location and deployment/application logs.
Improvement: Enabled secure boot for added security.
Improvement: Added the ability to update systems over-the-air (OTA). Note: For this release, v0.4.30 and v0.4.31 are functionally equivalent but users can choose v0.4.30 during location creation to try updating to v0.4.31.
Improvement: The Fleet Command Service Discovery system image is now accessible from the new “Settings” section of Fleet Command. Note, the Service Discovery system image is now separate from the Fleet Command Stack and neutral to the Stack version being deployed.
Improvement: Alerts significantly expanded for location, system and deployment/application messages and issues.
Improvement: Updated UI style and format.
Improvement: Added ability to access the Fleet Command terms and conditions.
Improvement: Improved security by preventing privileged applications.
Improvement: Increased platform security monitoring.
Fix: Removal workflow updated to prevent orphaned systems.
Fix: Resolved an issue where application gets stuck during deployment on Xavier-based systems (Technical preview).
Fix: Resolved an issue where activation code creation was delayed.
Fix: Resolved an issue where API key sometimes does not get set.
Fix: Implemented additional security controls to prevent ssh access.
Fix: Resolved an issue where a system could get stuck during the removal process.
Fix: Resolved an issue where a successfully deployed application shows as failed.
Fix: Resolved an issue where Fleet Command Operators had no access.
Fix: Resolved an issue where Xavier-based systems failed to provision (Technical preview)
Fix: Resolved an issue where a multi-disk system could not be provisioned.
Improvement: Additional security monitoring in-place.
Fix: Resolved an issue where HPE systems could not be provisioned.
Improvement: General improvements to security and operational performance and stability.
Note: Display name for locations, systems, and deployments have been deprecated to optimize searching and naming clarity.
Known issues:
The Fleet Command Stack installation on some ATOS Bull Sequana Edge servers may hang after the first boot at “Started Journal Service”. If the installation is stuck for more than 15mins, reboot the system and when the installation options (GNU GRUB) menu is presented, press e on your keyboard and then remove console=ttyS0 from the installation parameters and press F10 to reboot the system and continue the installation. 03/09/2021 9:49am PT
Improvement: Added support for user configured columns in Fleet Command CLI.
Improvement: Added support for systems in location information and updated associated help information.
Fix: Updated help text with proper case.
Fix: Resolved an issue where using an existing appconfig for a deployment would fail.
Improvement: Added CLI for accessing logging URLs.
Note: Displayname removed from location, system and deployment CLI due to deprecation of this feature.
Improvement: Added support for using NGC Private Registries from other orgs (requires permissions in both orgs).
Fix: Fixed an issue where successful deployments were showing an issue.
Fix: Resolved an issue where the system state was not fully retained between reboots on x86.
Improvement: Users can now view and subscribe to service status and maintenance notifications at fleetcommand.statuspage.io.
Improvement: Applications can now be deployed in any namespace instead of being restricted to the default namespace.
Improvement: Systems are now protected by secure boot.
Fix: Resolved an issue where the state was not retained between reboot on Xavier (Technical preview).
Improvement: Implemented minimal-downtime update, unless otherwise noted on fleetcommand.statuspage.io, at most users will experience small delays during ngc.nvidia.com Fleet Command updates.
Improvement: Fleet Command Stack updated to Kubernetes 1.17, Helm 3.3.3 and NVIDIA Container Runtime for Docker v2.5.0.
Fix: Resolved an issue where read-only users could create applications and add/change the API key.
Fix: Resolved an issue where master-worker could be removed if used in a deployment.
Improvement: Added message banner and improved messaging for offline mode.
Improvement: Show the last 4 digits of the API key for auditing key used.
Fix: Resolved an issue where adding the API key via CLI wasn’t showing in the UI.
Note: Users will be presented with Fleet Command terms of service (TOS) on next login.
Fix: Resolved an issue where messages were being shown on the same line in the message banner.
Fix: Resolved an issue where the application update CLI was not outputting an error when no arguments were passed.
Note: Location add-apikey has been removed, use application add-apikey.
Fix: resolved an issue where new locations and systems added to existing locations were not working as expected.
Improvement: When executing deployments via CLI, the command will now list deployment details in addition to a success message on completion.
Improvement: When listing location info via CLI, it will now include system status.
Improvement: EGX updated to Fleet Command in navigation panel.
Note: Monitoring disabled while capabilities are further integrated into the UI.
Fix: Resolved an issue where the UI reported an API key had been entered when it hadn’t been entered yet.
Improvement: Streamlined and more discoverable workflow for adding an NGC API key (Now located on the Applications, Create and Edit Deployments pages) including updates to the EGX CLI (Note: add-apikey is now deprecated).
Improvement: Updated look and instructional text for the Applications, Create and Edit Deployments pages.
Improvement: NGC Private Registry teams now supported.
Fix: Resolved an issue where some API keys that were entered were missing.
9/25/2020 12:45am PT (Updated 9/30/2020 12:30pm PT to reflect what was actually deployed 9/28/2020 and 9/30/2020)
Improvement: During EGX OS installation, a password can be set to access the system via a virtual console.
Fix: Resolved an issue whereby EGX Service Discovery USB could only be used once.
Fix: Removed options from EGX Service Discovery that are no longer valid.
Improvement: During EGX OS install, the activation code is validated before allowing the user to continue.
Fix: Resolved an issue where old EGX OS installations could prevent new EGX OS installations.
Improvement: Storage on EGX systems is now encrypted.
Improvement: Users are now warned if TPM is not enabled during EGX OS installation.
Fix: Resolved an issue where deployments failed due to the name having too many characters.
Improvement: EGX OS is now based on Ubuntu 20.04 and the CUDA 11 NVIDIA 450.51.06 driver.
Fix: Resolved an issue where it was possible to create a duplicate system name.
Fix: Resolved an issue where some data could be lost if services stopped working.
Improvement: EGX Manager services are now more resilient and fault-tolerant.
Improvement: During the EGX Manager OS installation, the network interface is automatically selected, and the user is only prompted when there is no obvious automatic selection (multiple interfaces).
Improvement: During the EGX Manager OS installation, network connectivity requirements such as DNS, routing, and NTP are now checked and issues reported as applicable before the installation starts.
Fix: Resolved incompatibility issues with IE 11.
Improvement: When editing deployments, change the button to “Update” to reflect the action better.
Improvement: When using CLI, controller-worker or worker roles no longer need to be specified; the role is automatically selected.
Improvement: Add confirmation when creating and updating deployments.
Improvement: Empty fields consistently report back blank for empty fields, in some cases those fields would report back as none.
Improvement: Default to name if no display name is entered.
Fix: Resolved an issue where deployment create and update were failing in some situations.
Fix: Resolved an issue where application name length was not being enforced on creation resulting in failure when removing.
Fix: Resolved an issue where appconfigs were showing blank instead of default if no appconfig was entered.
Fix: Resolved an issue where location name validation was being applied to the combined location:system instead of individually which was resulting in location creation failures.
Fix: Resolved an issue where non-existent appconfigs were no longer being flagged as an error.
Improvement: System and location status via CLI is now in parity with UI.
Improvement: Deployment status via CLI is now in parity with UI.
Bug: Fixed a mistake in the location’s error message.
Bug: Fixed an issue where the application configuration was displaying as empty when it should have been populated.
Improvement: Added the ability to update the locations in a deployment when editing a deployment.
Improvement: Ensure the integrity of a deployment by preventing the last system in a location from being deleted if it is used in a deployment.
Improvement: Added a notification when a system is created and automatically expands the location where that system was added.
Improvement: System, location and deployment statuses and messages are now wired-up and active.
Bug: Solved an issue where systems that were deleted were still showing.
Bug: Solved an issue where clear all locations in a deployment were only working on first click.
Improvement: Improved security of communications between back-end components.
Bug: Resolved an issue where deployments were not being logged.
Improvement: Master-worker and worker roles are now automatically set on creation and no longer require user selection.
Improvement: Support NGC orgs with “-” in their name.
Improvement: Fixed an issue where activation code is showing after it was used.
Bug: Fixed an issue where newly created locations weren’t showing.
Improvement: Renamed GNAC to activation code and updated UI to reflect the change.
Bug: Resolved an issue where duplicate location names were allowed.
Bug: Resolved an issue where API key entry was generating an error on the first input.
Improvement: Updated the source for system count to improve accuracy.
Bug: Resolved an issue where deployment name length wasn’t being validated.
Bug: Resolved an issue where deployment names, location names, and other names were being truncated.
Improvement: Improved error message when using locations CLI.
Improvement: Added validation to display name and name for locations, applications, app configs and deployments.
Improvement: CLI uses environment variables to make EGX CLI available (until the product is officially released).
Improvement: Improved error message when appconfig download fails due to disk write issue.
Bug: Fixed an issue where deployment via CLI was not working.
Improvement: Added IP address to the information listed for systems.
Improvement: Added functionality to prevent locations from being removed if they are used in a deployment, or they contain systems - this prevents them from becoming orphaned.
Improvement: System role (master-worker or worker) is now automatically assigned on creation.
Improvement: Added functionality that prevents master-workers from being deleted if they are supporting additional workers.
Improvement: Added the ability to enter a display name for deployments.
Bug: Fixed an issue where the deployment configuration interface bleeds through the application pull down.
Bug: Fixed an issue where EGX read-only users had more permissions than expected.
Bug: Fixed an issue where the deployment update button wasn’t available when it should have been.
Improvement: Increased data resilience of Fleet Command.
Bug: Fixed an issue where Location Health in Monitoring was displaying N/A for metrics.
Bug: Fixed an issue where Location Health in Monitoring was not reporting Not Ready systems.
Bug: Fixed an issue where the data source was not being set automatically when using System Metrics in Monitoring.
Bug: Fixed an issue where logging displayed a confusing error on first use.
Bug: Fixed an issue where System Metrics in Monitoring were displaying N/A.
Improvement: Platform upgrade automation is now more resilient.
Improvement: Private registry supported on Xavier (Beta).
Bug: Fixed an issue where config changes were not being saved during edit deployment.
Improvement: “-” replaced with a connecting symbol.
Bug: Fixed an issue where systems were all displaying “-” or “connecting” symbols - healthy, unhealthy, or connecting statuses are now displayed properly.
Improvement: Application configuration entry during deployment creation has been streamlined.
Improvement: Streamlined edit deployment process and made it more discoverable - edit deployment is now available from the deployment ellipsis.
Bug: Resolved an issue where System Metrics on the Monitoring page was pointing to the wrong metrics.
(Preview) Improvement: Added Deployment Metrics to Monitoring. Select the default namespace and workload associated with your application.
Improvement: Added ability to edit locations and systems.
Bug: Fixed and further restricted EGX User permissions.
Improvement: Updated Location and Deployment card style to optimize information display and surface system issues to the card instead of only being available when expanded. System issues that affect the location will now be associated with the affected system and then surfaced at the card level - replacing “Location status” and “Location issues”.
Bug: Fixed Location, System, Application and Deployment name validation to prevent periods from being used.
Improvement: Applications now include a free form display name field instead of being restricted to the Helm chart application name.
Improvement: Application description is now editable along with newly added application display name.
Improvement: Systems now display “-” during pairing, healthy when all system parameters report within expected parameters and unhealthy if any system issues are present
Known issues:
Deployments page not fully “wired” yet - number of systems, deployment status, deployment issues and location status may show incorrect or no information at all.
Systems all displaying “-” even when in a known good state.
System issues that affect the location, formerly “Location status” or “Location issues” are not yet “re-wired” as issues displayed on the card.
Monitoring does not reflect updated system or location names.