Skip to content

Commit 0da802e

Browse files
committed
fix(installer): support stock DGX OS on Station
Signed-off-by: Aaron Erickson <[email protected]>
1 parent b37e142 commit 0da802e

10 files changed

Lines changed: 831 additions & 68 deletions

‎docs/get-started/prerequisites.mdx‎

Lines changed: 15 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -41,10 +41,13 @@ If the group change is not active in the current shell, the installer exits with
4141
If you choose the native Linux Ollama install path, the onboard wizard also requires `zstd` for Ollama archive extraction.
4242
The installer also requires `strings` from `binutils` to verify the OpenShell binary before it continues with OpenShell install work.
4343

44-
On a DGX Station GB300 running the generic Ubuntu 24.04 ARM64 image, accepting express install prepares the host with NVIDIA open driver `610.43.02`, Docker CE `29.6.1` with Buildx, and NVIDIA Container Toolkit `1.19.1`.
45-
DGX OS, NVIDIA BaseOS images, and other Station generations are outside this automatic preparation boundary and stop before host preparation.
46-
On those systems, set `NEMOCLAW_PROVIDER` or `NEMOCLAW_NO_EXPRESS=1` explicitly to continue without Station host automation.
47-
The preparation probes package and runtime state first, reuses exact matches, and installs only missing pinned packages, including the NVIDIA Container Toolkit libraries and `nvidia-ctk` CLI.
44+
DGX Station GB300 express install accepts either the generic Ubuntu 24.04 ARM64 image or stock DGX OS `7.2.0`, `7.4.0`, or `7.5.0`.
45+
A stock DGX OS image qualifies only when a safe, root-owned `/etc/dgx-release` marker identifies `DGX Server for GALAXY-GB300`.
46+
An unknown version, malformed or unsafe marker, NVIDIA BaseOS image, or other Station generation stops before host preparation.
47+
On an unqualified system, set `NEMOCLAW_PROVIDER` or `NEMOCLAW_NO_EXPRESS=1` explicitly to continue without Station host automation.
48+
49+
On the generic Ubuntu path, accepting express install prepares the host with NVIDIA open driver `610.43.02`, Docker CE `29.6.1` with Buildx, and NVIDIA Container Toolkit `1.19.1`.
50+
Preparation probes package and runtime state first, reuses exact matches, and installs only missing pinned packages, including the NVIDIA Container Toolkit libraries and `nvidia-ctk` CLI.
4851
It permits only the reviewed factory transition from `dkms` `3.0.11-1ubuntu13` to `1:3.4.0-1ubuntu1`.
4952
After reboot, preparation enables NVIDIA's packaged CDI refresh path and service, requires the `nvidia.com/gpu=all` device, and verifies it with a real container launch.
5053
If the packaged refresh fails or does not produce that device, preparation prints service diagnostics and stops for administrator repair.
@@ -59,9 +62,16 @@ It does not install a host CUDA toolkit or Docker Compose.
5962
If any other existing prerequisite version differs, preparation stops instead of changing it automatically.
6063
After changing pinned packages, the installer exits with status `10`; reboot, sign in, and run the printed command, which pins the exact accepted NemoClaw commit before resuming express setup.
6164

65+
On a qualifying stock DGX OS image, express install validates the factory stack in place against the local default Docker daemon.
66+
It requires a loaded GB300 driver with zero volatile corrected and uncorrected ECC errors, active Docker and containerd services, working Docker Buildx, `nvidia-ctk`, the `nvidia.com/gpu=all` CDI device, and successful `nvidia-smi` device-visibility probes through both CDI and `--gpus all`.
67+
This path does not install or replace host packages, enable or restart services, generate CDI configuration, or rewrite the Docker runtime.
68+
The preparation helper writes a private audit log under `~/station-bootstrap-logs`.
69+
Aside from the installer's separately disclosed `docker`-group membership grant when the current trusted account needs Docker access, its only persistent stock-runtime effect is that the digest-pinned visibility-test image can remain in the Docker cache.
70+
These probes establish local container GPU visibility; physical acceptance of the pinned managed-vLLM and Nemotron Ultra recipe remains a separate gate.
71+
6272
<Warning title="DGX Station Support Status">
6373
DGX Station remains Deferred.
64-
Full NemoClaw onboarding with this recipe has not completed end-to-end validation on physical DGX Station hardware.
74+
Product validation of this express-install boundary on physical DGX Station hardware is still pending.
6575
</Warning>
6676

6777
<Warning title="Docker Group Access">

‎docs/get-started/quickstart.mdx‎

Lines changed: 19 additions & 11 deletions
Original file line numberDiff line numberDiff line change
@@ -119,15 +119,19 @@ Use these details when your first-run path needs more control.
119119
If the installer adds you to the `docker` group, run the printed `newgrp docker` command before you rerun it.
120120
On macOS, start Docker Desktop or Colima first.
121121

122-
DGX Spark, Station GB300 hosts running the generic Ubuntu 24.04 ARM64 image, and Windows WSL offer an interactive express-install path that chooses a managed local inference option for the platform.
123-
On that Station configuration, accepting the express prompt selects the pinned `nemotron-3-ultra-550b-a55b` managed-vLLM recipe and completes onboarding without more provider, model, policy, or sandbox-name choices.
124-
DGX OS, NVIDIA BaseOS images, and other Station generations stop before host preparation.
122+
DGX Spark, qualifying Station GB300 hosts, and Windows WSL offer an interactive express-install path that chooses a managed local inference option for the platform.
123+
Station accepts the generic Ubuntu 24.04 ARM64 image and stock DGX OS `7.2.0`, `7.4.0`, or `7.5.0` when a safe, root-owned `/etc/dgx-release` marker identifies `DGX Server for GALAXY-GB300`.
124+
Accepting the Station prompt selects the pinned `nemotron-3-ultra-550b-a55b` managed-vLLM recipe and completes onboarding without more provider, model, policy, or sandbox-name choices.
125+
Unknown DGX OS versions, malformed or unsafe release markers, NVIDIA BaseOS images, and other Station generations stop before host preparation.
125126
Set `NEMOCLAW_PROVIDER` or `NEMOCLAW_NO_EXPRESS=1` explicitly to continue on an unqualified Station without host automation.
126-
It probes the pinned driver, Docker, Buildx, and NVIDIA Container Toolkit versions, reuses exact matches, installs missing pins, and permits only the reviewed factory `dkms` transition.
127+
On generic Ubuntu, host preparation probes the pinned driver, Docker, Buildx, and NVIDIA Container Toolkit versions, reuses exact matches, installs missing pins, and permits only the reviewed factory `dkms` transition.
127128
It establishes NVIDIA CDI through the packaged refresh service, then proves both CDI and `--gpus all` with real container launches.
128129
If the packaged refresh fails or does not advertise `nvidia.com/gpu=all`, preparation prints service diagnostics and stops for administrator repair instead of generating CDI configuration directly.
129130
After it changes pinned packages, the installer exits with status `10` at the required reboot boundary; reboot, sign in, and run the printed exact-commit command to resume the accepted recipe without another prompt.
130-
This automation does not change Station's Deferred support status; physical end-to-end validation remains open.
131+
On stock DGX OS, the installer instead validates the existing GB300 driver and volatile ECC state, local Docker and Buildx, `nvidia-ctk` and CDI, and `nvidia-smi` device visibility through both GPU container launch paths.
132+
It does not install or replace host packages, enable or restart services, or rewrite the Docker runtime; aside from the private preparation log and separately disclosed `docker`-group membership grant when needed, only the digest-pinned visibility-test image can remain cached.
133+
The visibility probes do not replace physical acceptance of the pinned managed-vLLM and Nemotron Ultra recipe.
134+
This automation does not change Station's Deferred support status; product validation on physical DGX Station hardware remains pending.
131135
Pass `--station-deepseek` to use DeepSeek V4 Flash for a Station demo instead.
132136
The flag selects the interactive express prompt and requires terminal access.
133137
Refer to [Platform Support](../reference/platform-support) and [Choose an Inference Provider](../inference/learn-and-choose/choose-inference-provider) for the current platform behavior.
@@ -194,19 +198,23 @@ Use these details when your first-run path needs more control.
194198
curl -fsSL https://www.nvidia.com/nemoclaw.sh | bash
195199
```
196200

197-
On DGX Spark, Station GB300 hosts running the generic Ubuntu 24.04 ARM64 image, and Windows WSL, interactive installation offers express install after you accept the third-party software notice.
201+
On DGX Spark, qualifying Station GB300 hosts, and Windows WSL, interactive installation offers express install after you accept the third-party software notice.
198202
After you confirm the interactive express prompt, the installer switches the remaining onboarding to non-interactive mode, allows `sudo` password prompts for required host changes, and selects the managed local inference path for that platform.
199203
DGX Spark uses managed vLLM with `qwen3.6-35b-a3b-nvfp4` by default.
200204
DGX Station express install explicitly selects `nemotron-3-ultra-550b-a55b` instead of the Station managed-vLLM profile default, `deepseek-v4-flash`, and discloses the approximately `352 GB` model download before confirmation.
201-
Before onboarding, the Station path requires Station GB300 with the generic Ubuntu 24.04 ARM64 image and checks for NVIDIA open driver `610.43.02`, Docker CE `29.6.1` with Buildx, and NVIDIA Container Toolkit `1.19.1`.
202-
DGX OS, NVIDIA BaseOS images, and other Station generations are outside this automatic preparation boundary.
205+
Before onboarding, the Station path accepts either the generic Ubuntu 24.04 ARM64 image or stock DGX OS `7.2.0`, `7.4.0`, or `7.5.0` with a safe, root-owned `/etc/dgx-release` marker for `DGX Server for GALAXY-GB300`.
206+
Unknown versions, malformed or unsafe release markers, NVIDIA BaseOS images, and other Station generations are outside the automatic preparation boundary.
203207
Set `NEMOCLAW_PROVIDER` or `NEMOCLAW_NO_EXPRESS=1` to continue without Station host automation.
204-
Preparation reuses exact versions, installs missing pinned packages, permits only the reviewed `dkms` transition from `3.0.11-1ubuntu13` to `1:3.4.0-1ubuntu1`, and refuses every other mismatched version.
208+
On generic Ubuntu, preparation checks for NVIDIA open driver `610.43.02`, Docker CE `29.6.1` with Buildx, and NVIDIA Container Toolkit `1.19.1`.
209+
It reuses exact versions, installs missing pinned packages, permits only the reviewed `dkms` transition from `3.0.11-1ubuntu13` to `1:3.4.0-1ubuntu1`, and refuses every other mismatched version.
205210
It requires the packaged NVIDIA CDI refresh service to advertise `nvidia.com/gpu=all`, and verifies CDI and `--gpus all` with real container launches.
206211
If the packaged refresh fails or omits that device, preparation prints service diagnostics and stops for administrator repair instead of generating CDI configuration directly.
207212
If NVIDIA Docker runtime registration or a post-change acceptance probe fails, preparation restores the prior Docker daemon configuration.
208213
Changing pinned packages exits with status `10` for a reboot; after you sign in and run the printed exact-commit command, the accepted express recipe resumes without another prompt.
209-
This automation does not change Station's Deferred support status; physical end-to-end validation remains open.
214+
On stock DGX OS, the path instead validates the existing GB300 driver and zero-error volatile ECC state, local Docker and Buildx, `nvidia-ctk` and CDI, and `nvidia-smi` device visibility through both CDI and `--gpus all`.
215+
It does not install or replace host packages, enable or restart services, generate CDI configuration, or rewrite the Docker runtime; aside from the private preparation log and separately disclosed `docker`-group membership grant when needed, only the digest-pinned visibility-test image can remain cached.
216+
The visibility probes do not replace physical acceptance of the pinned managed-vLLM and Nemotron Ultra recipe.
217+
This automation does not change Station's Deferred support status; product validation remains pending.
210218
To select DeepSeek V4 Flash while retaining the one-confirmation Station express flow, run `curl -fsSL https://www.nvidia.com/nemoclaw.sh | bash -s -- --station-deepseek`.
211219
The `--station-deepseek` flag requires an interactive terminal; in a `curl | bash` pipeline, `/dev/tty` must be available.
212220
Without terminal access, the installer stops before it installs Docker or build dependencies instead of ignoring the flag.
@@ -229,7 +237,7 @@ Use these details when your first-run path needs more control.
229237

230238
<Warning>
231239
Express install automates the configuration but does not change DGX Station's Deferred support status.
232-
Full NemoClaw onboarding with this recipe has not completed end-to-end validation on physical DGX Station hardware.
240+
Product validation of this express-install boundary is still pending.
233241
</Warning>
234242

235243
The installer auto-launches `nemoclaw onboard` when it can find the new binary.

‎scripts/checks/vitest-project-overlap.ts‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -46,6 +46,7 @@ const INSTALLER_INTEGRATION_TESTS = new Set([
4646
"test/install-openshell-version-check.test.ts",
4747
"test/install-preflight-docker-bootstrap.test.ts",
4848
"test/install-preflight.test.ts",
49+
"test/install-station-dgx-os.test.ts",
4950
"test/install-station-host-preparation.test.ts",
5051
]);
5152

‎scripts/install.sh‎

Lines changed: 42 additions & 11 deletions
Original file line numberDiff line numberDiff line change
@@ -2928,8 +2928,21 @@ is_wsl_host() {
29282928
# and Windows WSL from the host environment. Echoes "DGX Spark",
29292929
# "DGX Station", "Windows WSL", or empty. Used to gate the express install
29302930
# prompt; only platforms with a known sensible default are offered.
2931+
is_station_gb300_product() {
2932+
local product=${1:-}
2933+
[[ "$product" =~ (^|[^[:alnum:]])[Pp]3830([^[:alnum:]]|$) ]] \
2934+
|| [[ "$product" == *[Ss][Tt][Aa][Tt][Ii][Oo][Nn]* &&
2935+
"$product" == *[Gg][Bb]300* ]]
2936+
}
2937+
2938+
classify_dgx_station_release() {
2939+
local helper="${SCRIPT_DIR}/prepare-dgx-station-host.sh"
2940+
[[ -f "$helper" ]] || error "DGX Station host preparation helper is missing: ${helper}"
2941+
bash "$helper" --classify-dgx-release
2942+
}
2943+
29312944
detect_express_platform() {
2932-
local model=""
2945+
local model="" release_state=""
29332946
if is_wsl_host; then
29342947
printf "Windows WSL"
29352948
return
@@ -2941,14 +2954,20 @@ detect_express_platform() {
29412954
model="$(tr -d '\0' </sys/firmware/devicetree/base/model 2>/dev/null || true)"
29422955
fi
29432956
case "$model" in
2944-
*DGX*Spark*) printf "DGX Spark" ;;
2945-
*Station*GB300*)
2946-
if [ -e /etc/dgx-release ] || [ -L /etc/dgx-release ]; then
2947-
printf "Unsupported DGX Station OS"
2948-
else
2949-
printf "DGX Station"
2950-
fi
2957+
*DGX*Spark*)
2958+
printf "DGX Spark"
2959+
return
29512960
;;
2961+
esac
2962+
if is_station_gb300_product "$model"; then
2963+
release_state="$(classify_dgx_station_release)"
2964+
case "$release_state" in
2965+
generic-ubuntu | supported-dgx-os) printf "DGX Station" ;;
2966+
*) printf "Unsupported DGX Station OS" ;;
2967+
esac
2968+
return
2969+
fi
2970+
case "$model" in
29522971
*DGX*Station*) printf "Unsupported DGX Station generation" ;;
29532972
*) ;;
29542973
esac
@@ -2958,7 +2977,7 @@ validate_express_platform_boundary() {
29582977
case "${1:-}" in
29592978
"Unsupported DGX Station OS")
29602979
if [ "${NEMOCLAW_NO_EXPRESS:-}" = "1" ] || [ -n "${NEMOCLAW_PROVIDER:-}" ]; then return 0; fi
2961-
error "DGX OS/BaseOS is outside the validated Station express boundary. Use the generic Ubuntu 24.04 ARM64 image."
2980+
error "This DGX Station OS image is outside the validated Station express boundary. Use generic Ubuntu 24.04 ARM64 or stock DGX OS 7.2.0, 7.4.0, or 7.5.0 on Station GB300."
29622981
;;
29632982
"Unsupported DGX Station generation")
29642983
if [ "${NEMOCLAW_NO_EXPRESS:-}" = "1" ] || [ -n "${NEMOCLAW_PROVIDER:-}" ]; then return 0; fi
@@ -3222,7 +3241,11 @@ run_station_host_preparation() {
32223241
ensure_station_express_host() {
32233242
[[ "${_SELECTED_EXPRESS_PLATFORM:-}" == "DGX Station" ]] || return 0
32243243

3225-
info "Checking pinned DGX Station host prerequisites. Exact matches are reused."
3244+
if [[ "$(classify_dgx_station_release)" == "supported-dgx-os" ]]; then
3245+
info "Validating stock DGX OS GPU device visibility and the local container runtime contract. Host packages and runtime configuration will not be changed."
3246+
else
3247+
info "Checking pinned DGX Station host prerequisites. Exact matches are reused."
3248+
fi
32263249
local status=0
32273250
run_station_host_preparation || status=$?
32283251
case "$status" in
@@ -3250,6 +3273,10 @@ prepare_installer_host() {
32503273
# Intentional ordering: Station preparation owns the reboot boundary before
32513274
# generic Docker bootstrap; ensure_station_express_host is a no-op elsewhere.
32523275
ensure_station_express_host
3276+
if [[ "${_SELECTED_EXPRESS_PLATFORM:-}" == "DGX Station" ]]; then
3277+
unset DOCKER_HOST
3278+
export DOCKER_CONTEXT=default
3279+
fi
32533280
ensure_docker
32543281
ensure_openshell_build_deps
32553282
}
@@ -3287,7 +3314,11 @@ describe_express_install() {
32873314
inference_summary="managed local vLLM with NVIDIA Nemotron 3 Ultra 550B"
32883315
inference_disclosure="Managed vLLM pulls the pinned Station image and approximately 352 GB model, then runs a local inference container."
32893316
fi
3290-
printf " Station host setup reuses exact prerequisite versions, applies the reviewed factory DKMS transition when present, installs missing pinned driver, Docker, and NVIDIA Container Toolkit packages, and may require one reboot.\n"
3317+
if [[ "$(classify_dgx_station_release)" == "supported-dgx-os" ]]; then
3318+
printf " Stock DGX OS setup reuses the factory driver and container stack after local GPU device-visibility, CDI, Docker, and Buildx validation. It does not install or replace host packages or rewrite the Docker runtime.\n"
3319+
else
3320+
printf " Station host setup reuses exact prerequisite versions, applies the reviewed factory DKMS transition when present, installs missing pinned driver, Docker, and NVIDIA Container Toolkit packages, and may require one reboot.\n"
3321+
fi
32913322
printf " Host setup may add this trusted local account to the docker group, which grants root-equivalent control. This flow is only for trusted single-user development hosts; shared or managed hosts require an organization-approved Docker access path.\n"
32923323
printf " DGX Station remains Deferred; this recipe has not completed end-to-end validation on physical hardware.\n"
32933324
sandbox_summary="${NEMOCLAW_SANDBOX_NAME:-my-assistant}"

0 commit comments

Comments
 (0)