Merge branch 'main' into feat/platform-and-nvl-records

Author: jayzhudev

AGENTS.md

@@ -32,7 +32,7 @@ ncx-infra-controller-core/
 │                        # [workspace] members list in `Cargo.toml` — each
 │                        # crate's own `Cargo.toml` has a `description` field.
 │                        # Note: the directory name does NOT always equal the
-│                        # crate name (e.g. crates/api/ → crate carbide-api).
+│                        # crate name (e.g. crates/api/ → crate nico-api).
 │                        # Use `grep '^name =' crates/<dir>/Cargo.toml | head -1`
 │                        # to get the actual crate name before running
 │                        # `cargo test -p <name>` or similar.
@@ -42,7 +42,7 @@ ncx-infra-controller-core/
 ├── helm/                # Helm chart for Kubernetes deployment
 ├── bluefield/           # BlueField DPU-specific components
 ├── pxe/                 # PXE boot artifact generation
-├── lints/               # Custom Clippy lints (carbide-lints crate)
+├── lints/               # Custom Clippy lints (nico-lints crate)
 ├── include/             # Shared Makefile fragments
 ├── .github/             # GitHub Actions workflows and templates
 ├── Cargo.toml           # Workspace dependency management
@@ -104,7 +104,7 @@ cargo make pre-commit-verify-workspace
 
 # Individual checks:
 cargo make clippy              # Clippy linter (warnings = errors)
-cargo make carbide-lints       # Custom carbide lints (requires nightly setup)
+cargo make nico-lints       # Custom nico lints (requires nightly setup)
 cargo make check-format-flow   # Check rustfmt formatting
 cargo make check-format-nightly # Check import grouping/sorting (requires nightly)
 cargo make check-workspace-deps # Validate dependency declarations in Cargo.toml
@@ -117,7 +117,7 @@ cargo make format-nightly      # Also sort imports
 ```
 
 > **Note:** The nightly toolchain is used only for `check-format-nightly` and
-> `carbide-lints`. The stable toolchain pinned in `rust-toolchain.toml` is used
+> `nico-lints`. The stable toolchain pinned in `rust-toolchain.toml` is used
 > for everything else.
 
 ### Top-level Makefile (rest-api entrypoint)

Cargo.lock

@@ -20,8 +20,8 @@ of the bare-metal lifecycle to fast-track building next generation AI Cloud offe
 | Layer | What it installs | Helm release |
 |-------|-----------------|--------------|
 | **Common services** | MetalLB, cert-manager, Vault, external-secrets, PostgreSQL | via `helmfile` in `helm-prereqs/` |
-| **Carbide Core** | NVIDIA Infra Controller (this repo's `helm/` chart) | `carbide` in `forge-system` |
-| **Carbide REST** | NVIDIA Infra Controller's REST API, Temporal, Keycloak, site-agent | `carbide-rest` + `carbide-rest-site-agent` in `carbide-rest` |
+| **NICo Core** | NVIDIA Infra Controller (this repo's `helm/` chart) | `nico` in `nico-system` |
+| **NICo REST** | NVIDIA Infra Controller's REST API, Temporal, Keycloak, site-agent | `nico-rest` + `nico-rest-site-agent` in `nico-rest` |
 
 ### Prerequisites
 
@@ -33,20 +33,20 @@ of the bare-metal lifecycle to fast-track building next generation AI Cloud offe
 
 ```bash
 # 1. Build and push images to your registry
-#    Carbide Core image: <your-registry>/nvmetal-carbide:<tag>  (this repo)
-#    Carbide REST images: <your-registry>/carbide-rest-api:<tag>, etc.  (infra-controller-rest)
+#    NICo Core image: <your-registry>/nvmetal-nico:<tag>  (this repo)
+#    NICo REST images: <your-registry>/nico-rest-api:<tag>, etc.  (infra-controller-rest)
 
 # 2. Set environment variables
 export KUBECONFIG=/path/to/kubeconfig
 export REGISTRY_PULL_SECRET=<your-registry-pull-secret-or-ngc-api-key>
 export NCX_IMAGE_REGISTRY=<your-registry>        # e.g. my-registry.example.com/infra-controller
-export NCX_CORE_IMAGE_TAG=<carbide-core-tag>     # e.g. v2025.12.30
-export NCX_REST_IMAGE_TAG=<carbide-rest-tag>     # e.g. v1.0.4
+export NCX_CORE_IMAGE_TAG=<nico-core-tag>     # e.g. v2025.12.30
+export NCX_REST_IMAGE_TAG=<nico-rest-tag>     # e.g. v1.0.4
 
 # 3. Customize site-specific values
 #    Edit helm-prereqs/values/ncx-core.yaml:
-#      carbide-api.hostname      — your site's external API hostname
-#      carbide-api.siteConfig    — network pools, VLAN ranges, IB config, MetalLB VIPs
+#      nico-api.hostname      — your site's external API hostname
+#      nico-api.siteConfig    — network pools, VLAN ranges, IB config, MetalLB VIPs
 #    Edit helm-prereqs/values/metallb-config.yaml:
 #      IPAddressPool, BGPPeer    — your site's VIP ranges and TOR switch config
 #    Edit helm-prereqs/values.yaml:
@@ -55,7 +55,7 @@ export NCX_REST_IMAGE_TAG=<carbide-rest-tag>     # e.g. v1.0.4
 # 4. Point NCX_REPO at infra-controller-rest (auto-detected if a sibling directory)
 export NCX_REPO=/path/to/infra-controller-rest   # optional
 
-# 5. Run setup — installs common services, Carbide Core, and Carbide REST in order
+# 5. Run setup — installs common services, NICo Core, and NICo REST in order
 cd helm-prereqs
 ./setup.sh        # interactive
 ./setup.sh -y     # non-interactive (CI/CD)

README.md

@@ -129,18 +129,18 @@ using interpolation if it makes sense.
 
 ### Core API handler Errors
 
-Inside API handlers, the `CarbideError` data type should be used to construct errors. It should then be converted into
-`tonic::Status` using `.into()`. All errors being derived from `CarbideError` assures that the errors will look uniform
+Inside API handlers, the `NicoError` data type should be used to construct errors. It should then be converted into
+`tonic::Status` using `.into()`. All errors being derived from `NicoError` assures that the errors will look uniform
 to tenants.
 
-The `CarbideError` variant that is used should be selected based on whether the error gets returned due to the user
+The `NicoError` variant that is used should be selected based on whether the error gets returned due to the user
 passing invalid arguments or due to the system not being able to handle the request correctly. Error variants that
 should be used if the user passing invalid arguments can be `InvalidArgument`, `InvalidConfiguration`, `NotFoundError`
 or `ConcurrentModificationError` - these will map to "4xx-like" gRPC error codes. An example of a system-side error
-would be `CarbideError::Internal`.
+would be `NicoError::Internal`.
 
 ```rust
-// Avoid — constructing Status directly, bypassing `CarbideError` error mapping
+// Avoid — constructing Status directly, bypassing `NicoError` error mapping
 pub async fn create_resource(
     api: &Api,
     request: Request<rpc::Resource>,
@@ -151,15 +151,15 @@ pub async fn create_resource(
         .ok_or_else(|| Status::invalid_argument("id is required"))?;
 }
 
-// Prefer — uses `CarbideError::InvalidArgument`
+// Prefer — uses `NicoError::InvalidArgument`
 pub async fn create_resource(
     api: &Api,
     request: Request<rpc::Resource>,
 ) -> Result<Response<()>, Status> {
     let resource = request.into_inner();
     let id = resource
         .id
-        .ok_or(CarbideError::InvalidArgument("id is required".into()))?;
+        .ok_or(NicoError::InvalidArgument("id is required".into()))?;
 }
 ```
 
@@ -172,14 +172,14 @@ checks for each meaningful combination of feature flags we support, which scales
 
 Cases where features *are* warranted:
 
-- For shared crates when only a subset of dependents need certain code: For example, the `carbide_uuid` is used by
-  several dependents, but only the `carbide_api` crate needs the sqlx conversions. We don't want e.g.
-  `carbide_admin_cli` to take a dependency on `sqlx`, so the sqlx conversions are behind a `sqlx` crate feature. But
+- For shared crates when only a subset of dependents need certain code: For example, the `nico_uuid` is used by
+  several dependents, but only the `nico_api` crate needs the sqlx conversions. We don't want e.g.
+  `nico_admin_cli` to take a dependency on `sqlx`, so the sqlx conversions are behind a `sqlx` crate feature. But
   this is covered by CI tests, since CI builds both the admin-cli and the api crate, both sets of features are
   exercised.
 
-- For supporting non-linux builds: The `carbide_api` crate needs to use types from the `tss-esapi` crate to support
-  validating secure-boot keys, but `tss-esapi` only builds on Linux. To support developers running `carbide_api` on
+- For supporting non-linux builds: The `nico_api` crate needs to use types from the `tss-esapi` crate to support
+  validating secure-boot keys, but `tss-esapi` only builds on Linux. To support developers running `nico_api` on
   their Mac for testing, the parts which require `tss-esapi` are carefully carved out into a `linux-build` feature
   (which is enabled by default). We do not run CI tests with this feature disabled, so supporting a build without
   `linux-build` enabled is best-effort.
@@ -217,10 +217,10 @@ Avoid spawning background tasks without joining them. Any panics that happen in
 the rest of the process unless you join them via `JoinHandle::join()` or add them to a `JoinSet` which is later awaited
 with `JoinSet::join_all()`.
 
-For carbide-api, we use a single `JoinSet` to spawn all background tasks, and call `join_all()` to block "forever" until
+For nico-api, we use a single `JoinSet` to spawn all background tasks, and call `join_all()` to block "forever" until
 the process is shut down. This makes it so any panics in the JoinSet will propagate to the main task, and crash the
 process (which is what we want.) If you want to spawn background work, prefer accepting a `&mut JoinSet` and spawn your
-background task into it. Your task can be constructed it inside `carbide::setup::initialize_and_spawn_controllers`,
+background task into it. Your task can be constructed it inside `nico::setup::initialize_and_spawn_controllers`,
 which has a JoinSet it can pass to your `start()` function.
 
 Avoid using `oneshot::Sender<()>` as a cancellation signal, and prefer tokio_util's `CancellationToken`, which can
@@ -286,7 +286,7 @@ impl ClientlessBackgroundJob {
 ```
 
 Avoid mixing the approaches and returning an RAII handle for "client-less" background tasks, if it only exists to stop
-the task when dropped. In carbide-api, there are many such client-less background jobs, and storing each of their
+the task when dropped. In nico-api, there are many such client-less background jobs, and storing each of their
 handles for the correct lifetime is awkward and error-prone. Propagating a single top-level CancellationToken to each of
 them is the preferred approach.
 

STYLE_GUIDE.md

@@ -17,8 +17,8 @@ The punt stats receiver generates metrics like the following:
 ### Punt Stats
 
 ```
-punt_stats_bytes_total{component="punt_stats",dropped="false",host_name="10-217-170-242.local.forge",protocol="dhcp"} 862206
-punt_stats_bytes_total{component="punt_stats",dropped="true",host_name="10-217-170-242.local.forge",protocol="dhcp"} 0
-punt_stats_packets_total{component="punt_stats",dropped="false",host_name="10-217-170-242.local.forge",protocol="dhcp"} 2686
-punt_stats_packets_total{component="punt_stats",dropped="true",host_name="10-217-170-242.local.forge",protocol="dhcp"} 0
+punt_stats_bytes_total{component="punt_stats",dropped="false",host_name="10-217-170-242.local.nico",protocol="dhcp"} 862206
+punt_stats_bytes_total{component="punt_stats",dropped="true",host_name="10-217-170-242.local.nico",protocol="dhcp"} 0
+punt_stats_packets_total{component="punt_stats",dropped="false",host_name="10-217-170-242.local.nico",protocol="dhcp"} 2686
+punt_stats_packets_total{component="punt_stats",dropped="true",host_name="10-217-170-242.local.nico",protocol="dhcp"} 0
 ```

bluefield/otel/puntstatsreceiver/README.md

@@ -1,7 +1,7 @@
 # Admin CLI Development Guide
 
 This guide covers how to develop new subcommands, and work on existing
-subcommands, in the `admin-cli` crate (for `carbide-admin-cli`).
+subcommands, in the `admin-cli` crate (for `nico-admin-cli`).
 
 ## Table of Contents
 
@@ -78,14 +78,14 @@ handler in `cmd.rs`.
 pub mod args;
 pub mod cmd;
 
-use crate::errors::CarbideCliResult;
+use crate::errors::NicoCliResult;
 pub use args::Args;
 
 use crate::cfg::run::Run;
 use crate::cfg::runtime::RuntimeContext;
 
 impl Run for Args {
-    async fn run(self, ctx: &mut RuntimeContext) -> CarbideCliResult<()> {
+    async fn run(self, ctx: &mut RuntimeContext) -> NicoCliResult<()> {
         cmd::show(&ctx.api_client, self, ctx.config.format).await
     }
 }
@@ -114,7 +114,7 @@ Contains the actual command handler. Receives parsed arguments and
 only the specific dependencies it needs (not the full RuntimeContext):
 
 ```rust
-use ::rpc::admin_cli::{CarbideCliResult, OutputFormat};
+use ::rpc::admin_cli::{NicoCliResult, OutputFormat};
 
 use super::args::Args;
 use crate::rpc::ApiClient;
@@ -123,7 +123,7 @@ pub async fn show(
     api_client: &ApiClient,
     args: Args,
     format: OutputFormat,
-) -> CarbideCliResult<()> {
+) -> NicoCliResult<()> {
     // Implementation here.
     Ok(())
 }
@@ -146,7 +146,7 @@ pub(crate) trait Dispatch {
     fn dispatch(
         self,
         ctx: RuntimeContext,
-    ) -> impl std::future::Future<Output = CarbideCliResult<()>>;
+    ) -> impl std::future::Future<Output = NicoCliResult<()>>;
 }
 ```
 
@@ -161,7 +161,7 @@ pub(crate) trait Run {
     fn run(
         self,
         ctx: &mut RuntimeContext,
-    ) -> impl std::future::Future<Output = CarbideCliResult<()>>;
+    ) -> impl std::future::Future<Output = NicoCliResult<()>>;
 }
 ```
 
@@ -236,7 +236,7 @@ those specific values to `cmd.rs`:
 
 ```rust
 impl Run for Args {
-    async fn run(self, ctx: &mut RuntimeContext) -> CarbideCliResult<()> {
+    async fn run(self, ctx: &mut RuntimeContext) -> NicoCliResult<()> {
         // Simple handler -- just needs API client.
         cmd::show(&ctx.api_client, self).await
 
@@ -289,12 +289,12 @@ Create `src/my_command/show/cmd.rs`:
  * ..etc etc.
  */
 
-use crate::errors::CarbideCliResult;
+use crate::errors::NicoCliResult;
 
 use super::args::Args;
 use crate::rpc::ApiClient;
 
-pub async fn show(api_client: &ApiClient, args: Args) -> CarbideCliResult<()> {
+pub async fn show(api_client: &ApiClient, args: Args) -> NicoCliResult<()> {
     // Your implementation here.
     Ok(())
 }
@@ -312,14 +312,14 @@ Create `src/my_command/show/mod.rs`:
 pub mod args;
 pub mod cmd;
 
-use crate::errors::CarbideCliResult;
+use crate::errors::NicoCliResult;
 pub use args::Args;
 
 use crate::cfg::run::Run;
 use crate::cfg::runtime::RuntimeContext;
 
 impl Run for Args {
-    async fn run(self, ctx: &mut RuntimeContext) -> CarbideCliResult<()> {
+    async fn run(self, ctx: &mut RuntimeContext) -> NicoCliResult<()> {
         cmd::show(&ctx.api_client, self).await
     }
 }

crates/admin-cli/DEVELOPMENT.md

@@ -48,23 +48,23 @@ pub struct CliOptions {
     )]
     pub cloud_unsafe_op: Option<String>,
 
-    #[clap(short, long, env = "CARBIDE_API_URL")]
+    #[clap(short, long, env = "API_URL", visible_alias = "carbide-url")]
     #[clap(
-        help = "Default to CARBIDE_API_URL environment variable or $HOME/.config/carbide_api_cli.json file or https://carbide-api.forge-system.svc.cluster.local:1079."
+        help = "Default to API_URL environment variable or $HOME/.config/carbide_api_cli.json file or https://carbide-api.forge-system.svc.cluster.local:1079."
     )]
-    pub carbide_api: Option<String>,
+    pub api_url: Option<String>,
 
     #[clap(short, long, value_enum, default_value = "ascii-table")]
     pub format: OutputFormat,
 
     #[clap(short, long)]
     pub output: Option<String>,
 
-    #[clap(long, env = "FORGE_ROOT_CA_PATH")]
+    #[clap(long, env = "ROOT_CA_PATH", visible_alias = "forge-root-ca-path")]
     #[clap(
-        help = "Default to FORGE_ROOT_CA_PATH environment variable or $HOME/.config/carbide_api_cli.json file."
+        help = "Default to ROOT_CA_PATH environment variable or $HOME/.config/carbide_api_cli.json file."
     )]
-    pub forge_root_ca_path: Option<String>,
+    pub root_ca_path: Option<String>,
 
     #[clap(long, env = "CLIENT_CERT_PATH")]
     #[clap(

crates/admin-cli/src/cfg/cli_options.rs

@@ -170,6 +170,45 @@ pub enum DeviceTargetArgs {
     Rack(RackTargetArgs),
 }
 
+/// Component-power-control target subset: no rack target since
+/// `ComponentPowerControlRequest` only supports machines, switches and
+/// power shelves.
+#[derive(Subcommand, Debug)]
+pub enum PowerControlTargetArgs {
+    #[clap(about = "Target NVLink switches")]
+    Switch(SwitchTargetArgs),
+
+    #[clap(about = "Target power shelves")]
+    PowerShelf(PowerShelfTargetArgs),
+
+    #[clap(about = "Target compute trays")]
+    ComputeTray(MachineTargetArgs),
+}
+
+#[derive(Copy, Clone, Debug, ValueEnum)]
+#[clap(rename_all = "kebab_case")]
+pub enum PowerActionArg {
+    On,
+    GracefulShutdown,
+    ForceOff,
+    GracefulRestart,
+    ForceRestart,
+    ACPowercycle,
+}
+
+impl From<PowerActionArg> for ::rpc::common::SystemPowerControl {
+    fn from(action: PowerActionArg) -> Self {
+        match action {
+            PowerActionArg::On => Self::On,
+            PowerActionArg::GracefulShutdown => Self::GracefulShutdown,
+            PowerActionArg::ForceOff => Self::ForceOff,
+            PowerActionArg::GracefulRestart => Self::GracefulRestart,
+            PowerActionArg::ForceRestart => Self::ForceRestart,
+            PowerActionArg::ACPowercycle => Self::AcPowercycle,
+        }
+    }
+}
+
 pub fn component_result_status_name(status: i32) -> &'static str {
     match rpc::forge::ComponentManagerStatusCode::try_from(status) {
         Ok(rpc::forge::ComponentManagerStatusCode::Success) => "success",