Enum ServeProcessAllocation 

pub enum ServeProcessAllocation {
    ModelRank {

        process: String,
        role: String,
        role_kind: ServeRoleKind,
        replica: u32,
        rank: u32,
        rank_count: u32,
        machine: String,
        devices: Vec<u32>,
        model_locator: String,
        auxiliary_model_locators: Vec<AuxiliaryModelLocator>,
        endpoint: Option<EndpointAssignment>,
        ports: BTreeMap<String, EndpointAssignment>,
        cache: String,
        capture_storage: Option<String>,
        launch: AllocationLaunch,
        effective_settings: BTreeMap<String, SettingValue>,
        effective_parallelism: Parallelism,
        links: Vec<ServeRoleLink>,
        dependencies: Vec<String>,
        discovery: Option<String>,
        render_inputs: Vec<SuppliedRenderInput>,
    
},
    Frontend {

        process: String,
        process_role: FrontendProcessRole,
        components: FrontendComponents,
        machine: String,
        devices: Vec<u32>,
        endpoint: EndpointAssignment,
        ports: BTreeMap<String, EndpointAssignment>,
        cache: String,
        launch: AllocationLaunch,
        gateway: Box<GatewayPlan>,
        pd_router: Option<Box<PdRouterPlan>>,
        links: Vec<ServeRoleLink>,
        dependencies: Vec<String>,
        discovery: Option<String>,
        render_inputs: Vec<SuppliedRenderInput>,
    
},
    Discovery {
        process: String,
        process_role: DiscoveryProcessRole,
        components: DiscoveryBinding,
        machine: String,
        devices: Vec<u32>,
        endpoint: EndpointAssignment,
        ports: BTreeMap<String, EndpointAssignment>,
        cache: String,
        data_directory: String,
        launch: AllocationLaunch,
        discovery: Box<DiscoveryRequirement>,
        render_inputs: Vec<SuppliedRenderInput>,
    },
}

One concrete process allocation supplied to RenderServe. Model-rank and process-only frontend identities are distinct so a frontend cannot acquire model coordinates or a model locator by construction.

Variants

ModelRank

Fields

process: String

role: String

role_kind: ServeRoleKind

replica: u32

rank: u32

rank_count: u32

machine: String

devices: Vec<u32>

model_locator: String

auxiliary_model_locators: Vec<AuxiliaryModelLocator>

The machine-resolved auxiliary weight locators for each declared kind ([[RFC-0003:C-SERVE-AUXILIARY-MODELS]]).

endpoint: Option<EndpointAssignment>

ports: BTreeMap<String, EndpointAssignment>

cache: String

capture_storage: Option<String>

The control-plane-assigned persistent trace directory; present only when this rank belongs to an engine-trace capture target ([[RFC-0004:C-WORKLOAD-PROFILING]]).

launch: AllocationLaunch

effective_settings: BTreeMap<String, SettingValue>

effective_parallelism: Parallelism

links: Vec<ServeRoleLink>

dependencies: Vec<String>

discovery: Option<String>

The discovery process this process registers with, when the topology has one ([[RFC-0003:C-SERVE-TOPOLOGY]]).

render_inputs: Vec<SuppliedRenderInput>

Frontend

Fields

process: String

process_role: FrontendProcessRole

components: FrontendComponents

machine: String

devices: Vec<u32>

endpoint: EndpointAssignment

ports: BTreeMap<String, EndpointAssignment>

cache: String

launch: AllocationLaunch

gateway: Box<GatewayPlan>

pd_router: Option<Box<PdRouterPlan>>

links: Vec<ServeRoleLink>

dependencies: Vec<String>

discovery: Option<String>

The discovery process this process registers with, when the topology has one ([[RFC-0003:C-SERVE-TOPOLOGY]]).

render_inputs: Vec<SuppliedRenderInput>

Discovery

The discovery service a Gateway result requires ([[RFC-0003:C-SERVE-TOPOLOGY]]): process-only, zero devices, placed on the Gateway’s machine, with an internal (not public) endpoint.

Fields

process: String

process_role: DiscoveryProcessRole

components: DiscoveryBinding

machine: String

devices: Vec<u32>

endpoint: EndpointAssignment

ports: BTreeMap<String, EndpointAssignment>

cache: String

data_directory: String

The control-plane-assigned empty data directory under the machine’s runtime cache.

launch: AllocationLaunch

discovery: Box<DiscoveryRequirement>

render_inputs: Vec<SuppliedRenderInput>

Trait Implementations

impl Clone for ServeProcessAllocation

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ServeProcessAllocation

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for ServeProcessAllocation

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl JsonSchema for ServeProcessAllocation

fn schema_name() -> Cow<'static, str>

The name of the generated JSON Schema.

fn schema_id() -> Cow<'static, str>

Returns a string that uniquely identifies the schema produced by this type.

fn json_schema(generator: &mut SchemaGenerator) -> Schema

Generates a JSON Schema for this type.

fn inline_schema() -> bool

Whether JSON Schemas generated for this type should be included directly in parent schemas, rather than being re-used where possible using the $ref keyword.

impl PartialEq for ServeProcessAllocation

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for ServeProcessAllocation

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for ServeProcessAllocation