Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
21 changes: 0 additions & 21 deletions api/v1alpha1/groupversion_info.go
Original file line number Diff line number Diff line change
Expand Up @@ -8,15 +8,6 @@
// NodeClaim - one provisioned external instance and its lifecycle. Owns the
// terminate finalizer so a paid instance is never leaked.
//
// On top of that provisioning core sit the workload types, each synthesizing
// Pods onto the same placement path rather than bypassing it:
//
// Sandbox - one interactive remote box (agent workspace, shell, scratch GPU),
// reachable with the same kubectl exec/logs as a local Pod.
// SandboxSet - maintains N Sandboxes, and owns /scale so `kubectl scale` and HPA
// drive the count. Keeping boxes ready ahead of demand is a USE of
// this, not its definition — there are no lease semantics here.
//
// +kubebuilder:object:generate=true
// +groupName=nebula.inftyai.com
package v1alpha1
Expand Down Expand Up @@ -75,18 +66,6 @@ const (
// value is the NodePool name, so the key mirrors the CRD kind.
PoolLabel = "nebula.inftyai.com/nodepool"

// SandboxLabel records which Sandbox a Pod belongs to. Its value is the Sandbox
// name, so the key mirrors the CRD kind. The Sandbox controller selects its own
// Pod by it, and it is what makes `kubectl get pods -l
// nebula.inftyai.com/sandbox=alice` work.
SandboxLabel = "nebula.inftyai.com/sandbox"

// SandboxSetLabel records which SandboxSet created a Sandbox. Its value is the
// set name. It is the selector the set's /scale subresource publishes in status
// (so HPA can find the set's members) and how the set controller enumerates the
// boxes it owns — ownerReferences alone would not support a label-selector query.
SandboxSetLabel = "nebula.inftyai.com/sandboxset"

// AcceleratorTypeLabel carries the accelerator TYPE only (e.g. "a100-40gb",
// "h100"). The COUNT is a standard container resource request/limit
// (nvidia.com/gpu today), so scheduling fit and provisioning read the same number
Expand Down
242 changes: 0 additions & 242 deletions api/v1alpha1/sandbox_types.go

This file was deleted.

Loading
Loading