Add unikernel runtime: run services as Nanos VMs under QEMU/KVM

Implements the hypervisor design's Phase 1: a second runtime.Runtime
backend (QEMU) that runs each service component as a Nanos unikernel VM
instead of a podman container, selected per-component via a new
runtime = "unikernel" service-def field.

- internal/runtime/qemu.go: QEMURuntime. Pull extracts the ELF from the
  OCI image; Run does `ops build` + boots qemu-system-x86_64 with KVM,
  user-mode net port-forwards, QMP control socket and serial console log;
  Stop/Remove/Inspect/List/Logs map onto VM lifecycle + state dir.
- proto/registry/servicedef: add runtime, memory_mb, vcpus fields
  (registry migration 5).
- agent: holds both runtimes; runtimeFor() selects per component;
  listAllContainers() merges containers + VMs so drift/status see both.
  Unikernel runtime auto-enables on nodes with /dev/kvm + ops.

Validated end-to-end on straylight: a test service deploys via
`mcp deploy --direct`, boots as a Nanos unikernel, serves HTTP through
the agent port-forward, and reports running via `mcp status`/`mcp logs`.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Kyle Isom
2026-06-11 00:54:49 -07:00
parent 3b08caaa0a
commit d56f224359
30 changed files with 949 additions and 152 deletions

View File

@@ -51,7 +51,7 @@ func (a *Agent) StopService(ctx context.Context, req *mcpv1.StopServiceRequest)
}
}
if err := a.Runtime.Stop(ctx, containerName); err != nil {
if err := a.runtimeFor(c.Runtime).Stop(ctx, containerName); err != nil {
a.Logger.Info("stop container (ignored)", "container", containerName, "error", err)
}
@@ -132,12 +132,14 @@ func startComponent(ctx context.Context, a *Agent, service string, c *registry.C
containerName := ContainerNameFor(service, c.Name)
r := &mcpv1.ComponentResult{Name: c.Name, Success: true}
rt := a.runtimeFor(c.Runtime)
// Remove any pre-existing container; ignore errors for non-existent ones.
_ = a.Runtime.Stop(ctx, containerName)
_ = a.Runtime.Remove(ctx, containerName)
_ = rt.Stop(ctx, containerName)
_ = rt.Remove(ctx, containerName)
spec := componentToSpec(service, c)
if err := a.Runtime.Run(ctx, spec); err != nil {
if err := rt.Run(ctx, spec); err != nil {
r.Success = false
r.Error = fmt.Sprintf("run container: %v", err)
return r
@@ -156,11 +158,12 @@ func restartComponent(ctx context.Context, a *Agent, service string, c *registry
containerName := ContainerNameFor(service, c.Name)
r := &mcpv1.ComponentResult{Name: c.Name, Success: true}
_ = a.Runtime.Stop(ctx, containerName)
_ = a.Runtime.Remove(ctx, containerName)
rt := a.runtimeFor(c.Runtime)
_ = rt.Stop(ctx, containerName)
_ = rt.Remove(ctx, containerName)
spec := componentToSpec(service, c)
if err := a.Runtime.Run(ctx, spec); err != nil {
if err := rt.Run(ctx, spec); err != nil {
r.Success = false
r.Error = fmt.Sprintf("run container: %v", err)
_ = registry.UpdateComponentState(a.DB, service, c.Name, "", "stopped")
@@ -177,14 +180,16 @@ func restartComponent(ctx context.Context, a *Agent, service string, c *registry
// componentToSpec builds a runtime.ContainerSpec from a registry Component.
func componentToSpec(service string, c *registry.Component) runtime.ContainerSpec {
return runtime.ContainerSpec{
Name: ContainerNameFor(service, c.Name),
Image: c.Image,
Network: c.Network,
User: c.UserSpec,
Restart: c.Restart,
Ports: c.Ports,
Volumes: c.Volumes,
Cmd: c.Cmd,
Name: ContainerNameFor(service, c.Name),
Image: c.Image,
Network: c.Network,
User: c.UserSpec,
Restart: c.Restart,
Ports: c.Ports,
Volumes: c.Volumes,
Cmd: c.Cmd,
MemoryMB: c.MemoryMB,
VCPUs: c.VCPUs,
}
}