jschoubben
5c91c0ecd2
Every scenario that places a container runtime gives each of its machines `egress: true` — a node that runs modules pulls images from the internet, which is what a node does. The underlay-only scenarios (bootstrap-single, behind-nat, segmented-and-unforwardable, the-ordinary-shape, two-on-a-segment) stay sealed on purpose: an extra NIC would change the very reachability they are asserting about. `images:` keeps only the mesh's own — 55 third-party entries leave whole-mesh-full alone, and the machine fetches them itself by the digest its module.json already pins. The whole-mesh beds also say per machine which of ours they get: novox the substrate control plane and its own fifteen runtimes, ace its twenty-four, the two workstations one each. That is not a lab economy. An operator's workstation holds the images its own modules need, and giving these two the union would put some thirty gigabytes onto a thirty-gigabyte disk. bootstrap-with-registry.yml is deleted. It existed only to demonstrate the lab's registry, nothing referenced it, and there is nothing left for it to demonstrate. Claude-Session: https://claude.ai/code/session_01LrgweAeERJYBg88c5cKDzF
38 lines
1.5 KiB
YAML
38 lines
1.5 KiB
YAML
# One machine that becomes a mesh, then proves model-access answered by a NODE (novox/hq ADR 0055):
|
|
# a local model server (Ollama) on the node ANSWERS the model-access provision, and a consumer that
|
|
# requires model-access is handed the server's ENDPOINT — base URL + model — instead of a vendor
|
|
# credential. No licence, no key: the consumer points an OpenAI-compatible client at the mesh's own
|
|
# model.
|
|
#
|
|
# The flow the test drives:
|
|
# the ollama provider is assigned (it `provides: [model-access]` at node scope, `serves` its port and
|
|
# model) -> the local-model consumer requires model-access and is answered by the local node ->
|
|
# the mesh fills the consumer's bound facts with ollama's at/port/model -> the consumer's templated
|
|
# openai.env carries OPENAI_BASE_URL=http://<at>:<port>/v1. The test asserts that URL and that a
|
|
# request to it reaches the running model server.
|
|
#
|
|
# MESH_LAB_HOST_BINARY=.../mesh-host MESH_LAB_BUNDLE=.../examples/substrate-first-node.lock
|
|
# Both modules are pure declaration (a server container + a templated file) — no module runtime image
|
|
# is built; the bed only stocks the substrate and the ollama server image.
|
|
scenario: local-model-bed
|
|
|
|
segments:
|
|
hosting:
|
|
kind: public
|
|
cidr: [192.0.2.0/24]
|
|
|
|
machines:
|
|
anchor:
|
|
at: { segment: hosting, address: [192.0.2.10] }
|
|
egress: true
|
|
inbound: allow
|
|
memory: 4GiB
|
|
cpus: 4
|
|
disk: 40GiB
|
|
|
|
images:
|
|
- mesh-control:development
|
|
|
|
place:
|
|
all: [host, runtime]
|