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
44 lines
1.9 KiB
YAML
44 lines
1.9 KiB
YAML
# One machine that becomes a mesh and is then assigned schedtest — the bed that proves the
|
|
# scheduled-container primitive (novox/hq ADR 0053) end to end.
|
|
#
|
|
# schedtest is the thinnest possible carrier of the ADR's claim: no service, no listener, no
|
|
# provisioner, no runtime, no tools, no events — just a `container` marked `schedule: "* * * * *"`
|
|
# that appends a timestamp line to a file in a mounted data directory each time the host fires it.
|
|
# It is the recurring twin of the run-once bed (catalogue-mqtt), with the gating rule deliberately
|
|
# reversed: installing the schedule must NOT run the container and must NOT gate convergence — the
|
|
# node reports current as soon as the schedule is state that is present, and the clock (not the
|
|
# apply) fires the runs thereafter.
|
|
#
|
|
# The sharp points this bed proves, straight from ADR 0053 "How each claim is checked":
|
|
# - installing it does not run it, and the node is current WITHOUT a run;
|
|
# - the container fires when the cron is due (top of the next minute);
|
|
# - it fires AGAIN on the following minute — recurrence, not a one-shot.
|
|
#
|
|
# MESH_LAB_HOST_BINARY=.../mesh-host MESH_LAB_MODULES=.../mesh-control/examples/modules
|
|
# alpine:latest must be in the local daemon; the machine pulls it from the internet over its
|
|
# uplink, and the scheduled container declares it exactly as the catalogue writes it.
|
|
# There is no runtime image: schedtest carries no code of its own — the scheduled container is a
|
|
# bare alpine that runs `date >> /data/runs.log` and exits.
|
|
scenario: schedule-tick
|
|
|
|
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: 3GiB
|
|
cpus: 2
|
|
|
|
images:
|
|
- mesh-control:development
|
|
|
|
place:
|
|
# Only the host — schedtest has no mesh-runtime to place. The tick image is pulled from the
|
|
# internet by the host over its uplink, not placed inside the machine.
|
|
all: [host]
|