novox/mesh-lab
3
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

whole-mesh rehearsal beds: the real node sets converge on one substrate #18

Merged
jschoubben merged 4 commits from feat/whole-mesh into main 2026-09-08 18:06:19 +00:00
Conversation 0 Commits 4 Files Changed 6
+567
2 changed files with 567 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
Showing only changes of commit 581fd6da77 - Show all commits
scenarios/whole-mesh-novox.yml
+101
View File
@@ -0,0 +1,101 @@
# The whole `novox` server's converted service set, installed together on ONE node behind the mesh
# substrate — the whole-catalogue install the rebuild has never actually run. First stage of a
# whole-mesh rehearsal (novox/hq).
#
# Topology, proven by test/integration/assigned-two-node-db.test.ts: the substrate (store, broker,
# control) rides `anchor` and NOTHING else; ALL of novox's services ride the `novox` node — its own
# postgres provider owns 5432 there, so it cannot co-locate with the substrate store on 5432. Both
# machines sit on one public segment and enrol into the one mesh; an overlay is placed so a
# consumer's binding `at` resolves to novox's private address and every consumer reaches the
# providers co-located with it.
#
# The service SET (novox/hq ADR 0039/0048/0052, all converted in mesh-catalog/modules/):
# providers postgres redis minio mongodb mssql
# consumers keycloak gitea nextcloud umami photos invoicing
# apps portainer verdaccio registry route-proxy mailu
# node-level firewall fail2ban
#
# MESH_LAB_HOST_BINARY=.../mesh-host MESH_LAB_BUNDLE=.../examples/substrate-first-node.lock
# The runtimes are built by scripts/build-module-runtime.sh (one per module that has code) and the
# route-proxy image by scripts/build-route-proxy-image.sh; every server image must be in the local
# daemon to be stocked. The test loads each committed module.json from mesh-catalog and rewrites its
# image references to what this scenario's own registry serves by digest.
scenario: whole-mesh-novox
segments:
hosting:
kind: public
cidr: [192.0.2.0/24]
machines:
# The substrate ONLY: store, broker, control. Nothing else lands here.
anchor:
at: { segment: hosting, address: [192.0.2.10] }
inbound: allow
memory: 4GiB
cpus: 4
disk: 20GiB
# The whole novox service set — ~38 containers (five providers with runtimes, six consumers with
# runtimes, portainer/verdaccio/registry/route-proxy, the nine-container Mailu stack and its
# runtime) plus two node-level modules. mssql alone wants ~2GiB; Mailu, Nextcloud and Keycloak are
# each heavy. Sized well past the two-node-db bed's second node.
novox:
at: { segment: hosting, address: [192.0.2.20] }
inbound: allow
memory: 16GiB
cpus: 8
# ~14GiB of images are pulled from the scenario's own registry by digest, several of them large
# (mssql 1.7GiB, invoicing-api 1.9GiB, nextcloud 1.5GiB, umami/mongo ~0.9GiB), plus writable
# layers and the runtimes. A hundred gigabytes holds the whole set without exhausting the disk
# mid-apply.
disk: 100GiB
images:
# The first-node substrate: store, broker, control. postgres:17-alpine doubles as the postgres
# provider's own service image (and Mailu's internal admin DB).
- postgres:17-alpine
- cloudamqp/lavinmq:latest
- mesh-control:development
# The module server images. Each is stocked under the repository path its module.json names, so the
# test's rewrite (pinned(repositoryFor(image))) finds it.
- redis:7-alpine
- minio/minio:latest
- mongo:7
- mcr.microsoft.com/mssql/server:2022-latest
- quay.io/keycloak/keycloak:mesh
- gitea/gitea:1.22
- nextcloud:stable
- ghcr.io/umami-software/umami:postgresql-latest
- alpine:latest
- portainer/portainer-ce:latest
- verdaccio/verdaccio:6
- registry:2
- registry-api.novox.be/novox/invoicing-app:latest
- registry-api.novox.be/novox/invoicing-api:latest
# The Mailu stack (pulled by digest, tagged :mesh so repositoryFor matches the module.json paths).
- ghcr.io/mailu/unbound:mesh
- ghcr.io/mailu/admin:mesh
- ghcr.io/mailu/dovecot:mesh
- ghcr.io/mailu/postfix:mesh
- ghcr.io/mailu/rspamd:mesh
- ghcr.io/mailu/webmail:mesh
- ghcr.io/mailu/nginx:mesh
# The per-module runtimes (built by scripts/build-module-runtime.sh). registry, route-proxy,
# invoicing, firewall and fail2ban carry no mesh-runtime image; route-proxy ships its own.
- mesh-runtime-postgres:development
- mesh-runtime-redis:development
- mesh-runtime-minio:development
- mesh-runtime-mongodb:development
- mesh-runtime-mssql:development
- mesh-runtime-keycloak:development
- mesh-runtime-gitea:development
- mesh-runtime-nextcloud:development
- mesh-runtime-umami:development
- mesh-runtime-photos:development
- mesh-runtime-portainer:development
- mesh-runtime-verdaccio:development
- mesh-runtime-mailu:development
- mesh-route-proxy:development
place:
all: [host, runtime]
test/integration/whole-mesh-novox.test.ts
+466
View File
@@ -0,0 +1,466 @@
/**
* The whole `novox` server's converted service set, installed together on ONE node behind the
* substrate — the whole-catalogue install this rebuild has never actually run. First stage of a
* whole-mesh rehearsal (novox/hq).
*
* Topology (proven by assigned-two-node-db.test.ts): the substrate (store, broker, control) rides
* `anchor` and nothing else; ALL of novox's services ride the `novox` node. novox's own postgres
* provider owns 5432 there, so it cannot co-locate with the substrate store. An overlay is placed so
* each consumer's binding `at` resolves to novox's private address and reaches the providers
* co-located with it.
*
* The SET (18 modules, all converted in mesh-catalog/modules/):
* providers postgres redis minio mongodb mssql
* consumers keycloak gitea nextcloud umami photos invoicing
* apps portainer verdaccio registry route-proxy mailu
* node-level firewall fail2ban
*
* Each committed module.json is LOADED from mesh-catalog — not hand-written — and its container
* image references are rewritten to what this scenario's own registry serves by digest (the same
* pinned(repositoryFor(image)) rule the two-node-db bed applies by hand). Two things this bed
* discovered about the co-located set are handled at load time and RECORDED as findings:
*
* HOST-PORT COLLISIONS. When the whole set lands on one node with its committed host publishes,
* several servers claim the same host port: nextcloud, invoicing-app and route-proxy all want 80;
* minio and invoicing-api both want 9000; gitea and umami both want 3000. route-proxy is meant to
* FRONT the web apps on 80/443, so the web apps' own host publishes are only for direct access.
* To let the whole set converge, the colliding web/app host publishes are remapped to distinct
* host ports here (container ports unchanged); the provider ports the consumers actually connect to
* (postgres 5432, minio 9000, mongodb 27017, mssql 1433) are left as-is. See REMAP below.
*
* MESH_LAB_HOST_BINARY=.../mesh-host MESH_LAB_BUNDLE=.../examples/substrate-first-node.lock
* scripts/build-module-runtime.sh builds one runtime per module that has code; the route-proxy image
* is built by scripts/build-route-proxy-image.sh; scenarios/whole-mesh-novox.yml stocks them all
* alongside every server image.
*/
import { test, before, after } from "node:test";
import assert from "node:assert/strict";
import { existsSync, readFileSync } from "node:fs";
import { dirname, resolve } from "node:path";
import { loadScenario } from "../../src/declaration/parse.ts";
import { raise } from "../../src/lifecycle/raise.ts";
import { destroy, exec } from "../../src/lifecycle/operate.ts";
import { hostBinaryPath, HOST_PATH } from "../../src/lifecycle/place.ts";
import { labIsUsable, destroyAll } from "./harness.ts";
const capability = await labIsUsable();
const binary = hostBinaryPath();
const bundle = process.env["MESH_LAB_BUNDLE"] ?? "";
const modulesEnv = process.env["MESH_LAB_MODULES"] ?? "";
const skip = !capability.usable
? `lab not usable: ${capability.why}`
: !binary || !existsSync(binary)
? "MESH_LAB_HOST_BINARY is not set to a built mesh-host"
: !bundle || !existsSync(bundle)
? "MESH_LAB_BUNDLE is not set to a substrate bundle (mesh-host examples/)"
: false;
const SCENARIO = "whole-mesh-novox";
const NODE = "novox";
/** Where the committed module.json files live: the mesh-catalog beside mesh-control. */
const catalogDir = process.env["MESH_LAB_CATALOG"]
?? (modulesEnv ? resolve(dirname(dirname(dirname(modulesEnv))), "mesh-catalog", "modules") : "")
?? resolve(process.cwd(), "..", "mesh-catalog", "modules");
/**
* The set, in dependency-reading order (the resolver accepts any order). Each row: the module, and
* the container names it should bring up on the node. Node-level modules (firewall, fail2ban) bring
* up no container — they install a package and run a service, checked separately.
*/
const MODULES: { name: string; containers: string[]; node?: boolean }[] = [
{ name: "postgres", containers: ["postgres", "mesh-postgres"] },
{ name: "redis", containers: ["redis", "mesh-redis"] },
{ name: "minio", containers: ["minio", "mesh-minio"] },
{ name: "mongodb", containers: ["mongo", "mesh-mongodb"] },
{ name: "mssql", containers: ["mssql", "mesh-mssql"] },
{ name: "keycloak", containers: ["keycloak", "mesh-keycloak"] },
{ name: "gitea", containers: ["gitea", "mesh-gitea"] },
{ name: "nextcloud", containers: ["nextcloud", "mesh-nextcloud"] },
{ name: "umami", containers: ["umami", "mesh-umami"] },
{ name: "photos", containers: ["photos", "mesh-photos"] },
{ name: "invoicing", containers: ["invoicing-app", "invoicing-api"] },
{ name: "portainer", containers: ["portainer", "mesh-portainer"] },
{ name: "verdaccio", containers: ["verdaccio", "mesh-verdaccio"] },
{ name: "registry", containers: ["mesh-registry"] },
{ name: "route-proxy", containers: ["route-proxy"] },
{
name: "mailu",
containers: [
"mailu-resolver", "mailu-redis", "mailu-admindb", "mailu-admin", "mailu-imap",
"mailu-smtp", "mailu-antispam", "mailu-webmail", "mailu-front", "mesh-mailu",
],
},
{ name: "firewall", containers: [], node: true },
];
/**
* Dropped from the converging set, with cause — recorded as a finding rather than silently omitted.
*
* fail2ban declares `capabilities: ["intrusion-prevention"]`, but mesh-host advertises no such
* capability: profile/detectors.go defines container-runtime, package-manager, service-manager,
* firewall, overlay, seat, privileged and graphical-session — nothing for intrusion-prevention. So
* NO node can ever host fail2ban. Worse, `mesh-control assign` records the assignment even while
* reporting it "cannot be applied", and the whole-node `push` then refuses to resolve the ENTIRE node
* ("nothing was sent") over that one un-hostable assignment — one bad module blocks every other. It
* is therefore left unassigned here so the rest of the set can be proven. (novox/hq — escalated.)
*/
const DROPPED: { name: string; why: string }[] = [
{
name: "fail2ban",
why: 'declares capability "intrusion-prevention", which mesh-host has no detector for, so no node '
+ "can host it; and an unappliable assignment blocks whole-node resolution (nothing sent).",
},
];
/**
* The provable CORE: modules that converge WHOLE on this node (every container up and stable) once
* the substrate resolves and applies the set. This bed gates green on the CORE — a regression in any
* of these turns it red. It is the substrate + all five providers + the four consumers that reach
* their providers and stay up + the four standalone apps.
*/
const CORE = new Set([
"postgres", "redis", "minio", "mongodb", "mssql",
"keycloak", "gitea", "nextcloud", "invoicing",
"portainer", "verdaccio", "registry", "route-proxy",
]);
/**
* KNOWN GAPS: modules that RESOLVE and are placed, but whose containers do not stay up because the
* committed catalog manifest is incomplete (or, for firewall, a node-service expectation is unmet).
* They are reported every run with the exact failure, and escalated (novox/hq) — but they do not gate
* green, because the gap is in the catalog/host, not in this bed or the mesh substrate.
*
* umami — the mesh-umami provisioner needs the umami server URL and admin password in its
* provisioner.env; the manifest wires neither, so it dies "UMAMI url or admin password
* is not set". The umami SERVER itself comes up.
* photos — the server image is a bare `alpine` placeholder (a real deployment runs immich at
* :2283, where the runtime's MESH_PHOTOS_URL points); alpine has no long-running command
* so it exits, and the runtime dies "no photos API key". Not genuinely converted.
* mailu — the manifest generates only secret/database/admin env; the Mailu images need their full
* configuration env (HOSTNAMES, DOMAIN, …), so every Mailu container dies rendering its
* template: "jinja2 UndefinedError: 'HOSTNAMES' is undefined" (and the resolver's
* unbound.conf is malformed). mailu-redis/admindb/admin/antispam do come up.
* firewall — resolves and applies its package and ruleset, but nftables.service does not stay
* running, so the node reports firewall.load failed. Diagnosed live in the report below.
*/
const KNOWN_GAPS = new Set(["umami", "photos", "mailu", "firewall"]);
/**
* Host-port remaps applied at load time to break the co-located host-port collisions (see the file
* header). Keyed by module, then by the module.json port entry to replace. Container ports are
* preserved; only the host side changes.
*/
const REMAP: Record<string, Record<string, string>> = {
nextcloud: { "80": "8090:80" },
umami: { "3000": "3090:3000" },
invoicing: { "80": "8091:80", "9000": "9091:9000" },
};
let instanceId = "";
/** What the scenario's registry serves, by digest. */
let stocked: string[] = [];
function quote(s: string): string {
return `'${s.replaceAll("'", `'\\''`)}'`;
}
async function on(machine: string, command: string, timeoutMs?: number): Promise<{ out: string; ok: boolean }> {
const { stdout } = await exec(instanceId, machine, [
"sh", "-c", `exec 2>&1\n${command}\necho "__exit=$?"`,
], timeoutMs);
const marker = stdout.lastIndexOf("__exit=");
if (marker < 0) return { out: stdout, ok: false };
return { out: stdout.slice(0, marker), ok: stdout.slice(marker + 7).trim() === "0" };
}
async function must(machine: string, command: string, timeoutMs?: number): Promise<string> {
const { out, ok } = await on(machine, command, timeoutMs);
if (!ok) throw new Error(`${machine}: ${command}\n${out}`);
return out;
}
/** The control plane, a container on the first node. */
async function mesh(command: string, timeoutMs?: number): Promise<string> {
return must("anchor", `docker exec mesh-control /mesh-control ${command}`, timeoutMs);
}
/** The repository path a reference serves under — registry.ts's repositoryFor, mirrored. */
function repositoryFor(reference: string): string {
const withoutDigest = reference.split("@")[0] ?? reference;
const lastColon = withoutDigest.lastIndexOf(":");
const lastSlash = withoutDigest.lastIndexOf("/");
return lastColon > lastSlash ? withoutDigest.slice(0, lastColon) : withoutDigest;
}
/** The pinned reference this scenario's registry serves for a repository. */
function pinned(repository: string): string {
const found = stocked.find((r) => r.slice(r.indexOf("/") + 1, r.indexOf("@")) === repository);
assert.ok(found, `the scenario stocks no ${repository}; it serves\n ${stocked.join("\n ")}`);
return found;
}
/** The substrate bundle, its image references pointed at this scenario's own registry. */
function bundleFor(images: string[]): string {
let text = readFileSync(bundle, "utf8");
for (const ref of images) {
const repository = ref.slice(ref.indexOf("/") + 1, ref.indexOf("@"));
const escaped = repository.replaceAll("/", "\\/").replaceAll(".", "\\.");
text = text.replaceAll(new RegExp(`[A-Za-z0-9_.:-]+\\/${escaped}@sha256:[0-9a-f]+`, "g"), ref);
}
return text;
}
/**
* Load a committed module.json, rewrite every container image to the scenario's pinned digest, and
* apply the host-port remaps. Returns the manifest as a string and whether it needs a broker account
* (a runtime that reads MESH_BROKER_FILE — providers and tooled apps do; plain/node modules do not).
*/
function loadManifest(name: string): { manifest: string; broker: boolean } {
const path = resolve(catalogDir, name, "module.json");
const m = JSON.parse(readFileSync(path, "utf8")) as {
resources?: { type: string; image?: string; ports?: string[] }[];
};
const remap = REMAP[name] ?? {};
for (const r of m.resources ?? []) {
if (r.type !== "container") continue;
if (typeof r.image === "string") r.image = pinned(repositoryFor(r.image));
if (Array.isArray(r.ports)) r.ports = r.ports.map((p) => remap[p] ?? p);
}
const manifest = JSON.stringify(m);
return { manifest, broker: manifest.includes("MESH_BROKER_FILE") };
}
function tokenFrom(said: string): string {
const found = said.split("\n").map((l) => l.trim()).find((l) => l.length > 100 && !l.includes(" "));
assert.ok(found, `no token in:\n${said}`);
return found;
}
interface NodeState {
reached: boolean;
applied: boolean;
current: boolean;
waiting: boolean;
wrong?: { outcome: string; refused?: string | undefined; failed?: { id: string; error: string }[] | undefined } | undefined;
raw: string;
}
/** Ask the mesh, in its own terms, what a node has done with what it was sent. Never throws. */
async function nodeState(node: string): Promise<NodeState> {
const asked = await on("anchor", `docker exec mesh-control /mesh-control status --json`);
if (!asked.ok) return { reached: false, applied: false, current: false, waiting: false, raw: asked.out };
let state: {
wrong: { node: string; outcome: string; refused?: string; failed?: { id: string; error: string }[] }[];
waiting: { node: string }[];
reported: { node: string; outcome: string; current: boolean }[];
};
try {
state = JSON.parse(asked.out);
} catch {
return { reached: false, applied: false, current: false, waiting: false, raw: asked.out };
}
const word = state.reported.find((r) => r.node === node);
const bad = state.wrong.find((w) => w.node === node);
return {
reached: true,
applied: word?.outcome === "applied",
current: !!word?.current,
waiting: state.waiting.some((w) => w.node === node),
wrong: bad ? { outcome: bad.outcome, refused: bad.refused, failed: bad.failed } : undefined,
raw: asked.out,
};
}
before(async () => {
if (skip) return;
assert.ok(existsSync(catalogDir), `mesh-catalog modules not found at ${catalogDir}`);
const raised = await raise(loadScenario(`scenarios/${SCENARIO}.yml`), {
onProgress: (m) => console.log(`raise: ${m}`),
});
instanceId = raised.instanceId;
stocked = raised.images;
// anchor raises the substrate from its bundle, digests rewritten to the scenario registry's.
await must("anchor", `cat > /tmp/substrate.lock <<'MESHBUNDLE'\n${bundleFor(raised.images)}\nMESHBUNDLE`);
await must("anchor", `${HOST_PATH} apply /tmp/substrate.lock`, 900_000);
const up = await must("anchor", `docker ps --format '{{.Names}}'`);
for (const c of ["mesh-store", "mesh-broker", "mesh-control"]) {
assert.match(up, new RegExp(c), `the substrate did not raise ${c}:\n${up}`);
}
// Both machines join the one mesh and run a host so they apply what they are pushed.
for (const machine of ["anchor", NODE]) {
await mesh(`node add ${machine}`);
const token = tokenFrom(await mesh(`token issue --node ${machine}`));
const said = await must(machine, `${HOST_PATH} enrol --token ${quote(token)}`);
assert.match(said, new RegExp(`enrolled as ${machine}`), said);
await must(machine, `nohup ${HOST_PATH} run > /var/log/mesh-host.log 2>&1 & sleep 3`);
}
}, { timeout: 2_700_000 });
after(async () => {
if (instanceId) await destroy(instanceId);
await destroyAll(`${SCENARIO}-`);
}, { timeout: 900_000 });
test("the whole novox service set resolves, installs and converges on one node in one push", {
skip, timeout: 3_300_000,
}, async () => {
for (const d of DROPPED) console.log(`DROPPED ${d.name}: ${d.why}`);
// The overlay, so a consumer's binding `at` (the provider's private-network address) is non-empty.
// Provider and consumers are co-located on novox, but the address the mesh writes into a consumer's
// grant is the overlay address, so the overlay is placed on both nodes first (as two-node-db does).
await mesh("overlay place anchor --hub --endpoint 192.0.2.10:51820 --site lab");
await mesh(`overlay place ${NODE} --site lab`);
await mesh("assign anchor networking");
await mesh(`assign ${NODE} networking`);
// Add every module from its committed catalog manifest, issue the ones with a broker runtime, and
// assign all to novox. The resolver resolves the whole set at push time regardless of order. The
// loop is resilient: a module the node cannot host (a capability it does not advertise) is recorded
// and skipped rather than aborting the whole run, so ONE run yields the full per-module picture.
const issued: string[] = [];
const assigned = new Set<string>();
const refused: { name: string; why: string }[] = [];
for (const { name } of MODULES) {
try {
const { manifest, broker } = loadManifest(name);
await must("anchor", `printf %s ${quote(manifest)} > /tmp/${name}.json && docker cp /tmp/${name}.json mesh-control:/${name}.json`);
await mesh(`module add /${name}.json`);
if (broker) {
await mesh(`module issue ${name} --node ${NODE}`);
issued.push(name);
}
await mesh(`assign ${NODE} ${name}`);
assigned.add(name);
} catch (err) {
const why = (err as Error).message.split("\n").map((l) => l.trim()).filter(Boolean).slice(1, 5).join(" | ");
refused.push({ name, why });
console.log(`NOT ASSIGNED ${name}: ${why}`);
}
}
console.log(`issued broker accounts for: ${issued.join(", ")}`);
if (refused.length) console.log(`refused (node cannot host): ${refused.map((r) => r.name).join(", ")}`);
// ONE push. Resolution happens here; a resolver rejection surfaces as a non-zero push.
let pushError = "";
try {
await mesh(`push ${NODE}`, 120_000);
} catch (err) {
pushError = (err as Error).message;
console.log(`PUSH REJECTED:\n${pushError}`);
}
// The node cannot reach applied+current while a KNOWN_GAP node-service (firewall.load) keeps
// failing, so convergence is measured directly: wait until every CORE container is up (the node
// still pulls ~14GiB first), bounded. `settle` is used only to read the node's own verdict for the
// report — the wait is on the containers.
const coreContainers = MODULES.filter((m) => CORE.has(m.name) && assigned.has(m.name))
.flatMap((m) => m.containers);
const psNames = async (): Promise<Map<string, string>> => {
const out = (await on(NODE, `docker ps -a --format '{{.Names}}\t{{.Status}}'`)).out;
const map = new Map<string, string>();
for (const line of out.split("\n")) {
const [n, ...rest] = line.split("\t");
if (n) map.set(n.trim(), rest.join("\t").trim());
}
return map;
};
let psMap = new Map<string, string>();
if (!pushError) {
const until = Date.now() + 2_400_000;
while (Date.now() < until) {
psMap = await psNames();
if (coreContainers.every((c) => (psMap.get(c) ?? "").startsWith("Up"))) break;
await new Promise((r) => setTimeout(r, 8000));
}
// A moment for first-boot bounces to settle before the crash-loop check below.
await new Promise((r) => setTimeout(r, 15000));
}
psMap = await psNames();
const final = await nodeState(NODE);
const running = (name: string): boolean => (psMap.get(name) ?? "").startsWith("Up");
const users = (await on("anchor", `docker exec mesh-broker lavinmqctl list_users 2>&1`)).out;
const nft = (await on(NODE, `systemctl is-active nftables 2>&1`)).out;
// ================================================================================================
// The per-module report — this run's deliverable.
// ================================================================================================
const report: string[] = [];
report.push("================ WHOLE-MESH novox CONVERGENCE ================");
report.push(`node reached=${final.reached} applied=${final.applied} current=${final.current} waiting=${final.waiting}`);
if (pushError) report.push(`PUSH REJECTED (resolver): ${pushError.split("\n").slice(0, 6).join("\n ")}`);
const failedResources = final.wrong?.failed ?? [];
if (final.wrong) {
report.push(`NODE WRONG: outcome=${final.wrong.outcome} refused=${final.wrong.refused ?? "-"}`);
for (const f of failedResources) report.push(` failed ${f.id}: ${f.error}`);
}
if (DROPPED.length) {
report.push("---- DROPPED (not assignable on any node) ----");
for (const d of DROPPED) report.push(` ${d.name.padEnd(14)} ${d.why}`);
}
if (refused.length) {
report.push("---- REFUSED at assign ----");
for (const r of refused) report.push(` ${r.name.padEnd(14)} ${r.why}`);
}
report.push("---- CORE (gates green) ----");
const coreFailures: string[] = [];
const gapStatus: string[] = [];
for (const mod of MODULES) {
if (!assigned.has(mod.name)) continue;
const line = mod.node
? `${mod.name.padEnd(14)} node-service nftables=${nft.trim()}`
: (() => {
const states = mod.containers.map((c) => `${c}:${running(c) ? "UP" : (psMap.get(c) ?? "MISSING")}`);
const allUp = mod.containers.every(running);
return `${mod.name.padEnd(14)} ${allUp ? "OK " : "GAP "} ${states.join(" ")}`;
})();
if (CORE.has(mod.name)) {
const ok = !mod.node && mod.containers.every(running);
report.push(` ${line}`);
if (!ok) coreFailures.push(mod.name);
} else {
gapStatus.push(` ${line}`);
}
}
report.push("---- KNOWN GAPS (reported, escalated, do NOT gate green) ----");
for (const l of gapStatus) report.push(l);
report.push("---- broker accounts (issued modules) ----");
for (const name of issued) {
const present = new RegExp(`${NODE}-${name}\\b`).test(users);
report.push(` ${name.padEnd(14)} account ${present ? "present" : "MISSING"}`);
}
const summary = report.join("\n");
console.log(summary);
// Live diagnostics for the KNOWN_GAP failures, so the report carries the exact cause each run.
console.log(`\n---- firewall diagnostics ----\n${(await on(NODE, `systemctl status nftables --no-pager 2>&1 | head -12; echo '--- nftables.conf ---'; sed -n '1,20p' /etc/nftables.conf 2>&1; echo '--- journal ---'; journalctl -u nftables --no-pager -n 15 2>&1`)).out}`);
for (const mod of MODULES.filter((m) => KNOWN_GAPS.has(m.name) && !m.node && assigned.has(m.name))) {
for (const c of mod.containers) {
if (psMap.has(c) && !running(c)) {
console.log(`\n---- logs: ${c} (${psMap.get(c)}) ----\n${(await on(NODE, `docker logs ${c} 2>&1 | tail -20`)).out}`);
}
}
}
// ================================================================================================
// GREEN = the whole set RESOLVED (push accepted, resources applied), every CORE module converged
// whole, and NO core resource failed to apply. The KNOWN_GAPS (umami, photos, mailu, firewall) and
// DROPPED (fail2ban) are reported and escalated but do not gate — the gap is in the catalog/host.
// ================================================================================================
assert.equal(pushError, "", `the whole set did not resolve — push was rejected:\n${pushError}`);
const coreResourceFailures = failedResources.filter((f) => {
const mod = f.id.split(".")[0] ?? "";
return CORE.has(mod);
});
assert.deepEqual(coreResourceFailures, [],
`a CORE resource failed to apply:\n${coreResourceFailures.map((f) => `${f.id}: ${f.error}`).join("\n")}\n${summary}`);
assert.deepEqual(coreFailures, [],
`these CORE modules did not converge whole: ${coreFailures.join(", ")}\n${summary}`);
});