Closes#4282. mara: "scope looks good" — approving the plan posted
there (manifest + icons + minimal shell-only service worker + iOS meta
tags) and both explicit questions (network-first-with-offline-fallback,
never cache /api/*).
docs/web-ui/design-guide.md's "Layout & viewport" section already
asserted swarm-ui is installable as a PWA — this is what actually backs
it.
- manifest.webmanifest: name/icons/start_url/standalone display, theme
#cba6f7 / background #1e1e2e matching the mocha --purple/--bg values.
- sw.js: plain JS, not TypeScript — the DOM lib swarm-ui's own tsconfig
uses and the WebWorker lib a service worker's globals need are
mutually exclusive in one tsc program, not worth a second tsconfig for
a self-contained ~100-line file. Scoped to the app shell only, never
touches /api/* at all, network-first with offline-fallback-to-cache
(not cache-first) since main.js/main.css are unhashed filenames and a
cache-first SW would risk wedging an operator on stale JS after a
deploy.
- index.html: manifest link, theme-color meta, iOS
apple-mobile-web-app-* tags (Safari ignores the manifest spec).
- main.tsx: feature-detected SW registration.
- branding/hyperhive-maskable.svg: hyperhive.svg's own artwork already
fills nearly its whole canvas, so a maskable icon needs a padded
variant or an OS mask crops the outer ring/corner brackets — embeds
the original via a scaled <image> ref rather than duplicating markup.
- nix/packages/swarm-ui.nix: rasterizes icon-192/512/512-maskable.png
from the branding SVGs at build time via librsvg, rather than
checking in static PNGs.
Verified for real: typecheck+build green, and a real headless-chromium
tab driven over CDP confirms the service worker registers and becomes
the active controller, and a simulated-offline reload still serves the
full cached shell rather than a browser error page. nix build .#swarm-ui
also verified green, including the rasterized icon output.
Five cases: the leaf and the pairing that points at it, that the push
hands bao a path rather than the secret, a path per hive in the roster,
that a publisher holding an identity renders on a host with no store,
and the control that renders none without an identity.
They fail on this branch, and that is the finding. The suite was green
at 100 cases with the publisher already committed, because no fixture
enabled both authelia and a store identity — so the module's `config`
never activated and its script was never evaluated. An imported module
whose config never fires is as unmeasured as an unimported one.
What they surface: `swarm.authelia.agentClientSuffix` does not exist on
main. The agent principal these secrets belong to is minted by the PR
for the agent queue principal, whose branch adds both the option and the
`agentClients` list authelia mints from. Delivery of a secret nothing
mints cannot evaluate, let alone run — so that PR lands first and this
one rebases onto it.
The argv case strips comments before matching, which it earned: a
`script` renders its own comments into the text, and this unit's
comments name the hazard verbatim so the next editor does not
reintroduce it. Matching the raw text read that warning and failed —
a check the artifact defeats by describing the thing it is checked for.
Refs #3853
A hive that does not host authelia has no path to its own agent queue
client secret. The mint writes the plaintext to a host directory whose
other reader lives in a different container, so the host that mints is
the only place both trees are addressable — which is where this unit
runs.
Four pieces, in the order they depend on each other: the leaf
(glue-bao-tls.nix signs it, because the thing that owns a private key
owns issuing from it), the module declaring its own cert/key options,
the one-pairing glue file pointing them at that leaf, and the imports.
The unit is gated on holding a client identity, never on
deploy.bao.enable — that option is the co-location assumption itself,
and the publisher is the case that assumption excludes.
The secret is passed to bao as `value=@<path>`, never as an argv
element: bao is an external binary, so an argument is world-readable in
/proc for the life of the call.
Refs #3853
A sibling unit rather than more script in swarm-bao-controller-policy, because
that unit's name is an operator-facing string: docs/getting-started/setup.md
tells a reader to run `systemctl status swarm-bao-controller-policy`. Widening
it to two principals makes the name wrong; renaming it makes the instruction
wrong.
`after` and not `requires`. The controller's unit creates the KV and cert-auth
mounts this one writes into, so the ordering is real — but a failed oneshot
still counts as finished, so `requires` would neither wait for its success nor
re-run this unit when the sibling's own retry eventually lands. Ordering plus
this unit's `Restart=on-failure` is what converges.
Four module-eval cases, because the unit arrived with every claim about it in
prose and the suite still reporting the same count: the grant is write-only and
reaches the hive prefix alone (pinned as the whole capability list, since an
added capability is what a presence check misses, with negative arms for the
agent prefix, the bare swarm prefix and the policy path); it is ordered after
the unit that creates the mounts; it renders on the host; and the control, that
it does not render inside the store's container.
Refs #3853
Write-only `create`/`update` on `secret/data/swarm/hives/*`, and nothing else.
It copies secrets in and never reads one back; a read capability would let a
file-copier recover every hive's credentials rather than merely replace them.
`hives/` and not `swarm/*` because this principal has no business with an
agent's or a service's credentials, and the hive prefix is the only one whose
paths it produces. `secret/data/` is KV v2's ACL prefix, inserted by the engine
rather than written by the caller — the same trap the controller's grant
documents one binding up.
Named outside `hive-*`: the controller may create policies under that prefix,
and a policy it can rewrite is not a constraint on it.
The unit that writes this lands next. Refs #3853
The unit that will copy authelia's minted OIDC client secrets into the store
needs an identity of its own. Not the controller's: that grant includes
rewriting every hive's policy and login role, which a unit whose whole job is
copying one file has no business holding.
The subject joins `certAuthCns`, so it is unrepresentable as a hive name for
the same reason the controller's is — cert auth trusts the CA, and a hive's
own leaf carries its name as the CN.
The module-eval case collides with the SECOND list element and leaves the
controller's subject at its default. A list with one consulted element and one
dead one is indistinguishable from the first element's case, so without this
the addition could be inert and nothing would say so.
Refs #3853
mara, PR review: "i want to see it over time - sounds like you built
bar chart?" — correct, it was a bargauge (instant/lastNotNull). Switched
to timeseries, matching the CPU/memory-by-agent panels' own treatment of
a plain live gauge (raw value plotted across the range, no rate()/
increase() since it's not a cumulative counter).
Closes#4284. hyperhive.agent.claude_md.lines is already exported (see
hive-agent's claude_md_watch module) but had no dashboard panel — this
is the display half.
Same bargauge shape as the two sibling per-agent gauges already in this
file (container storage, active time): sum by (agent), instant query,
lastNotNull reduction, since it's a continuously-live gauge rather than
a cumulative counter. Metric name uses the dotted OTLP form per this
file's own documented convention for this store.
A collector's `refused` / `failed` / queue-depth counters are the only
signal that says telemetry is being dropped, and nothing read them at
any tier — so a collector losing records looked exactly like a quiet
system.
The hive tier could not be scraped without first naming its port. 8888
is the collector's built-in default and appeared in no config, which is
also why nothing comparing configured ports could see it clash with a
co-located collector — swarm-otel.nix sidesteps 8888 by hand for that
reason, and says so. Declaring the port and binding it explicitly makes
the value comparable; wiring the scrape is then one entry per tier.
Extending the port-collision assertion to cover it is deliberately left
out: that belongs with the other port work, and coupling a collision
fix to a scraping fix makes both harder to review.
Gate: 101 module properties hold, was 95. The six cases pin the rendered
scrape job rather than the option; the metrics pipeline naming the
prometheus receiver, a path never emitted on any hive before this since
the hive tier's scrapeTargets was empty everywhere; the `readers`
spelling, with a control so a missing telemetry block cannot pass the
port check vacuously; the swarm tier's own entry; the two tiers not
claiming the same port; and the absence arm, a hive with no collector
declaring no target.
scrapeTargets' description said "Empty by default, and that is the
shipped case". This makes that false, so the paragraph moves with it.
`agent-<hive>` reads as "the agent named <hive>" — which is the one thing that
identity does not carry, since it is minted per hive. It becomes
`hive-<hive>-agent`: the hive's own id, extended.
The rename is not a string swap. `hive-foo-agent` satisfies the hive parse too
(it strips to a hive named `foo-agent`), so the responder's agent rule now runs
BEFORE its hive rule — most specific wins. Hive-first would have handed every
agent its hive's grant, including writing that hive's status key, with nothing
to report it: the client authenticates and is merely able to do more than it
should.
`Policy::new`'s overlap check goes with the prefix it was written for. The
invariant the suffix form needs instead is that the suffix is non-empty: an
empty one makes `strip_suffix` succeed on every hive id, so the two principals
become one string and whichever arm runs first answers for both.
The suffix form also introduces a collision the prefix form did not have: a hive
genuinely named `foo-agent` mints `hive-foo-agent`, which is hive `foo`'s agent
id. The responder cannot see it — it has no roster, deliberately — so
`swarm-authelia.nix` asserts at eval that no hive name ends with the suffix. The
existing duplicate-id assertion does not cover this: it fires only when both
`foo` and `foo-agent` are on the roster, and with `foo-agent` alone there is no
duplicate, just a hive quietly receiving its agents' grant.
A test written by analogy with `the_prefix_alone_names_no_hive` failed, correctly
— `hive--agent` is a hive named `-agent` under the hive parse, which this module
cannot rule out. It now asserts only the part this module owns: no empty hive
name is ever expanded into a subject.
Agents have authelia *users*; they had no machine identity at all, so an
agent could not authenticate to the swarm queue as anything. This mints
one `agent-<hive>` OIDC client per hive beside the existing
`hive-<hive>` one, teaches the auth-callout responder an agent arm, and
opens the queue's client port on the bridge so a container can reach it.
One client per HIVE, not per agent: agents are created at runtime, and a
per-agent client would make creating one a config change plus an
authelia reload. The cost is that agents on a hive are indistinguishable
to the broker, which is deliberate and tracked separately.
The agent grant is deny-by-default twice over. An agent id matches no
hive rule, so it gets a hive's status-key grant from neither; and with
no agent subject configured the responder returns no grant at all rather
than an empty publish list, which would be a denial wearing a grant's
shape. What an agent may publish is a deployment's decision, taken
through `--agent-publish-subject` the same way `--hive-publish-subject`
already works.
`Policy::new` now refuses two prefixes where one contains the other. The
arms are tried in order, so that overlap does not error at match time -
it silently hands one principal the other's grant.
Not shipped here, and neither is reachable without it: no subject is
configured for agents anywhere in nix, and nothing yet delivers
`agent-<hive>.secret` into an agent container. Both belong to the stream
that will be the first consumer.
The reader had one failure branch, and it could not fail. It named three
states — the store holds nothing, the store is sealed, the store is
unreachable — treated all three as success, and sent bao's stderr to
/dev/null, which is the only thing that tells them apart.
The degraded mode is right and is unchanged: a missing registration token
means new agent accounts cannot be provisioned and nothing else breaks,
so this still exits 0 and leaves the local token alone. What changes is
that the journal now carries the store's own message instead of a
sentence of ours asserting all three at once.
Deliberately not branching on that message. Distinguishing "no value
found" from "permission denied" programmatically would mean matching
prose from an external tool that I cannot reach a live instance of to
confirm, which is inventing a discriminator rather than reporting one.
Nothing asserted this script before — eleven module-eval cases cover the
unit's existence and its ordering, none its contents, which is how the
branch stayed. The case added here covers the property that was missing.
Gated: `96 module properties hold`, exit 0 (95 on the base commit, +1 =
the case added). `nix fmt` reported 0 changed over 723 files emitted in a
fresh worktree. Tracker-tag, comment-block and doc-pointer lints all exit
0 after staging.
Both places that warned about a hive named after the controller's cert-auth
subject still told the reader it was unmitigated, and the option's description
recommended a migration — "change this to something outside the hive-name
grammar, at the cost of a role rename in any store that has already run the
granting unit" — that is no longer the answer and is not cheap.
swarm.nix now feeds the subject into the guard on `swarm.hives`, so a
colliding roster fails evaluation. Both paragraphs get SHORTER saying so: a
guarded hazard needs the consequence ("reserved as a hive name") rather than
the threat model, and the write-site keeps only the sentence the next person
needs — a role added beside this one must join that list.
Found by sweeping for the claim rather than for the symbol: the change that
made these stale touched swarm.nix, swarm-otel.nix and module-eval.nix, so no
diff-context or doc-comment rule covers a paragraph two files away. Grepping
the tree for "would satisfy" and for prose about a hive named after a subject
turned up exactly these two and nothing in docs/.
The two hive-name guards lived in swarm-otel.nix, inside its
`config = lib.mkIf (… && deployCfg.swarm-otel.enable)`. A swarm running the
secret store and the controller but no collector therefore had no hive-name
check at all, while the names were still composed into OIDC client ids, bao
policies and cert-auth roles exactly the same way. They move to swarm.nix,
which declares `swarm.hives` and is unconditional. swarm-otel keeps the
assertion that its own entry is still in the shared list — that one is about
this module's stake in a file it no longer controls.
The equality guard also takes the store's cert-auth subjects now. Cert auth
trusts the CA, so `allowed_common_names` is the whole of what narrows a role
to one identity, and the same CA signs every hive's leaf with the hive's name
as its CN. A hive named after a role's subject presents a certificate that
role accepts, which for the controller is write access to every hive's
credentials and policies.
A list rather than the one string, because the next role added beside it
widens what a hive name must not collide with, and because the subject is an
option an operator sets — a literal deny entry covers the default and nothing
else.
Four module-eval cases, two of them controls. The fixture overrides the
subject to `ctl` on purpose: the default contains `swarm`, which the substring
guard catches whatever the new arm does, so a fixture using it could not tell
the two apart. The controls are that a legal roster trips neither guard, and
that all three fixtures really do have the collector disabled — without the
second, every case would pass while testing the arrangement they exist to
rule out.
The crate had a single path convention and it was per-agent:
`swarm/agents/<agent>/matrix/<account>`. The secrets still to move into the
store do not fit it — one belongs to a hive, one to a swarm service, one to
the controller itself — so each would have picked its own shape, and each
would have been a separate grant to get wrong.
mara ruled the scheme on the epic: `swarm/<kind>/<name>/<secret>`, over
`agents`, `hives`, `services` and `controller`. This lands it.
`Kind` is an enum rather than free strings for one reason: the store's grant
is written in nix and cannot be reached from Rust, so a misspelled kind is a
403 at provision time and not a compile error. `Kind::ALL` lets a test
enumerate the set instead of restating it, which is what makes adding a kind
a deliberate edit rather than an accidental grant.
Note `Kind` sits beside `checked_segment`'s existing `kind` argument, which
means something else entirely — the label of the name being validated. They
are not the same concept and should not be merged.
Nothing about the rendered policy changes. `policy::render` still grants read
on the agent kind alone; the other kinds are absent on purpose, because what a
hive may read of its own kind is a boundary question and not a consequence of
the namespace growing. The controller's write grant likewise stays scoped to
`agents/` — it widens when a path outside it gains a writer, not when the
kinds are declared.
Verified: `cargo test -p swarm-secret-client` 23 passed, 0 failed. The two
tests pinning the rendered strings (`the_document_grants_read_over_the_whole_agent_prefix`
and matrix's path assertion) still assert the same literals they did before,
which is what shows this is a faithful port rather than a reshape. `nix fmt`
710 emitted, 10 formatted, 0 changed; the three scripts/check-*.sh lints pass
with the change staged. No reference to the removed `path::AGENT_PREFIX`
survives in the crate or in nix — checked with a scoped pattern, because the
unqualified name also belongs to hive-host-sock's container prefix and greps
for it are answering a different question.
The mode was declared twice in this file — the service unit's
RuntimeDirectoryMode and the socket unit's DirectoryMode — with only a
prose "must match" note tying them. Whichever unit activates first creates
the directory, so they cannot be allowed to disagree.
Both literals are in one file, so they become a `let`. Deleting a copy
beats checking it, and unlike rendering the mode into hive-priv it costs no
config knob for a value nobody should ever set.
hive-priv's tmpfiles.d entry for the same path is a third declaration that
cannot read this binding, and is left in step by hand. An earlier revision
of this branch added a CI check for exactly that pair; mara pointed out it
was keyed to one path rather than to the class, and looking for the general
case found two more paths declared by more than one mechanism — including
/run/hive-agent, where hive-gateway's tmpfiles rule and hive-priv's
generated one disagreed on the owner and the winner depended on systemd's
read order. That check is being reworked as a general one, tracked
separately, so nothing about it rides in here.
Verified: nix fmt (713 traversed, 5 formatted, 0 changed); the three
scripts/check-*.sh lints all exit 0 with the tree staged; .forgejo/ is now
byte-identical to main and the diff is this one file. checks.module-eval
reported 91 module properties hold on the previous revision of this branch
— the only nix change since is comment text inside the same let block,
which cannot affect evaluation.
`auth.disable_login_form` was gated on `ssoLocal` — `grafana.enable &&
authelia.enable`, i.e. "both of them run on THIS host". With authelia
elsewhere in the swarm that is false, so a deployment that is very much
using SSO still rendered grafana with its local login form enabled, on a
vhost the gateway publishes, for a product that ships an `admin`/`admin`
account.
The reason that matters was already in the module, three lines up
("Grafana ships an `admin`/`admin` account, and this vhost is on the
public gateway") — it was just attached to a conditional. Whether a
password box sits on a public login page is not a per-host question.
Per mara on the docs PR for this: "grafana requires sso - no local
login". The OIDC block below stays gated on locality; making that follow
the same swarm-wide question is a larger change with its own tracking.
The module-eval suite already had the fixture this needed: `grafanaOldPath`
enables grafana and not authelia, which is exactly the shape the login
form stayed enabled in, so the regression case needs no new hive. 90 -> 91
properties.
Closes#4218.
`docs/tools/subagent.md` and `docs/tools/bash.md` both described their MCP
server as injected "unconditionally". Both entries are `lib.mkDefault`, and
the module says why one line above each: "so an agent.nix can still
override/disable the entry", "so the operator's own agent.nix can override
the entry".
The word matters for the subagent one in particular. The same comment block
records the framing that it is default-on for now and should become a real
capability gate later, so "can I turn this off today?" is a question an
operator has — and "unconditionally" answers it as "patch nix/" when the
answer is one override in agent.nix.
Both pages now say default, and say what the default yields to.
The other direction on the same page: `subagentHttpPort`'s option
description and the unit comment beside it both listed three tools,
`start`/`continue`/`interrupt`. The daemon serves four. #4101, which
introduced it, is titled with the three-verb phrasing, so `status` landed
afterwards and never reached either description — while `subagent.md` had
the full set all along. The option description renders into the generated
options doc, so it is the one an operator reads.
Closes#4231.
`nix/host-modules/swarm-peers-removed.nix` exists to turn an "option does
not exist" error into a warning that says where the entries went. That
warning is its whole deliverable, and no fixture set `swarm.peers` — the
string appeared 0 times in module-eval.nix, so the shim was never
evaluated by anything.
It differs from its ten siblings in what a broken shim looks like. The
others re-route a value, so a failure shows up as a wrong rendered
config. This one renders prose nothing reads back, so a `mkIf` that
stops matching or a rename of the `swarm.hives` it points at fails
silently, and lands on the one operator who needed it.
The check already carries an old-path fixture for ten migrated
namespaces (wireguard, forge, matrix, nats, authelia, controller, ui,
stores, grafana, statusPublish). `swarm.peers` was the eleventh and the
only uncovered one.
Two peers, only one carrying `caCert`, because the module emits a second
warning filtered on exactly that attribute — with a single peer the
filtered list and `attrNames` are the same list, so a `withCaCert` that
had collapsed into `attrNames` would still read green. The third case is
the control: a hive that never set `peers` must get neither warning,
without which the other two pass on any config whose warning list
happens to carry the string.
Closes#4188.
`deploy.allSwarmServices` derives nine service enables and appeared
nowhere in the check, so the tier a service sits on was prose only. The
tenth, the swarm controller, already had this exact pair of cases — it
rides `singleHostSwarm` instead, and swarm-ui follows the controller.
Four cases: the switch turns its nine on, a hive that does not host them
runs none, an operator placing one elsewhere still wins over the
`mkDefault`, and hosting the shared services does not make a hive the
swarm's control plane.
The roster is counted before it is read: `lib.all` over an empty set
holds vacuously, so a roster that lost a member would otherwise turn the
case green by measuring nothing.
Closes#4186
Nothing in this suite evaluated nix/agent-modules/ at all: every fixture
was a host, so a typo in a rendered container config surfaced on a real
deploy and nowhere else. This adds an `agent` constructor beside `hive`,
off the same `nixosModules.agent-base` the meta flake hands a container.
It also adds the suite's first two-hive fixture. Every existing one
declares `swarm.hives.h1` alone, so a per-hive arm written against those
passes on a hardcoded literal — which is exactly what the new per-hive
logs pipeline needed covering.
Eight cases, each paired with the control that makes it mean something:
the absence arms with a presence half, the per-hive arm with a roster
length check, because `lib.all` over an empty roster holds vacuously.
Each was then shown to fail. Eight mutations across the three files —
`directory` back to the runtime default, a pipeline naming no receiver,
an exporter aimed at a loopback literal, an exporter name that stops
reading `protocol`, a missing hive-tier logs pipeline, the metrics-only
processor inside it, a per-hive pipeline hardcoded to one hive, and logs
pipelines exporting to the metrics store — all caught, none survived,
none skipped, each run's baseline green.
Part of #3940.
An agent container writes a complete journal — 991 MB and nine days deep
on this hive — that nothing outside it can read: the host-side
per-container journal directory is an id-mapped bind mount, and journald
writes nothing into it. So the reader has to run inside the container,
and the path it would push to did not exist.
Three tiers, one vertical slice, because any two of them alone are
silent:
- the agent container gains an `opentelemetry-collector` with a
`journald` receiver aimed at its own journal and an exporter aimed at
the same base address every in-process producer already exports to.
- the hive collector gains `service.pipelines.logs`. Without it the
`otlp` receiver answers 404 on `/v1/logs` — measured, and
indistinguishable from a route that was never meant to exist.
- the swarm collector gains a per-hive `logs/<hive>` pipeline beside
`metrics/<hive>`. Without it the push is accepted, answered 200, and
routed nowhere.
The receiver's `directory` is stated rather than inherited, and that is
the load-bearing line: its default is the RUNTIME journal
(`/run/log/journal`), which in an agent container is empty. Left at the
default this whole path validates, starts, reports healthy and forwards
nothing. The assertion beside it covers the same silence from the other
end — a `volatile` or `none` journald storage empties the directory the
receiver reads.
Attribution follows the tier that can prove it. The forwarder stamps
`agent`, which no host-side reader could supply; `hive` is deliberately
left to the swarm tier, which upserts it from whichever receiver
accepted the sample, precisely so the label comes from something the
sender cannot write.
No `units` allowlist, unlike the swarm tier's journald receiver. That
one needs one because the host's journal also holds an operator's own
session; a container's journal is the harness and what the harness
spawns. Measured volume is 20827 entries / 6.3 MB per agent per day,
with nothing logging below `info` — so the receiver's `info` default
filters nothing and there is no bill to justify a knob.
Agent containers only, per the ruling on the issue: swarm services need
one forwarder per service container and get re-measured once this works.
Part of #3940.
The bootstrap unit writes a policy granting `secret/data/swarm/agents/*` and
nothing creates that engine. A fresh OpenBao has no `secret/` — only a dev-mode
one does — so `swarm-controller`'s first credential write answers `no handler
for route "secret/data/swarm/agents/<agent>/matrix/<name>". route entry not
found.` Measured on the live host at 21:27:27Z; #4171.
`git grep` for `secrets enable`, `kv-v2`, `kv_v2` and `sys/mounts` returned zero
across the whole tree. Control, so the zero means something: `auth enable` in
this same file returns 2 — the same defect was already found and fixed once, for
the cert auth mount, with a comment that states the principle. This is the other
half of it.
The mount name is now bound once and interpolated into both the policy text and
the new step, because a grant and a mount that disagree is exactly the failure
being fixed.
Placed outside the client-CA block: the controller writes *through* this mount
regardless of whether anything can log in by certificate. `module-eval` asserts
that, since one indentation level decides it.
Grants, measured against a real openbao 2.6.2 rather than derived:
`-output-policy` asks for `sys/mounts/secret` create+update, and a token holding
exactly `sys/mounts` read + `sys/mounts/<path>` create/update enabled the engine
— **no `sudo`**, unlike `sys/auth/cert`. Negative control: the same token on an
ungranted path got 403, so the grant is what made it work. `setup.md`'s
documented policy gains those two.
Also from that session, each deciding how this is written: re-enabling an
existing path errors (exit 2), so this asks first like the auth mount does;
`secrets list -format=json` keys look like `"secret/"`, so the `case` idiom
ports over; and `kv put -mount=<p>` reports `<p>/data/...`, confirming v2 — the
prefix the policy grants and the client writes.
setup.md also drops a check that cannot work: it told the operator to confirm
with `bao read auth/cert/…`, which 403s because the host wrapper carries no
token. `systemctl status swarm-bao-controller-policy` needs no credential and
names the three success lines. The first-attempt-after-rebuild race is now
written down too — the store is still coming up, and the 30s retry is what
lands.
Refs #4171.
Both claims are load-bearing prose, and both are wrong in a way nothing in
the tree reads (#4168).
swarm-nats says the queue is "reachable from every agent container on the
hive" because the container shares the host netns — in a comment, and again
in the operator-facing `calloutUserPublicKey` description, which renders into
the options doc. Agent containers do not share it: `PRIVATE_NETWORK=1` is
written unconditionally (hive-priv/src/main.rs, and hive-priv-sock says
"isolation is the only supported mode"), and hive-network.nix states the
shared-netns mode was removed. The bridge firewall opens 53/67/80/443 plus
`exposeHostPorts`, whose only consumer tree-wide is otel — the queue's port is
in none of them, and no gateway route exists either (`grep -c nats` in
hive-gateway/default.nix -> 0; control `forge` -> 3, so the zero means
something). Its actual clients are host-side: HIVE_C0RE_NATS_URL and
SWARM_CONTROLLER_NATS_URL, both 127.0.0.1 on a single-host swarm, plus each
remote hive dialling a routable address.
swarm-victorialogs says the ingest endpoint has "no authentication of its
own". Upstream offers Basic Auth via -httpAuth.username / -httpAuth.password
(and -metricsAuthKey / -deleteAuthKey / -pprofAuthKey); this module sets none
of it. "The software offers nothing" and "we configure nothing" send a later
reader to different places, so the wording now says the second one.
Neither conclusion changes. The queue must still refuse everyone until the
callout responder exists, and the logs endpoint must still be pinned to
loopback — only the reasons were false.
Checked while here: swarm-authelia's identical "no authentication of its own"
is TRUE (upstream's telemetry.metrics has exactly enabled, address, buffers,
timeouts), and otel.nix's "reachable from agent containers and nowhere else"
is true and better-founded than it claims — the receiver binds the bridge IP,
not just a firewall hole.
Refs #4168.
`swarm-bao-controller-policy` creates the `swarm-controller` policy and
cert-auth role — the credential every hive logs in with. It has never
succeeded on any deployment, and the reason is where it ran.
Inside the container it had neither of the two things the store demands.
Its `BAO_ADDR` was the public DNS name, which from that netns resolves to
the hive bridge: `dial tcp 10.42.0.1:8200: connect: connection refused`.
And every API listener carries `tls_require_and_verify_client_cert`, while
`tlsDir` holds the server's leaf and the CA that signs clients — no client
identity at all. Fixing only the address moves the failure one hop.
The comment above the unit asserted the opposite — that in there the store
is "reachable without a client certificate at all, which is the point". The
listener config decides that, and says otherwise. That belief is what put
the unit in the container, so it goes with it.
On the host all four coordinates already exist: `baoCli` carries the
address, the CA, the certificate and the key, so the unit needs no
`environment` block at all. `bootstrapTokenFile` was always a host path —
the container only saw it through a bind mount. Nothing new crosses the
boundary; the mount gets no wider.
The retry bound is resized with it. 10 attempts at 30s is five minutes,
and under `seal = "shamir"` an operator unseals by hand, so it would give
up before a human arrived — permanently, because `start-limit-hit` does
not self-heal. That is the same silent no-bootstrap this issue is about.
2880 × 30s covers a day, inside a 25h window.
module-eval follows the unit to the host and gains an arm asserting it is
NOT rendered inside the container: the move is the fix, so the side it
landed on is worth pinning.
`StartLimitBurst` sat in `serviceConfig`, so it rendered into `[Service]`,
where systemd silently ignores it — the unit retried every 30s forever.
Measured on a live store: the journal reports `restart counter is at 18`
against a burst of 10.
This repo already states the rule and pins it with a test:
`hive-priv/src/main.rs` renders its drop-in with `StartLimit*` under
`[Unit]` and says why — "systemd silently ignores them under `[Service]`,
so a bound that moved sections would look configured and do nothing".
That is exactly what happened here, in another module.
Moving the burst alone would not have fixed it. systemd's default window
is 10s while `RestartSec = 30`, so at most one restart falls inside it and
a burst of 10 is unreachable; the interval has to exceed `RestartSec` times
the burst. 600 matches the value hive-priv already uses.
Uses the NixOS service-level options rather than a hand-written
`unitConfig`: nixpkgs renders `startLimitBurst` / `startLimitIntervalSec`
into `unitConfig` itself (`nixos/lib/systemd-lib.nix`), and `hive-ci.nix`
already sets `startLimitIntervalSec` that way.
The module-eval case asserts placement where nixpkgs puts it, and that
`serviceConfig` does not carry it — so moving it back fails the build.
Reading a role out of the store took four round-trips of environment
guessing: the certificate has no IP SAN so loopback cannot verify, the
DNS name resolves to the bridge from inside the container, and the CA
the client needs lives on the host, which had no `bao` at all.
The wrapper carries the address, the server CA and — where the PKI glue
minted one — the reader's own leaf, so cert-auth login needs nothing
typed. `--set-default` throughout: an operator pointing BAO_ADDR
elsewhere still wins.
Only the wrapper reaches PATH. `wrapProgram` renames the real binary,
so there is no unwrapped `bao` to reach by accident, and the module-eval
case asserting the package's absence is what keeps a later "install it
too" from undoing that.
`clientCn` fell back to `cfg.domain` when `hiveName` was unset. That branch
cannot run: `hive-network.nix` asserts `hiveName != null` under
`mkIf services.hyperhive.enable`, and this file's `config` is gated on the same
predicate, so any host that evaluates the conditional has already failed the
assertion.
Worse than dead, it read as a second supported spelling of a hive's identity —
which is what a cert-auth role matches on. It was not even the hive's own
domain: `cfg` here is `services.hyperhive.swarm.bao`, so the fallback resolved
to the store's address, one string shared by every hive in the swarm and the
same CN the server leaf carries.
Reading the option directly matches what other modules needing the name already
do (`hive-c0re/environment.nix`).
Creating a hive's cert-auth role means writing the authority into the role
by value -- the store matches a presented certificate against the role's own
copy -- and nothing gave this daemon that file.
Named separately from deploy.bao.clientCaFile rather than read off it: that
option is the store's, saying which readers the store trusts on the host
that runs it, while a controller runs anywhere. The glue module supplies it
where the two are co-located, which is the same split baoClientCertFile
already makes against the hive reader's leaf.
Gated on the identity as well as the CA. Without a leaf there is nothing to
write a role with, so the file would reach a daemon that cannot act on it.
The module-eval arm needed a fixture of its own: a deployment that
self-signs both ends points clientCaFile and serverCaFile at one file, so on
the existing fixture the two authorities are the same string and wiring
either into the other's slot passes. controllerTwoCas is where they differ.
Closes#4124.
The controller's policy granted only the bootstrap paths -- hive cert-auth
roles and hive ACLs. #4113 then made it a secret WRITER, and nothing related
the grants to the paths the code writes, so every matrix token provision
answered 403. The two halves landed on different issues and neither looked
wrong on its own.
`secret/data/` is KV v2's ACL prefix and is absent from the path the code
passes, so matching `swarm-secret-client`'s spelling literally would have
granted nothing. Write-only: the controller mints these and never reads one
back, and a read capability would let it recover every agent's credentials
rather than only replace them.
The gate is the point. Two module-eval arms -- the grant exists and is not
a broader wildcard, and its capability list is pinned whole, because an
ADDED capability is what a presence check misses -- plus a test in path.rs
pinning MOUNT/AGENT_PREFIX and naming the nix file, since renaming either
constant is a silent 403 rather than a compile error.
setup.md carried two warnings this makes false: that nothing in the tree had
ever authenticated to the store, and that no deployment shape mints a leaf
whose CN reads swarm-controller. glue-bao-tls.nix has minted one since #3726
item 1.
The swarm UI had nowhere to POST an external matrix account to: this daemon
had no matrix-account code at all and no `swarm-secret-client` dependency, so
the last leg of #3726 — a credential reaching an agent — had no entry point.
`PUT /api/hives/{hive}/agents/{agent}/matrix-accounts/{account}` writes the
credential to the store under the agent's own path and publishes a
`CredentialNotice` on that hive's credential subject. All three path names are
load-bearing: agent + account locate the secret, hive routes the notice. The
account is a path segment rather than a body field so that splitting the 1:1
account-to-agent mapping later is a new route, not a changed payload.
Store first, notify second, and the order cannot be swapped: a notice that
overtakes its own write reaches a hive that reads nothing, and the hive
deliberately does not retry. The publish is followed by a flush for the reason
`publish_deploy` flushes — `publish` hands the message to the connection's
write buffer and returns, so the response could otherwise outrun the notice it
reports as sent.
The store client is built per request rather than held in `AppState`, matching
what the hive side does inside `deliver`: a login that expires is not worth
caching for a route this cold.
`swarm_hive` is `declaration_target`'s two name checks, extracted so this
handler makes them identically rather than in a second copy free to drift.
`declaration_target` still tests the writer first, so a deployment with no
queue answers 503 whatever the caller spelled.
## The nix half
#4081 minted the controller's leaf and gave it `baoClientCertFile` /
`baoClientKeyFile`, deliberately stopping there — the leaf is minted whether or
not a controller runs on that host. Nothing consumed those options, so the
identity never reached the process. Measured before writing: `git grep
baoClientCertFile` returned 5 sites and zero consumers, against a control
(`tokenEndpoint`, 4 hits in the same file) proving the search can see
consumption where it exists.
The unit now gets `BAO_ADDR` / `BAO_CLIENT_CERT` / `BAO_CLIENT_KEY` /
`BAO_CACERT` and the matching `LoadCredential` entries, following
`hive-c0re/environment.nix`'s `%d` credential shape.
The gate is `deploy.swarm-controller.baoClientCertFile`, NOT
`deploy.bao.clientCertFile`. The latter is the hive reader's identity and its
policy scopes a hive's own secrets; wiring it here would evaluate, deploy, and
fail only when the daemon tried to write an agent's credential.
Two `module-eval` arms cover exactly that. The presence arm asserts the
`LoadCredential` *source path* (`…:/var/lib/swarm-bao-pki/controller.pem`) and
not just the `%d` name, because a `%d`-only assertion passes while the daemon
holds the wrong policy. The absence arm (`controllerNoStore`) is what makes the
presence arm mean anything.
`RestrictAddressFamilies` already covers the store client; its own comment asks
for the family to be added with the client, and AF_INET/AF_INET6 are present.
Contributes to #3726
The log-store board reports rows ingested, disk size, free disk and errors —
every one of them a fact about VictoriaLogs itself. None of them can say
whether a particular unit or host is contributing, so aggregate ingestion
reads healthy on host-tier units while a whole tier ships nothing, and there
was no way to tell those apart from a dashboard.
Add four panels to that board rather than a second one. Every other board here
is per service — agents, authelia, openbao, forge, queue, the two stores — so
a second board about the same service would have made an operator guess which
of the two answers their question.
Nothing in the new panels names a unit: both breakdowns discover their rows
from the data, so a source that starts shipping appears without an edit, and
one that never existed is simply not there.
The ungrouped total is a control, not a summary. An empty breakdown renders
the same whether the query is malformed or the source genuinely never shipped;
with the total beside it, nonzero-and-empty is a broken query and
zero-and-empty is an empty store. Being on one board buys a second reading it
could not have alone: that total and `Log rows ingested` are the same quantity
measured by querying and from the store's own metrics, so the two disagreeing
means rows arrived that no query can reach — which is the failure this
pipeline actually had once.
The two breakdowns are bargauges rather than tables. A `stats` query returns
one frame per group, and a table panel renders one frame at a time behind a
series picker; bargauge consumes multi-series natively, for the same reason
the timeseries beside it always did. The alternative was a table plus
labelsToFields plus merge plus organize — three assumptions where this needs
none, in a spot I cannot render to check.
Series are named by their label rather than `rows{_SYSTEMD_UNIT="x.service"}`,
which also cleans up the timeseries legend. The datasource supports
legendFormat; its query editor's own placeholder is `{{label}}`.
These are the first panels to query the logs datasource at all; the other
eight boards are prometheus, including this one until now, which reads
VictoriaLogs' self-metrics out of VictoriaMetrics. So `renderDashboard` grows
a `@logsDatasourceUid@` substitution alongside the metrics one. The binding it
points at already existed, and its comment claiming a dashboard panel named it
only becomes true with this commit.
The uid is unchanged, so existing links and bookmarks still resolve; only the
title widens to match what the board now covers.
The query model was read out of the plugin in the store rather than guessed:
`queryType` is one of hits/instant/logs/stats/statsRange, and Stats/StatsRange
are the two that consume `expr`. The stream field names come from the module
that builds them — swarm-otel.nix's `_stream_fields=_HOSTNAME,_SYSTEMD_UNIT`.
The queries themselves are confirmed against the live store: mara ran the
by-unit one over seven days and it returned rows.
Refs #4084
glue-bao-tls.nix signs a third leaf. It is minted whether or not a
controller runs here, because the case it serves is the one where it
does not: a controller elsewhere needs a leaf from this CA and cannot
sign one, so issuing it here turns "obtain a certificate out of band"
into "copy this file".
glue-controller-bao-identity.nix holds the pairing and nothing else --
which paths this host's controller reads. Gated on the leaf existing
rather than on deploy.bao.enable, so a controller on the store's host
and one three networks away with an out-of-band leaf get the same
wiring; gating on the store would have made the co-located case the
only supported shape. The directory comes from deploy.bao.clientCertFile
rather than repeating glue-bao-tls.nix's literal, so moving the PKI
moves both.
module-eval gains three cases and two fixtures, because nothing asserted
the PKI script before: an earlier commit added a leaf to that rendered
unit and left the derivation unchanged. The fixture's CN is deliberately
a value no default could supply, so "the role and the leaf both carry
it" says they read one option rather than that both happen to say
swarm-controller.
Gates: 62 module properties hold (59 before, plus these three), on a
derivation hash that actually moved -- this suite is a cache hit when
only fixtures change, so an unchanged hash would have meant the cases
never ran. nix fmt clean, all three scripts/check-*.sh exit 0.
Two halves of one interface that currently exists in one file only.
The CN was a `let` binding in swarm-bao.nix. Whatever mints the
controller's leaf has to spell it identically, and that lives in another
file, so a literal in one place is an interface with no name. It becomes
`deploy.bao.controllerCommonName` -- under bao because it is a property
of the role this module writes, not a credential belonging to the
controller.
The certificate gets `deploy.swarm-controller.baoClientCertFile` /
`baoClientKeyFile` rather than reusing `deploy.bao.clientCertFile`. That
one means "this host as a reader" and carries the hive's name, while the
controller's policy lets it create roles for every hive; one certificate
serving both would hand that power to whatever else reads the store
here.
Both default to null. The glue that mints the leaf sets them with
`mkDefault`, the same way glue-bao-tls.nix already supplies the reader's
paths -- so the controller module names no path of bao's, and deleting
the glue leaves a controller that takes operator-provided ones.
Nothing reads the new options yet; the minting half is the next commit.
`module-eval`'s derivation is unchanged, which is the honest result for
a change that adds options without altering any asserted value -- it
proves the tree still evaluates, not that anything behaves differently.
Four lines about "no store here, so no bind mount and no unit" sat
above the swarm-ui case, which they have nothing to do with. The case
they describe -- a bootstrap token on a host that runs no store -- had
no comment at all, so this is a move rather than a deletion.
Reworded on the way, because it opened with "the arm that makes the one
above mean something". A comment that names a POSITION retargets when
the file is reordered, which is how it came loose in the first place;
it now names the granting-unit cases.
Noticed while adding a case directly above it, which opened with the
same sentence.
The policy the granting unit already writes grants paths under
`auth/cert/certs/*`, and nothing in the tree creates that mount. Every
certificate login therefore fails against a path that is not there --
the controller's own, and the per-hive ones it is meant to issue
against the same mount.
Same unit, same bootstrap token: check whether cert auth is mounted,
enable it if not, then write a role binding CN `swarm-controller` to
the `swarm-controller` policy.
Idempotency is a read rather than a tolerated error. `auth enable`
fails on an existing mount, and recognising that would tie a rebuild to
an error string no run of this store has ever produced, so the unit
asks `bao auth list` and mounts only on absence. That read is why the
token policy in setup.md gains `sys/auth`.
Each grant came from `bao <cmd> -output-policy`, which prints what a
command requires without running it -- the same way controllerPolicyText
was derived. Enabling an auth method needs `sudo` on `sys/auth/cert`,
which the documented token did not have.
Gated on `clientCaFile`, not on the token alone: `swarm-bao-certs`
installs `client-ca.pem` only under that condition, and a role's
`certificate=` has to name a real CA. The policy write, which needs no
CA, is unchanged in that case.
Nothing can present a certificate for this role yet -- the only client
leaf the tree mints carries CN = the hive's name -- and none of this has
been run against a live store. Both are stated in setup.md.
The last of the twelve. `swarm.forge.package` becomes
`deploy.forgejo.package`, beside the `behindGateway` / `mirrors` /
host-secret options that were already there.
⚠️ `deploy.forgejo`, not `deploy.forge`. The forge joins the controller
and the UI as a mover where the two sides simply do not share a name,
so the shim comment says which one.
Four readers, and one of them is why an anchored grep was the wrong
instrument: `:678` reads `cfg.package.data`, a sub-attribute, which
`cfg.package}` and `cfg.package/` both miss.
Three copies of the same sentence claimed the package stays under
`swarm.forge` — `deploy.nix`, this module's own `deploy.forgejo`
header, and `docs/integrations/forge.md`. The docs one is outside
`nix/` entirely. A namespace split gets described wherever the split is
explained, which is not only where the option is declared.
`forgeOldPath` gains the old path, and its comment's count of the
host-side options goes five → six with it.
Both are regressions from earlier commits in this same slice, not
pre-existing drift:
- the nats comment listed "the responder package" among what stays
under `swarm.nats`; `authPackage` moved in a17286c4
- the matrix comment listed "its package" among what `swarm.matrix`
keeps; both matrix packages moved in 1b2da54c
Each move edited the prose in the module that declares the option —
`swarm-nats.nix` carries the same "responder package" sentence and it
was corrected there — but deploy.nix keeps its own one-paragraph
description of every service, next to that service's shims. It is a
per-module prose site that does not live in the module, so a sweep
scoped to the module cannot reach it.
The counts in both comments were checked and are correct as written:
the matrix block has six shims and the nats block five. The `*.package`
entries append to the separate block further down, which already has a
comment saying what it is, so no count moved. Each fix is one clause.
`package` (the daemon build) and `swarmctlPackage` (the operator CLI
installed beside it) are host decisions, so they join the
`deploy.swarm-controller` block that already holds `enable`, the socket
path and the three local-disk credentials.
⚠️ The target namespace is spelled out in the shim comment because
getting it wrong here does not fail. `deploy.hive-controller` also
exists — it is hive-c0re's, carrying `tls.*` and `statusPublish.*`, and
is referenced from four modules. A rename pointing at it would land on
a live, unrelated option and evaluate cleanly. The module already warned
about this above its own deploy block; the warning now sits next to the
entries it governs too.
`swarm-ui.nix`'s package description cited `swarm.controller.package` as
the precedent for "wired from flake.nix rather than defaulted to pkgs".
That was true when written and stops being true here, so it moves with
it — it lives in a module this slice already finished, which is exactly
where a per-module sweep stops looking.
These were the last two `services.hyperhive.swarm.*` writers in
flake.nix; that file now has none.
Riding along, and not a defect fix: `controllerOldPath` spelled `enable`
the NEW way while every other entry used the old one, so six of the
seven controller shims were exercised and the seventh was not. It is
named for testing old paths, so it now uses the old path throughout.
Unlike the nats fixture there was no comment claiming otherwise — this
is coverage the fixture was already shaped to provide.
`package` (which authelia build runs in the container) and
`bridgePackage` (which `swarm-authelia-bridge` build writes the users
file) are both host decisions, so they join the `deploy.authelia` block
that already holds `enable`, `usersFile` and the two host-side paths.
Six readers, and one of them is why this commit touches a second module:
`swarm-controller.nix` reads `autheliaCfg.package` for
`SWARMCTL_AUTHELIA_BIN`. A per-module sweep never sees that — it is a
cross-reference between two movers, and whichever landed second would
have inherited a dangling alias. The alias itself stays: it still has
three live uses for `url` and `bridgeUrl`.
Two comments needed real edits rather than a rename:
- The `deploy.authelia` header said "all three of these are the same
kind of thing — a filesystem path". It is five options now and two
are packages, so the sentence distinguishes the builds from the
paths, and keeps "a hive that does not run it has nothing at any of
those paths" scoped to the paths it is true of.
- `module-eval.nix` said "Only `usersFile` has a rename entry",
explaining that authelia's other slice-7 movers are `readOnly` and a
rename module cannot contribute a definition to those. That reasoning
is worth keeping, but the claim stopped being true one line above the
two entries this commit adds.
The fixture gains both old paths; no new case. `autheliaOldPath`'s
existing arms already fail the eval if either shim goes missing.
`natsOldPath` exists to prove every old `swarm.nats.*` path still
resolves through its rename shim, and it defined six of the seven —
`enable` was spelled the NEW way (`deploy.nats.enable`), so nothing in
the suite ever exercised that shim. Deleting it would have gone
unnoticed.
The comment above the fixture claimed more than the fixture delivered:
"dropping any single shim entry fails the eval" reads as a guarantee
over all seven. Closing the gap rather than narrowing the sentence, so
the claim becomes true instead of merely careful.
`autheliaOldPath` keeps `deploy.nats.enable` deliberately — it needs the
queue on, it is not an old-path test for nats — and the gate controls
both spellings so a future edit cannot silently swap them.
Found by asking the fixture whether a planned mutation could fail before
spending an eval on it: no fixture defined the old path, so the arm was
guaranteed to survive. A grep answered in a second what a mutation would
have taken ~6 minutes to report.
`swarm.nats.authPackage` is the callout responder's build — a host
decision like every other package in this slice — so it joins the
`deploy.nats` block that already holds `enable`, the seed paths and the
auto-mint toggle. Its one reader is the `ExecStart` in this file.
Three comments moved with it, and none of them survived the move
unedited:
- The "deliberately NO `package` option" note now says *anywhere* —
it argued against a `package` under `swarm.nats`, and after the move
a reader's next question is why there isn't one under `deploy.nats`
either, where every other service's build now lives.
- The description's "see the note above" stopped resolving: the note
is ~80 lines up and in a different option block, so it names
`options.services.hyperhive.swarm.nats` explicitly.
- The split-rationale comment listed "the responder package" among
what the queue IS to every hive. It's what the host running it
decides, so it moved to that half of the sentence.
`flake.nix` writes this option (`lib.mkDefault`), so the writer is
repointed too — three of those left, all in the same block.
The eval fixture gains the old path rather than a new case: the shim
is what a dropped rename would break, and `natsOldPath` already fails
the eval if any single entry goes missing.
Same rule as the four movers before it. What makes this one different is
that `package` was the LAST option under `services.hyperhive.swarm.ui` —
`enable` moved in an earlier slice — so the namespace now holds nothing
and survives only as its two rename entries.
Two consequences worth naming:
flake.nix set `swarm.ui.package` directly. Left alone the shim would
have kept it working, which is exactly the trap: this repo's own flake
would then be setting the option through its own deprecation shim and
warning on every eval. It is repointed here. Four sibling assignments in
that file name movers this commit does not touch; they move with theirs.
The `cfg` alias is deleted. With the one code reader repointed it bound
nothing but an empty attrset — which still evaluates, so nothing would
have failed; it would just have sat there implying `swarm.ui` still held
something. Three comments naming `cfg.package` are repointed with it.
`uiOldPath` sets both old paths, so dropping either rename fails the
eval. Its case reads the vhost's rendered `root` rather than the option,
because that is the half a resolving-but-unwired shim would break.
Not touched: swarm-ui.nix's description says its default is wired "the
same way `swarm.controller.package` is", which is true until the
controller moves. It belongs to that commit, not this one.
Refs #3772.
Same rule as matrix and grafana: which build a service runs is a decision
of the host that runs it. Both stores already had a `deploy.<store>`
option for retention, so the package joins something rather than opening
a namespace.
The prose in both modules claimed the package as part of "what the store
IS from any hive's point of view" — a client hive needs the domain and
the port to reach a store, never the build it runs. deploy.nix's own
comment made the same claim about the pair and is corrected with them.
Separately, and the reason this commit adds a fixture rather than a line:
NEITHER STORE HAD AN OLD-PATH FIXTURE AT ALL. `swarm.victorialogs.` and
`swarm.victoriametrics.` had zero hits in module-eval.nix, so the
`enable` shims from the first slice and both `retentionPeriod` shims have
been uncovered since they landed — the suite would have gone green with
any of them deleted. That is precisely what the wireguard fixture's own
comment warns about: a missing shim reads as a clean tree and breaks
every existing operator config.
`storesOldPath` therefore sets all six old paths, not just the two this
commit moves. The case reads the package the CONTAINER renders rather
than the option, so a shim that resolves but stops reaching the module
fails too.
Refs #3772.
Same rule as the matrix pair: which build a service runs is a decision of
the host that runs it, so `package` follows `enable` out of the namespace
every hive must agree on byte for byte.
Grafana already had a `deploy.grafana` block holding the datasources, the
plugin list and the socket directory, so this is the package joining
options it belongs with rather than a new namespace.
Prose: the swarm-side header listed "its package" among what stays; the
deploy-side header listed what the host decides without it. Both edited,
which is the fix — a substitution on the first alone would have left the
second silently incomplete.
`swarm-grafana.nix:219` was on the sweep list and is NOT edited: it says
what stays above is what the service is to every hive, and that stays
true once the package leaves. The stale half was the sentence after it.
Refs #3772.
Which build a service runs is a decision of the host that runs it, so
`*.package` belongs with `enable` rather than in the namespace every hive
must agree on byte for byte.
The `gui` pair was the sharpest case and the reason this starts here:
`gui.enable` moved to `deploy.*` in an earlier slice while `gui.package`
stayed behind, so one web client's two halves sat in two namespaces —
and hive-matrix.nix said so, in a comment directly above the option.
The diff removes a stated inconsistency rather than introducing a rule.
Both old paths keep working through `mkRenamedOptionModule`; both are
settable (neither is `readOnly`), so both get a shim. The `matrixOldPath`
fixture in module-eval.nix now sets eight options through their
pre-rename paths instead of six, so dropping either new shim fails the
eval rather than silently dropping a definition.
Prose swept in the same commit: three comments claimed the package was
swarm-wide or named the pre-rename path, and docs/web-ui/dashboard.md
named `swarm.matrix.gui.package` two lines above
`deploy.matrix.gui.enable` — the same asymmetry this commit exists to
remove, displayed in the docs. A shim keeps the old spelling valid, so
no build, lint or broken-reference search would have found that line.
Refs #3772.
Cert auth answers a role, so nothing in the swarm can authenticate to
the store until some role exists. Creating the first one therefore
cannot itself use a certificate — the credential has to come from
outside that cycle, and an operator places it.
`deploy.bao.bootstrapTokenFile` names that token. A one-shot unit inside
the store's container reads it and writes the `swarm-controller` policy;
`local-defaults.nix` supplies the path (never the file) on an all-local
deploy, since co-location makes only the location derivable. `bao
operator init` stays an operator action in every shape.
The unit runs in the container rather than on the host because writing
the first grant is a store-side operation: it reaches the store locally
and needs no client certificate. Gating it on the store being here is
therefore not the co-location assumption glue-matrix-bao-token.nix
warns about — a reader has to work from anywhere, the first write never
does.
Policy only, deliberately: a cert-auth role binds a certificate and the
controller has no bao identity yet — it holds no leaf and contains no
bao code at all. The two certificates that do exist are both wrong to
bind. `clientCertFile` is the store host's own reader leaf rather than
the controller's, and the CA that signed it also signs every other
reader's, so binding that would let any reader authenticate as the
controller. Whoever gives the controller an identity writes the role.
The grants are scoped to `hive-*`. Without the prefix the controller
could rewrite the policy that constrains it.
Two things the module-eval arms pin: the unit renders inside the
container with the token path in both its script and its
ConditionPathExists, and a host that names a token while running no
store gets neither the unit nor the host-side directory.
The capabilities were derived with `bao write -output-policy` rather
than written from memory, and the setup.md commands were run against a
real binary for shape before being written down.
Refs #3726.
