test_compose asserts the new decnet.topology.* labels land on both base
deckies (role=base, no service marker) and service fragments
(service=true). The stub docker client in test_deploy grew a filters
kwarg so it keeps matching the real .networks.list(filters=...) call
signature now used by the deployer.
Legacy fleet deckies live in decnet-state.json; MazeNET topology
containers don't. Tag them at compose-time with
decnet.topology.service=true and let the collector match on that label.
Spin up the agent's log collector on the first successful /topology/apply
(not in the lifespan — that would break the no-docker-on-boot invariant)
and tear it down with the app. Land log lines in DECNET_AGENT_LOG_FILE,
separate from master-side DECNET_INGEST_LOG_FILE, so a dev box running
both roles can't forward its own ingest back to itself.
When master pushes a topology that differs from whatever is pinned
locally, teardown the predecessor and accept the new one. Refusing with
409 left the agent stranded after partial deploys. record_error now
persists the hydrated blob so a later teardown can still walk the LAN
list — otherwise a half-failed apply strands containers + bridges with
no breadcrumb back to them.
Two small observability follow-ups to the phase-1 agent/topology wiring:
TopologySummary now carries needs_resync so operators can see the
heartbeat's resync flag via the topology list/detail API without
dropping into the DB.
TopologyStore.record_error becomes an upsert — when a docker/compose
failure fires during the first materialise (put() never reached), we
still land a marker row so GET /topology/state surfaces the error and
the next heartbeat carries an empty applied_version_hash. That empty
hash is what master's heartbeat check relies on to flag the topology
for resync instead of assuming the apply succeeded.
Four regression tests guarding Step 8 of the agent/topology wiring:
- Lifespan startup must not call docker.from_env even with a populated
topology.db — replace docker with a boom-stub and assert zero calls.
- GET /topology/state returns the cached row verbatim without
re-materialising bridges/containers; live observation is read-only.
- Static guard: TopologyStore must not grow a restore/replay/reapply
method without someone re-reading the module docstring.
- Raw sqlite read + a second TopologyStore instance confirm the store
is passive — nothing scrubs stale rows on open, which is the
behaviour master's resync flow depends on.
Agent heartbeats now carry an applied-topology snapshot. The master
heartbeat handler compares the reported version_hash against what
canonical_hash yields for the hydrated topology pinned to that host
and flags Topology.needs_resync on divergence (or when the agent
reports no topology at all while master expects one).
The mutator watch loop gains reconcile_agent_resyncs, which re-pushes
the current hydrated blob via AgentClient.apply_topology without
touching status, then clears the flag on success. Push failures leave
the flag set so the next tick retries.
deploy_topology and teardown_topology now branch on
target_host_uuid. When set:
- Hydrate the topology locally (validator runs exactly as before).
- Compute canonical_hash; push {hydrated, version_hash} to the
pinned agent through AgentClient.apply_topology.
- Status machine still moves PENDING -> DEPLOYING -> ACTIVE on 2xx,
PENDING -> DEPLOYING -> FAILED on error; master remains the sole
owner of the row.
Teardown flips to TEARING_DOWN, fires /topology/teardown, then
TORN_DOWN — we log a warning on agent error but still settle to
TORN_DOWN so operators can delete the row (agent garbage is cleaned
on the next re-enroll).
Unihost deploys are unchanged — the field defaults to NULL so every
existing flow takes the local path.
Step 6 of the agent <-> topology integration.
Three new RPCs mirroring the existing deploy/teardown/status pattern:
- apply_topology(hydrated, version_hash) — long-timeout (600s) for
image pulls + compose up.
- teardown_topology(topology_id) — 300s timeout; enough for a
stubborn compose-down without hanging a heartbeat.
- get_topology_state() — short control-plane read for reconcile.
The per-call timeout swap uses the same trick as .deploy().
Step 5 of the agent <-> topology integration.
New mTLS-protected routes on the agent:
- POST /topology/apply — master pushes {hydrated, version_hash}.
Validates the hash matches locally (serialisation drift guard),
runs the topology through the same validator/composer pipeline
used master-side, then creates bridges + compose up + records the
apply in topology.db.
- POST /topology/teardown — dismantles compose, removes bridges,
clears topology.db. Idempotent.
- GET /topology/state — returns applied row + live docker
observation for the heartbeat.
Implementation lives in decnet/agent/topology_ops.py; it reuses the
private compose helpers from decnet.engine.deployer so we don't
duplicate compose/project-name plumbing. The apply path is sync
under the hood (docker SDK + subprocess); we hop to a thread so the
event loop keeps servicing other agent traffic.
v1 is one-topology-per-agent; cross-topology apply returns 409.
Step 4 of the agent <-> topology integration.
Single-row sqlite tracking which topology the agent last applied and
its version hash. Sync/stdlib, same pattern as the log-forwarder
offset store. v1 is one-topology-per-agent; attempting to apply a
different topology over a populated row raises AlreadyApplied so the
endpoint can return 409. observed() snapshots live docker state
(decnet-topology-* bridges + decnet-* containers) for the heartbeat.
The store is a cache, not authority — no auto-restore on boot.
Master remains the only source of truth.
Step 3 of the agent <-> topology integration.
Tiny pure helper both master and agent will use to answer "is the
applied state the one we expect?". SHA-256 of canonical JSON with
volatile keys (timestamps, status, version, canvas x/y/w/h) stripped
so the hash only captures deployment-relevant state.
Step 2 of the agent <-> topology integration.
Adds the `target_host_uuid` FK on `Topology` plus wiring through the
two create endpoints (`POST /topologies`, `POST /topologies/blank`).
Validates the mode/host pair: `mode='agent'` now requires a known,
routable host; `mode='unihost'` must leave the field unset.
Surfaced on `TopologySummary` so list/detail responses expose it.
Purely additive at the schema level — existing unihost flows unchanged
(field defaults to `NULL`).
Step 1 of the agent <-> topology integration.
GET /api/v1/topologies — paginated list with status filter. Extends
repo.list_topologies() to accept limit/offset and adds count_topologies()
for the total envelope field.
GET /api/v1/topologies/{id} — hydrated TopologyDetail; 404 if missing.
GET /api/v1/topologies/{id}/status-events — audit trail, limit-capped.
Catalog helpers for the phase-4 canvas UI:
* GET /topologies/services — full service catalog.
* GET /topologies/next-subnet?base=172.20 — wraps SubnetAllocator against
reserved_subnets across non-torn-down topologies.
* GET /topologies/{id}/lans/{lan_id}/next-ip — IPAllocator pre-seeded
with existing decky IPs in that LAN.
All read routes are viewer-or-admin. Sub-routers are included in an
order that keeps literal catalog paths (/services, /next-subnet) from
being shadowed by the /{topology_id} trie branch.
Add Pydantic DTOs in decnet/web/db/models.py covering every phase-3
endpoint shape: TopologyGenerateRequest, TopologySummary/Detail, child
create/update requests, MutationEnqueueRequest (Literal op guard),
MutationRow with JSON-payload decoder, validation/version/not-editable
error envelopes, and the three catalog responses.
Create decnet/web/router/topology/ as an import-safe package exporting
topology_router (prefix /topologies) — sub-routers land step-by-step in
subsequent commits. Mount under the main api router alongside swarm_mgmt.
tests/api/topology/test_models.py pins repo-dict ↔ DTO parity so future
repo-row drift breaks the contract test before the endpoints.
Adds the live-mutation pipeline for active/degraded topologies:
* TopologyMutation table with composite index (state, topology_id)
so the watch-loop guard query stays O(log n).
* claim_next_mutation is a single atomic UPDATE ... WHERE
state='pending' so racing reconcilers deterministically pick one
winner; losers see rowcount=0 and skip.
* reconcile_topologies drains pending rows per live topology, applies
via decnet.mutator.ops.dispatch, and on failure marks the mutation
failed + transitions topology to degraded.
* run_watch_loop gains a gated branch: flat-fleet mutate_all runs
every tick unchanged; the reconciler only enters when the cheap
has_pending_topology_mutation guard returns True.
* apply_* ops re-check hard invariants (names, IP collisions, subnet
overlap, known services, service_config shape) after every mutation
so the repo never lands in an invalid state.
* CLI: 'decnet topology mutate' / 'mutations' subcommands.
MazeNET phase 2 step 6. Equips the repo layer with the CRUD the web
editor needs before deploy.
- TopologyNotEditable exception: raised when a pending-only method hits
a non-pending topology. The intent is "free-form edits stop at deploy;
the mutator (step 7) takes over for live topologies."
- _assert_pending helper checks status inside the session.
- update_lan / update_topology_decky accept enforce_pending=True for
pre-deploy callers (existing internal callers default to False so
behavior is unchanged).
- delete_lan: cascades edges; refuses if any decky has only one edge
(= this LAN is its home) to prevent orphans.
- delete_topology_decky: cascades edges.
- delete_topology_edge: bare-bones removal.
All four mutators accept expected_version for optimistic concurrency.
Existing tests continue to pass (no behavior change for persist/deploy).
MazeNET phase 2 step 5. Pure storage — the generator emits None for
x/y and the web canvas fills them in later. No logic changes; no
compose, deploy, or validator impact.
MazeNET phase 2 step 4. Readies the repo layer for concurrent editors
(web canvas + CLI + mutator) without lost-write races.
- Topology.version: monotonically bumped on supervised child-row writes.
- VersionConflict exception carries {current, expected} for the UI.
- _check_and_bump_version helper reads Topology in the same session,
compares against expected_version, raises on mismatch, bumps on match.
Commit happens in the caller's existing transaction so check+bump+write
are atomic per mutation.
- add_lan / update_lan / add_topology_decky / update_topology_decky /
add_topology_edge accept expected_version=None by default, preserving
every existing caller's behavior.
When expected_version is None, no check runs and version stays put —
internal callers (persist) that don't care about concurrency keep
working unchanged.
MazeNET phase 2 step 3. Blocks deploys of hand-authored topologies that
would fail mid-bring-up (orphan deckies, duplicate IPs, overlapping
subnets, unknown services) with a structured error list instead of a
docker error at startup.
Rules (one function each, composable by the editor for inline hints):
- exactly one DMZ
- every LAN has a bridge chain to the DMZ (BFS via multi-homed deckies)
- no orphan deckies
- unique LAN and decky names per topology
- no IP collisions + IPs inside their LAN's subnet
- no LAN subnet overlaps
- every service in decnet.fleet.all_service_names()
- service_config keys match the decky's declared services
deploy_topology runs the validator after hydrate, before any status
transition or Docker call; errors raise ValidationError and status
stays at pending.
MazeNET phase 2 step 2. Mirrors the flat-fleet service_config pattern
(DeckyConfig.service_config → composer → svc.compose_fragment) into the
topology compose pipeline, so a hand-authored decky can carry overrides
like {"ssh": {"password": "megapassword"}} and the ssh fragment reads
them just like the flat path does.
- _PlannedDecky gains service_config: dict[str, dict].
- persist() stores it under decky_config["service_config"].
- topology/compose.py passes cfg.get("service_config", {}).get(svc, {})
to svc.compose_fragment(service_cfg=...).
Schema unchanged — service_config lives inside the existing
decky_config JSON blob. Zero changes in decnet/services/*.
MazeNET phase 2 step 1. Pulls inline IP/subnet allocation out of the
generator into decnet/topology/allocator.py so the editor + reconciler
can reuse the same primitives without duplicating logic.
- IPAllocator: stateful host-IP handout with reserve/release/is_free.
- SubnetAllocator: /24 handout under a base prefix, skips reservations.
- reserved_subnets(repo): collects claimed subnets across every
non-torn_down topology so concurrent drafts cannot collide.
- generate() accepts reserved_subnets= to skip existing claims.
Generator output is byte-identical under seed (behavior preserved).
Covers dry-run compose emission (no status change), FAILED transition
with reason logged on daemon errors, teardown from FAILED, and a
live-marked end-to-end test that creates/removes bridge networks
against a real docker daemon (skipped on CI).
Adds per-topology compose generation (one Docker bridge network per
LAN, multi-homed bridge deckies, ip_forward sysctl for L3 forwarders)
plus async deploy_topology/teardown_topology in the engine. Leaf-first
teardown via BFS-named LAN reverse sort; partial-state safe on failure.
Adds decnet/topology/ with:
- config.TopologyConfig: pydantic model driving generation (depth,
branching_factor, deckies_per_lan_min/max, bridge_forward_probability,
cross_edge_probability, subnet_base_prefix, service selection, seed).
Emits GeneratedTopology dataclass (lans, deckies, edges).
- status.TopologyStatus + assert_transition: seven-state machine with
an explicit legal-transition table. torn_down is terminal; degraded
is schema-reserved for future Healer use.
- generator.generate: deterministic DAG generation under config.seed.
Builds a tree of LANs (DMZ at root), plants deckies in each LAN,
promotes one decky per non-DMZ LAN to a parent bridge, and rolls
cross-edges per cross_edge_probability for DAG shape.
- persistence: persist() writes a plan to the repo as pending;
transition_status() enforces state-machine legality; hydrate() loads
topology + children into a single dict.
Covered by tests/topology/{test_status,test_generator,test_persistence}.
Adds topology CRUD to BaseRepository (NotImplementedError defaults) and
implements them in SQLModelRepository: create/get/list/delete topologies,
add/update/list LANs and TopologyDeckies, add/list edges, plus an atomic
update_topology_status that appends a TopologyStatusEvent in the same
transaction. Cascade delete sweeps children before the topology row.
Covered by tests/topology/test_repo.py (roundtrip, per-topology name
uniqueness, status event log, cascade delete, status filter) and an
extension to tests/test_base_repo.py for the NotImplementedError surface.
- tests/**: update templates/ → decnet/templates/ paths after module move
- tests/mysql_spinup.sh: use root:root and asyncmy driver
- tests/test_auto_spawn.py: patch decnet.cli.utils._pid_dir (package split)
- tests/test_cli.py: set DECNET_MODE=master in api-command tests
- tests/stress/conftest.py: run locust out-of-process via its CLI + CSV
stats shim to avoid urllib3 RecursionError from late gevent monkey-patch;
raise uvicorn startup timeout to 60s, accept 401 from auth-gated health,
strip inherited DECNET_* env, surface stderr on 0-request runs
- tests/stress/test_stress.py: loosen baseline thresholds to match hw
The 1,878-line cli.py held every Typer command plus process/HTTP helpers
and mode-gating logic. Split into one module per command using a
register(app) pattern so submodules never import app at module scope,
eliminating circular-import risk.
- utils.py: process helpers, _http_request, _kill_all_services, console, log
- gating.py: MASTER_ONLY_* sets, _require_master_mode, _gate_commands_by_mode
- deploy.py: deploy + _deploy_swarm (tightly coupled)
- lifecycle.py: status, teardown, redeploy
- workers.py: probe, collect, mutate, correlate
- inventory.py, swarm.py, db.py, and one file per remaining command
__init__.py calls register(app) on each module then runs the mode gate
last, and re-exports the private symbols tests patch against
(_db_reset_mysql_async, _kill_all_services, _require_master_mode, etc.).
Test patches retargeted to the submodule where each name now resolves.
Enroll-bundle tarball test updated to assert decnet/cli/__init__.py.
No behavioral change.
Uvicorn's h11/httptools HTTP protocols don't populate scope['extensions']['tls'], so /swarm/heartbeat's per-request cert pinning was 403ing every call despite CERT_REQUIRED validating the cert at handshake. Patch RequestResponseCycle.__init__ on both protocol modules to read the peer cert off the asyncio transport and write DER bytes into scope['extensions']['tls']['client_cert_chain']. Importing the module from swarm_api.py auto-installs the patch in the swarmctl uvicorn worker before any request is served.
New decnet.agent.heartbeat asyncio loop wired into the agent FastAPI
lifespan. Every 30 s the worker POSTs executor.status() to the master's
/swarm/heartbeat with its DECNET_HOST_UUID for self-identity; the
existing agent mTLS bundle provides the client cert the master pins
against SwarmHost.client_cert_fingerprint.
start() is a silent no-op when identity env (HOST_UUID, MASTER_HOST) is
unset or the worker bundle is missing, so dev runs and un-enrolled hosts
don't crash the agent app. On non-204 responses the loop logs loudly but
keeps ticking — an operator may re-enrol mid-session, and fail-closed
pinning shouldn't be self-silencing.
swarmctl CLI gains --tls/--cert/--key/--client-ca flags. With --tls the
controller runs uvicorn under HTTPS + mTLS (CERT_REQUIRED) so worker
heartbeats can reach it cross-host. Default is still 127.0.0.1 plaintext
for backwards compat with the master-CLI enrollment flow.
Auto-issue path (no --cert/--key given): a server cert signed by the
existing DECNET CA is issued once and parked under ~/.decnet/swarmctl/.
Workers already ship that CA's ca.crt from the enroll bundle, so they
verify the endpoint with no extra trust config. BYOC via --cert/--key
when the operator wants a publicly-trusted or externally-managed cert.
The auto-cert path is idempotent across restarts to keep a stable
fingerprint for any long-lived mTLS sessions.
New POST /swarm/heartbeat on the swarm controller. Workers post every
~30s with the output of executor.status(); the master bumps
SwarmHost.last_heartbeat and re-upserts each DeckyShard with a fresh
DeckyConfig snapshot and runtime-derived state (running/degraded).
Security: CA-signed mTLS alone is not sufficient — a decommissioned
worker's still-valid cert could resurrect ghost shards. The endpoint
extracts the presented peer cert (primary: scope["extensions"]["tls"],
fallback: transport.get_extra_info("ssl_object")) and SHA-256-pins it
to the SwarmHost.client_cert_fingerprint stored for the claimed
host_uuid. Extraction is factored into _extract_peer_fingerprint so
tests can exercise both uvicorn scope shapes and the both-unavailable
fail-closed path without mocking uvicorn's TLS pipeline.
Adds get_swarm_host_by_fingerprint to the repo interface (SQLModel
impl reuses the indexed client_cert_fingerprint column).
The reaper was being SIGTERM'd mid-rm because `start_new_session=True`
only forks a new POSIX session — it does not escape decnet-agent.service's
cgroup. When the reaper ran `systemctl stop decnet-agent`, systemd
tore down the whole cgroup (reaper included) before `rm -rf /opt/decnet*`
finished, leaving the install on disk.
Spawn the reaper via `systemd-run --collect --unit decnet-reaper-<pid>`
so it runs in a fresh transient scope, outside the agent unit. Falls
back to bare Popen for non-systemd hosts.
Decommissioning a worker from the dashboard (or swarm controller) now
asks the agent to wipe its own install before the master forgets it.
The agent stops decky containers + every decnet-* systemd unit, then
deletes /opt/decnet*, /etc/systemd/system/decnet-*, /var/lib/decnet/*,
and /usr/local/bin/decnet*. Logs under /var/log are preserved.
The reaper runs as a detached /tmp script (start_new_session=True) so
it survives the agent process being killed. Self-destruct dispatch is
best-effort — a dead worker doesn't block master-side cleanup.
Teardowns were synchronous all the way through: POST blocked on the
worker's docker-compose-down cycle (seconds to minutes), the frontend
locked tearingDown to a single string so only one button could be armed
at a time, and operators couldn't queue a second teardown until the
first returned. On a flaky worker that meant staring at a spinner for
the whole RTT.
Backend: POST /swarm/hosts/{uuid}/teardown returns 202 the instant the
request is validated. Affected shards flip to state='tearing_down'
synchronously before the response so the UI reflects progress
immediately, then the actual AgentClient call + DB cleanup run in an
asyncio.create_task (tracked in a module-level set to survive GC and
to be drainable by tests). On failure the shard flips to
'teardown_failed' with the error recorded — nothing is re-raised,
since there's no caller to catch it.
Frontend: swap tearingDown / decommissioning from 'string | null' to
'Set<string>'. Each button tracks its own in-flight state; the poll
loop picks up the final shard state from the backend. Multiple
teardowns can now be queued without blocking each other.
Submitting an INI with a single [decky1] was silently redeploying the
deckies from the *previous* deploy too. POST /deckies/deploy merged the
new INI into the stored DecnetConfig by name, so a 1-decky INI on top of
a prior 3-decky run still pushed 3 deckies to the worker. Those stale
decky2/decky3 kept their old IPs, collided on the parent NIC, and the
agent failed with 'Address already in use' — the deploy the operator
never asked for.
The INI is the source of truth for which deckies exist this deploy.
Full replace: config.deckies = list(new_decky_configs). Operators who
want to add more deckies should list them all in the INI.
Update the deploy-limit test to reflect the new replace semantics, and
add a regression test asserting prior state is dropped.
docker compose up is partial-success-friendly — a build failure on one
service doesn't roll back the others. But the master was catching the
agent's 500 and tagging every decky in the shard as 'failed' with the
same error message. From the UI that looked like all three deckies died
even though two were live on the worker.
On dispatch exception, probe the agent's /status to learn which deckies
actually have running containers, and upsert per-decky state accordingly.
Only fall back to marking the whole shard failed if the status probe
itself is unreachable.
Enhance agent.executor.status() to include a 'runtime' map keyed by
decky name with per-service container state, so the master has something
concrete to consult.
Two compounding root causes produced the recurring 'Address already in use'
error on redeploy:
1. _ensure_network only compared driver+name; if a prior deploy's IPAM
pool drifted (different subnet/gateway/range), Docker kept handing out
addresses from the old pool and raced the real LAN. Now also compares
Subnet/Gateway/IPRange and rebuilds on drift.
2. A prior half-failed 'up' could leave containers still holding the IPs
and ports the new run wants. Run 'compose down --remove-orphans' as a
best-effort pre-up cleanup so IPAM starts from a clean state.
Also surface docker compose stderr to the structured log on failure so
the agent's journal captures Docker's actual message (which IP, which
port) instead of just the exit code.
The nested list-comp `[f"{id}-{svc}" for svc in [d.services for d ...]]`
iterated over a list of lists, so `svc` was the whole services list and
f-string stringified it -> `decky3-['sip']`. docker compose saw "no such
service" and the per-decky teardown failed 500.
Flatten: find the matching decky once, then iterate its services. Noop
early on unknown decky_id and on empty service lists. Regression test
asserts the emitted compose args have no '[' or quote characters.
Agents already exposed POST /teardown; the master was missing the plumbing
to reach it. Add:
- POST /api/v1/swarm/hosts/{uuid}/teardown — admin-gated. Body
{decky_id: str|null}: null tears the whole host, a value tears one decky.
On worker failure the master returns 502 and leaves DB shards intact so
master and agent stay aligned.
- BaseRepository.delete_decky_shard(name) + sqlmodel impl for per-decky
cleanup after a single-decky teardown.
- SwarmHosts page: "Teardown all" button (keeps host enrolled).
- SwarmDeckies page: per-row "Teardown" button.
Also exclude setuptools' build/ staging dir from the enrollment tarball —
`pip install -e` on the master generates build/lib/decnet_web/node_modules
and the bundle walker was leaking it to agents. Align pyproject's bandit
exclude with the git-hook invocation so both skip decnet/templates/.
