Addresses review on #519 (@didericis 6143 + the fail-hard direction over the
codex skip).
Base owns the flow (6143 / ADR 0006). `attach_bottled_agents_to_gateway()` is
now a concrete method on `BottleBackend` that delegates to
`gateway_attach.reconcile_running_bottles`; backends override only three
primitives — `_gateway_attach_resources()`, `_running_bottles()`,
`_attach_bottle_to_gateway()`. The shared control flow + error policy live in
one place so a backend can't drift onto a bespoke loop or a silent skip. Keeps
the reconcile flow + `GatewayAttachResources` out of `backend/base.py` (the
size guardrail) in a dedicated `backend/gateway_attach.py`. New ADR 0006 records
the "shared behaviour in the base backend, subclasses override primitives"
theme.
Fail hard, never skip (the maintainer's direction over the codex review's
skip). Any attach failure now aborts bring-up instead of being logged and
skipped: a bottle that silently can't reach the fresh gateway (its egress just
starts failing TLS) is worse than a loud failure. Every bottle is attempted and
the failures are raised together (aggregate `InfraLaunchError`) so one bring-up
surfaces the full blast radius. Resource gathering (the CA fetch) also fails
hard. PRD 0081 goal + design updated to match (reversed from the earlier
"tolerate per-bottle failures" draft).
Bumps the base.py guardrail cap 580->600 for the new contract surface (the
delegator + 3 abstract primitives); the flow itself lives in gateway_attach.py.
Refs #516, #519.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Mechanical rename addressing review 6141 on #519: "consolidated launch" was
unclear about what it composes. Rename the per-backend module
`consolidated_launch.py` -> `infra_launch.py` and the error
`ConsolidatedLaunchError` -> `InfraLaunchError` across all three backends and
their callers/tests. Unify the three identical per-backend error classes into
one `InfraLaunchError` defined in `backend/base.py` (re-exported by each
`infra_launch`) so the base backend can raise and catch a single shared type —
groundwork for the base owning the gateway-attach flow.
Add a glossary entry defining **infra** = the per-host gateway + orchestrator
service pair every bottle attaches to.
No behavior change.
Refs #516, #519.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add `attach_bottled_agents_to_gateway()` as an `@abc.abstractmethod` on
`BottleBackend` and implement the first reconcile step across all three
backends: on a gateway cold boot, replace every already-running agent's
trusted CA with the freshly-minted gateway CA. Reconciling running bottles
is a backend responsibility (only the backend can enumerate its agents and
reach them — firecracker over SSH, docker/macOS over exec/cp), so making it
an abstract method keeps the fix cross-backend by construction.
The gateway rootfs is ephemeral, so a rebuild mints a new CA that every
running bottle distrusts (SSL verification fails — #510). This reconcile is
what distributes the fresh CA, so a routine gateway rebuild now doubles as a
free CA rotation. Per-bottle steps are best-effort: one unreachable or
malformed bottle is logged and skipped, never blocking the others.
Trigger — `Gateway.connect_to_orchestrator` now returns a cold-boot bool
(True when it actually (re)brought the gateway up, False when a healthy
current gateway was left untouched); each infra `ensure_running` gates the
reconcile on it, so it fires exactly on cold boot and never on an adopt.
Git-gate re-provision and egress-token restore fold into this same reconcile
on the same trigger in follow-up PRs (#516).
Refs #516. Addresses #510.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a backend-neutral `InfraService` ABC (backend/infra_service.py) so all three
backends present the same composer contract: `orchestrator()` / `gateway()`
accessors, `ensure_running(*, startup_timeout) -> str` (the host control-plane
URL), `stop()`, plus concrete `url()` / `is_healthy()` that delegate to the
orchestrator. `DockerInfraService` and `MacosInfraService` now subclass it.
Give firecracker the missing class: `FirecrackerInfraService`
(backend/firecracker/infra.py) wraps the `infra_vm` substrate as the pair
coordinator (adopt-or-boot-both under the singleton flock). `infra_vm` keeps only
the plane-agnostic substrate; its `ensure_running` / `_adopt` / `InfraEndpoint`
move to the class, and the coordinator helpers it now calls cross-module go
public (`singleton_lock` / `expected_version` / `adoptable` /
`record_booted_version`).
Unify the return type on the way: `ensure_running` returns the URL string
everywhere (was `str` for docker, two different `InfraEndpoint`s for
macOS/firecracker), and those `InfraEndpoint`s are deleted — the services are the
source of truth (callers read `gateway().address()` / `orchestrator().url()`).
docker's `url` property becomes the inherited `url()`. backend.py /
consolidated_launch / image_builder + tests updated.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Sweep for vestiges of the old combined-plane model and the pre-split shared
rootfs. Two are load-bearing, the rest are stale docs/comments:
- Bug: macOS `enumerate_active` only excluded the gateway container from the
agent list, so after the split the orchestrator container
(`bot-bottle-mac-orchestrator`, also `bot-bottle-`-prefixed) was enumerated as
a phantom agent. Exclude both infra containers; test covers it.
- Dead code: the gateway `bootstrap.py` still carried an `orchestrator` daemon
spec + `_OPT_IN_DAEMONS` + a signing-key/JWT env branch, all for the old
combined container where the gateway process could also run the control plane.
No backend ever requests it now — removed; the key-stripping stays as
defense-in-depth.
Stale-comment reframes: "the/single infra container" -> the orchestrator +
gateway pair (or the specific plane); "shared rootfs / bb_role init / one
published rootfs" -> the per-plane rootfs + `role_init`; the deleted
Dockerfile.infra references in Dockerfile.orchestrator/.gateway; and the macOS
"one infra container ... same address" docstring + its now-false
share-one-address test (the planes are distinct containers with distinct
addresses).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Brings in main's 8 commits — #470 (backend-agnostic CI guards: BackendStatus
enum, quiet status(), is_backend_available/is_backend_ready, tests/_backend.py
skip_unless_backend) plus firecracker status() readiness checks (kvm/kernel/
dropbear/mke2fs).
Reconciled #470's additions into the reorg'd backend structure:
* BackendStatus enum + the status(*, quiet=) ABC signature -> backend/base.py
* is_backend_available / is_backend_ready -> backend/selection.py
* exposed all three through the lazy backend facade (__init__).
Integration-test conflicts were our control_plane->orchestrator renames vs
main's skip_unless_docker -> skip_unless_backend swap: kept our refactored
import paths + main's guard. Repointed main's new is_backend_* tests to patch
selection._backends (our moved home) instead of the facade.
pyright: 0 errors. Full unit suite green (2272).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Add a module-level BackendStatus(IntEnum) with READY=0 to replace magic
0 comparisons. Extend BottleBackend.status() with a quiet parameter:
quiet=False (default) prints diagnostics; quiet=True returns the code
silently for programmatic checks.
Add is_backend_available() (cheap PATH check, alias for has_backend) and
is_backend_ready(name, *, quiet=False) (full status() check) to the
backend package for callers that need to distinguish availability from
readiness.
Update tests/_backend.py guards to use is_backend_ready(quiet=False) so
diagnostic output is printed during test discovery, giving the operator a
concrete reason for each skip rather than a bare "prerequisites
unavailable" message.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The codebase used "control plane" both as an architectural role term AND
as an identifier alias for the orchestrator component, producing
duplicate names for one thing (control_plane_url vs orchestrator_url,
CONTROL_PLANE_PORT, host_control_plane_token, …). Going forward the
concrete component is always named for what it is — Gateway or
Orchestrator — and the plane vocabulary is reserved for prose (module
descriptions, the security argument).
Renamed (identifiers + the in-repo env/wire/file string values, all
setters/getters are in this repo so the change is atomic):
ControlPlaneServer -> OrchestratorServer
control_plane_url -> orchestrator_url
probe_control_plane_url -> probe_orchestrator_url
host_control_plane_token -> host_orchestrator_token
CONTROL_PLANE_PORT -> ORCHESTRATOR_PORT
CONTROL_PLANE_TOKEN_ENV/FILE -> ORCHESTRATOR_TOKEN_ENV/FILENAME
BOT_BOTTLE_CONTROL_PLANE_TOKEN-> BOT_BOTTLE_ORCHESTRATOR_TOKEN
control-plane-token (file) -> orchestrator-token
control_auth (module) -> orchestrator_auth (stays top-level;
the gateway imports it and must not
import the orchestrator/ package)
CONTROL_AUTH_HEADER -> ORCHESTRATOR_AUTH_HEADER
x-bot-bottle-control-auth -> x-bot-bottle-orchestrator-auth
CONTROL_AUTH_JWT_ENV -> ORCHESTRATOR_AUTH_JWT_ENV
BOT_BOTTLE_CONTROL_AUTH_JWT -> BOT_BOTTLE_ORCHESTRATOR_AUTH_JWT
_control_auth_headers -> _orchestrator_auth_headers
Prose plane-terms ("control plane", "data plane") are preserved,
including the test name test_data_plane_daemons_get_jwt_not_key (it
names the security invariant). Gateway and orchestrator verified to
agree on the renamed wire header; full unit suite green (2243).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Give each service its own store package + manager, and cut the supervise module
along the control/data-plane boundary so nothing in the shared layer reaches up
into the orchestrator.
Stores, by owner:
- bot_bottle/store/ keeps only the shared base (DbStore, migrations) and the
concrete stores that aren't service-owned (audit_store, config_store).
- bot_bottle/orchestrator/store/ now houses the orchestrator-owned stores —
queue_store (supervise queue), secret_store, config_store — plus a new
orchestrator store_manager that migrates them (composing audit/config
downward from the base). The old shared store_manager is gone.
Supervise plane, by tier:
- bot_bottle/supervisor/ (NEUTRAL, importable by every tier including the
gateway): types.py (the Proposal/Response/AuditEntry wire types + the tool/
status/poll constants + the shared daemon constants moved out of
supervise.py) and plan.py (SupervisePlan, a pure DTO).
- bot_bottle/orchestrator/supervisor/ (orchestrator-only): queue.py (the
queue/audit I/O wrappers + render_diff + sha256_hex) and supervise.py (the
Supervise lifecycle that stages the DB via the store manager). Its __init__
re-exports the neutral vocabulary so orchestrator-side callers import from
one place.
The gateway now imports only bot_bottle.supervisor.types (never
bot_bottle.supervise), so the data plane holds no code dependency on the
orchestrator — it reaches the queue over the control-plane RPC. This removes
the circular import that moving queue_store under orchestrator introduced
(supervise -> orchestrator -> service -> supervise).
supervise_types.py -> supervisor/types.py; supervise.py deleted (split). Full
unit suite green (2251).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds the `nested_containers` bottle flag. On the macOS backend it starts
a rootless podman service inside the bottle and exposes its
Docker-compatible API socket, so the agent still runs `docker` and
`docker compose`. No host daemon socket is mounted and the guest gains
no capabilities; backends that cannot run a guest-local engine reject
the flag in the shared prepare template rather than ignoring it.
Podman rather than rootless Docker because Apple Container's capability
bounding set omits CAP_SYS_ADMIN, which the kernel requires to write a
multi-range uid_map via newuidmap. With no subordinate UID range podman
falls back to a single-UID self-mapping an unprivileged process may
write itself, so the image build strips /etc/subuid and /etc/subgid
entries rather than adding them.
That mapping is also why nested containers are not an isolation layer:
root inside one is the agent user outside it. They are a build/test
convenience; the bottle remains the boundary.
Ports the spike branch onto main, renaming docker_access — it implied
Docker and granted access to nothing on the host — and drops podman
from the derived layer now that every built-in image ships it (#451).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Addresses review comments 3098, 3099, 3100 on PR #336:
- Add BottleImages(agent, sidecar) dataclass to backend/__init__.py.
Docker/macOS backends use str image refs; smolmachines uses Path
artifacts. Replaces the singular `image` variable from the canonical
pattern in comment 3100.
- Replace _image_stale_checks/skip_stale with public prelaunch_checks().
CLI now calls backend.prelaunch_checks(plan) before backend.launch(plan);
if StaleImageError is raised and the operator confirms, launch proceeds
without re-checking. Removes the while-True/skip_stale retry loop.
- Add abstract _build_or_load_images(plan) -> BottleImages to
BottleBackend. launch() calls it then passes images to _launch_impl.
Each backend implements both methods.
- Fix comment 3098 (macos-container): _build_images is removed.
build_or_load_images() has separate fresh/cached code paths — the
cached path never calls a build helper.
- Update _start_bundle (smolmachines) to accept sidecar_artifact: Path
directly. Sidecar artifact resolution moves to _sidecar_from_path(),
called by build_or_load_images alongside _agent_from_path().
- image_cache: StaleImageError exception + check_stale/check_stale_path (raise instead of warn)
- BottleBackend.launch: template method (skip_stale flag) that calls _image_stale_checks then _launch_impl
- Each backend: _image_stale_checks delegates to a stale_checks() function in its launch module; _launch_impl replaces launch override
- macos_container: adds image_created_at to util, cached-image support in _build_images, stale_checks
- cli/start.py: catches StaleImageError, prompts interactively, retries with skip_stale=True; headless mode dies on it
Adopts the firecracker infra-VM pattern for macOS: the orchestrator control
plane and the gateway data plane now run in a SINGLE Apple container instead of
two. Apple Containers are lightweight VMs with separate kernels, so the prior
two-container design had both guests writing one bot-bottle.db over virtiofs,
where fcntl locks are not coherent across kernels — concurrent writes (the
orchestrator's registry vs the gateway supervise daemon's queue) could corrupt
it. One container = one kernel = coherent locking.
The DB moves onto a container-only Apple volume (bot-bottle-mac-db), never
bind-mounted from the host, so no host process opens the live file either. The
host CLI already reaches registry + supervise state over the control-plane HTTP
surface (cli/supervise.py uses OrchestratorClient), exactly as firecracker's
VM-only DB requires.
Two simplifications fall out of the single container:
- No DNS dance: the control plane and gateway daemons reach each other over
127.0.0.1, so the orchestrator-before-gateway ordering (a workaround for
Apple having no container DNS) is gone, along with the moved-IP recreate
logic it needed.
- Net -243 lines.
Mechanics: the infra container runs from the gateway image with the
control-plane source bind-mounted read-only (like the docker orchestrator, so a
code change needs no rebuild) and a small sh -c init that starts both processes
(mirrors firecracker's _infra_init). Also implements the macOS backend's
ensure_orchestrator() and adds it to discover_orchestrator_url, so operator
tools (supervise) can bring up / find the control plane on demand — previously
the macOS backend died with "no orchestrator control plane".
Verified end-to-end on real Apple Container 1.0.0: the single infra container
comes up healthy (one address for control plane + gateway), both processes run,
the DB is written on the container-only volume, host-side supervise works over
HTTP, and a registered agent gets 200 for an allowed host / 403 for a denied
one. 1824 unit tests pass with `container` absent (CI parity), pyright clean,
pylint 9.89.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Add an abstract teardown() classmethod to BottleBackend — the inverse of
setup(), surfaced as `./cli.py backend teardown [--backend=NAME]`
(uninstall). Symmetric with setup: it prints the privileged commands /
declarative config change to remove the host prerequisites.
- firecracker: NixOS-aware — disable the flake module (or drop the
import) and rebuild, or `firecracker-netpool.sh down` imperatively.
- docker / macos-container: nothing to undo (no privileged host state);
print a short note.
Not called by the launch path or the test suite. Extends test_cli_backend
for the new dispatch.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WBMWTEtQdJ4W5UrWuLHCck
Add abstract `setup()`/`status()` classmethods to BottleBackend (same
per-host, no-instance shape as is_available) so host provisioning is
part of the backend contract, not a per-backend command. Replace the
`./cli.py firecracker {setup,status}` command with a generic
`./cli.py backend {setup,status} [--backend=NAME]` that resolves a
backend (flag / $BOT_BOTTLE_BACKEND / host default) and dispatches —
swapping backends is just a different --backend.
Implementations:
- firecracker: moved out of cli/ into backend/firecracker/setup.py
(network pool module/script + range-overlap check), unchanged output.
- docker: new backend/docker/setup.py — reports docker on PATH, daemon
reachability, and gVisor runsc; setup notes no privileged pool is
needed. Minimal placeholder; richer version tracked in #345.
- macos-container: new setup.py — container CLI + system-service checks.
Also retarget the launch-preflight / isolation-probe pointers to the new
command. New test_cli_backend covers dispatch + docker setup/status.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01WBMWTEtQdJ4W5UrWuLHCck
BottleSpec.manifest was ManifestIndex | Manifest — a union encoding
two lifecycle stages in one field. The union was unjustifiable:
it forced a type-narrowing workaround (loaded_manifest property)
on every consumer.
Clean split:
- BottleSpec.manifest: ManifestIndex (always; CLI-supplied intent)
- BottlePlan.manifest: Manifest (always; loaded by _validate())
_validate() returns the loaded Manifest directly. prepare() passes
it to _resolve_plan(), which stores it on the plan. All provisioner
code now reads plan.manifest.agent / plan.manifest.bottle — no
union, no asserts, no type: ignore.
didericis-claude
didericis
didericis-codex