Files
raphael 54626ceb6a
tests / build-and-test (push) Successful in 2m59s
tests / thread-sanitizer (push) Successful in 3m28s
tests / tshark-dissector (push) Successful in 2m22s
tests / secs4j-interop (push) Successful in 2m6s
tests / python-interop (push) Failing after 3m8s
tests / libfuzzer (push) Successful in 3m44s
feat(daemon): Phase E — production hardening, complete
Exposure: --grpc default flipped from 0.0.0.0 to 127.0.0.1 (the API is
unauthenticated by design; auth belongs to the transport), Unix-domain-
socket support (--grpc unix:///run/secs_gemd/api.sock = zero network
surface), SECURITY.md documents the contract and ch42 gained a "Running it
in production" section (which also documents the HSMS-SS single-session
assumption).

Graceful shutdown: SIGTERM/SIGINT land on an asio::signal_set on the io
thread, which nudges grpc Shutdown with a 2s deadline (cancels open
Subscribe/WatchHealth streams); Wait() returns on the MAIN thread, which
stops the engine (rt->stop() joins the io thread, so it must not run on
it). Exit 0, journal-safe, the in-code TODO is gone. --spool-dir added so
host-bound events survive daemon restarts.

Observability: --metrics serves Prometheus gauges secsgem_link_selected /
secsgem_control_state / secsgem_spool_depth, wired via the Phase-0
add_link_observer/add_control_state_observer hooks + io-thread sampling.

Deployment: deploy/secs_gemd.service — hardened systemd unit (DynamicUser,
ProtectSystem=strict, StateDirectory for the spool, UDS for the API,
TimeoutStopSec aligned with the graceful-shutdown window).

Enforcement: tools/check_daemon_ops.sh proves all three operational claims
(unix-socket gRPC accepts, all gauges present on /metrics, SIGTERM -> exit
0 + clean-stop log) — green; wired into tools/run_interop.sh (now 11
steps) and CI. CI python-interop lane also gained the pyclient and
spool-restart steps, so every harness now runs in CI.

TODO sweep: the shutdown TODO is fixed; the four remaining TODOs (nested
list formats, C2-as-text, U8>2^63, CONNECTED link state) are deliberate
deferred edge cases, each marked in code with context.

Daemon suite re-verified green (175 assertions).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-06-11 00:07:37 +02:00

153 lines
6.4 KiB
C++

// secs_gemd — the SECS/GEM daemon.
//
// Runs the GEM engine (EquipmentRuntime + the default handlers) on a background
// thread, owning the HSMS link to the host, and exposes the name-based gRPC
// API (proto/secsgem/v1/equipment.proto) so a tool's software — in any language
// — can drive the equipment without linking C++ or knowing SEMI.
//
// Operations (Phase E):
// --grpc defaults to 127.0.0.1:50051 — NEVER expose the unauthenticated
// API on the equipment LAN. For same-host tool software prefer a
// Unix domain socket: --grpc unix:///run/secs_gemd.sock
// --metrics Prometheus endpoint port (0 = disabled). Gauges: HSMS link
// state, control state, spool depth.
// SIGTERM/SIGINT: graceful shutdown — gRPC server drains (open streams are
// cancelled), then the engine stops cleanly, so a supervised
// restart (systemd / docker stop) never kills the spool journal
// mid-write. Exit code 0.
//
// Deployment recipe: deploy/secs_gemd.service (systemd) and docs/42.
#include <grpcpp/grpcpp.h>
#include <asio.hpp>
#include <chrono>
#include <csignal>
#include <cstdint>
#include <functional>
#include <iostream>
#include <memory>
#include <string>
#include "secsgem/daemon/equipment_service.hpp"
#include "secsgem/gem/default_handlers.hpp"
#include "secsgem/gem/runtime.hpp"
#include "secsgem/metrics/prometheus.hpp"
namespace gem = secsgem::gem;
namespace metrics = secsgem::metrics;
using namespace std::chrono_literals;
namespace {
std::string arg(int argc, char** argv, const std::string& key, const std::string& def) {
for (int i = 1; i + 1 < argc; ++i)
if (key == argv[i]) return argv[i + 1];
return def;
}
} // namespace
int main(int argc, char** argv) {
const std::string hsms_port = arg(argc, argv, "--port", "5000");
const std::string grpc_addr = arg(argc, argv, "--grpc", "127.0.0.1:50051");
const std::string cfgdir = arg(argc, argv, "--config-dir", "data");
const std::string spool_dir = arg(argc, argv, "--spool-dir", "");
const uint16_t metrics_port =
static_cast<uint16_t>(std::stoi(arg(argc, argv, "--metrics", "0")));
gem::EquipmentRuntime::Config cfg;
cfg.equipment_yaml = cfgdir + "/equipment.yaml";
cfg.control_state_yaml = cfgdir + "/control_state.yaml";
cfg.process_job_yaml = cfgdir + "/process_job_state.yaml";
cfg.control_job_yaml = cfgdir + "/control_job_state.yaml";
cfg.port = static_cast<uint16_t>(std::stoi(hsms_port));
cfg.spool_dir = spool_dir;
cfg.log = [](const std::string& m) { std::cout << "[gemd] " << m << std::endl; };
std::unique_ptr<gem::EquipmentRuntime> rt;
try {
rt = std::make_unique<gem::EquipmentRuntime>(cfg);
} catch (const std::exception& e) {
std::cerr << "[gemd] config error: " << e.what() << std::endl;
return 1;
}
gem::register_default_handlers(*rt);
// ---- observability (before run_async: observers must land first) -------
std::shared_ptr<metrics::Registry> registry;
std::shared_ptr<metrics::PrometheusServer> exporter;
std::shared_ptr<asio::steady_timer> gauge_timer;
if (metrics_port != 0) {
registry = std::make_shared<metrics::Registry>();
registry->describe("secsgem_link_selected",
"1 when an HSMS session is SELECTED, else 0",
metrics::MetricType::Gauge);
registry->describe("secsgem_control_state",
"E30 control state (0=EquipOffline 1=AttemptOnline "
"2=HostOffline 3=OnlineLocal 4=OnlineRemote)",
metrics::MetricType::Gauge);
registry->describe("secsgem_spool_depth", "Queued spool messages",
metrics::MetricType::Gauge);
rt->add_link_observer([registry](bool selected) {
registry->set_gauge("secsgem_link_selected", selected ? 1.0 : 0.0);
});
rt->add_control_state_observer(
[registry](gem::ControlState, gem::ControlState to, gem::ControlEvent) {
registry->set_gauge("secsgem_control_state", static_cast<double>(to));
});
registry->set_gauge("secsgem_link_selected", 0.0);
registry->set_gauge("secsgem_control_state",
static_cast<double>(rt->control_state()));
// Spool depth: sampled on the io thread (the model's owner).
gauge_timer = std::make_shared<asio::steady_timer>(rt->io());
auto tick = std::make_shared<std::function<void(std::error_code)>>();
*tick = [registry, gauge_timer, tick, raw = rt.get()](std::error_code ec) {
if (ec) return;
registry->set_gauge("secsgem_spool_depth",
static_cast<double>(raw->model().spool.size()));
gauge_timer->expires_after(5s);
gauge_timer->async_wait(*tick);
};
gauge_timer->expires_after(1s);
gauge_timer->async_wait(*tick);
exporter = std::make_shared<metrics::PrometheusServer>(
rt->io(), metrics_port, registry);
exporter->start();
}
// Construct the service before starting the io thread: its constructor
// snapshots the name->id/format maps and registers its observers.
secsgem::daemon::EquipmentService service(*rt);
rt->run_async(); // engine + HSMS link on a background thread
grpc::ServerBuilder builder;
builder.AddListeningPort(grpc_addr, grpc::InsecureServerCredentials());
builder.RegisterService(&service);
std::unique_ptr<grpc::Server> server(builder.BuildAndStart());
if (!server) {
std::cerr << "[gemd] failed to start gRPC server on " << grpc_addr << std::endl;
return 1;
}
// Graceful shutdown: the signal lands on the io thread, which only nudges
// the gRPC server; Wait() then returns on the MAIN thread, which tears the
// engine down (rt->stop() joins the io thread — it must not run on it).
// The deadline cancels open Subscribe/WatchHealth streams instead of
// waiting for clients to hang up.
asio::signal_set signals(rt->io(), SIGINT, SIGTERM);
signals.async_wait([&server, &rt](std::error_code ec, int signo) {
if (ec) return;
rt->log("signal " + std::to_string(signo) + ": shutting down");
server->Shutdown(std::chrono::system_clock::now() + 2s);
});
std::cout << "[gemd] gRPC on " << grpc_addr
<< (metrics_port ? "; metrics on :" + std::to_string(metrics_port) +
"/metrics"
: "")
<< "; HSMS equipment on :" << hsms_port << std::endl;
server->Wait();
rt->stop();
std::cout << "[gemd] stopped cleanly" << std::endl;
return 0;
}