feat: ship fenris-monitor helper, polkit policy, and systemd units (#29)

src/fenris/collect.py

@@ -0,0 +1,132 @@
+#!/usr/bin/env python3
+"""fenris-collect: device interrogation and store writes.
+
+This is the root oneshot unit's ExecStart. It reads the device selector
+from /etc/fenris/fenris.conf, interrogates the drive via smartctl and sysfs,
+and writes the sample to the observation store.
+
+Spec: §8.4, §8.5
+
+When run as a script, uses the fenris package from the installed wheel.
+"""
+import json
+import subprocess
+import sys
+from datetime import datetime, timezone
+from pathlib import Path
+
+# Add the venv to path if running from the installed location
+VENV_DIR = Path("/opt/fenris")
+if VENV_DIR.exists():
+    site_packages = next((VENV_DIR / "lib").glob("python*/site-packages"), None)
+    if site_packages:
+        sys.path.insert(0, str(site_packages))
+
+from fenris.store import init_store, get_store_path
+from fenris.collector import run_collection
+
+
+CONFIG_PATH = Path("/etc/fenris/fenris.conf")
+
+
+def load_config() -> dict:
+    """Load configuration from /etc/fenris/fenris.conf.
+
+    The file holds exactly one key: the device selector.
+    Spec §8.3: re-read every run; no reload path.
+    """
+    if not CONFIG_PATH.exists():
+        raise RuntimeError(f"Configuration file not found: {CONFIG_PATH}")
+
+    config = {}
+    try:
+        with open(CONFIG_PATH, "r") as f:
+            for line in f:
+                line = line.strip()
+                if not line or line.startswith("#"):
+                    continue
+                if "=" in line:
+                    key, value = line.split("=", 1)
+                    config[key.strip()] = value.strip()
+    except Exception as e:
+        raise RuntimeError(f"Failed to read configuration: {e}")
+
+    if "device" not in config:
+        raise RuntimeError("Configuration error: missing 'device' key")
+
+    return config
+
+
+def interrogate_drive(device: str) -> dict:
+    """Interrogate the drive via smartctl.
+
+    Returns the smartctl JSON output.
+    Raises RuntimeError on failure.
+    """
+    result = subprocess.run(
+        ["smartctl", "-a", "-j", device],
+        capture_output=True,
+        text=True,
+    )
+
+    if result.returncode != 0:
+        raise RuntimeError(
+            f"smartctl failed for {device}: {result.stderr}"
+        )
+
+    try:
+        return json.loads(result.stdout)
+    except json.JSONDecodeError as e:
+        raise RuntimeError(f"Failed to parse smartctl output: {e}")
+
+
+def find_nvme_sysfs() -> Path | None:
+    """Find the NVMe controller sysfs path."""
+    nvme_ctrl = Path("/sys/class/nvme")
+    if not nvme_ctrl.exists():
+        return None
+
+    for ctrl in sorted(nvme_ctrl.iterdir()):
+        if ctrl.name.startswith("nvme"):
+            return ctrl
+    return None
+
+
+def main() -> None:
+    """Run one collection cycle."""
+    try:
+        config = load_config()
+        device = config["device"]
+
+        # Interrogate the drive
+        smartctl_data = interrogate_drive(device)
+
+        # Find sysfs path
+        sysfs_path = find_nvme_sysfs()
+        if sysfs_path is None:
+            raise RuntimeError("No NVMe controller found in sysfs")
+
+        # Inject a simple clock
+        class SimpleClock:
+            def utcnow(self):
+                return datetime.now(timezone.utc)
+
+        clock = SimpleClock()
+
+        # Run collection
+        result = run_collection(smartctl_data, sysfs_path, config, clock)
+
+        if result["ok"]:
+            print(f"Collection successful: {result['sample_count']} sample(s)")
+            sys.exit(0)
+        else:
+            print(f"Collection failed: {result['error']}", file=sys.stderr)
+            sys.exit(1)
+
+    except Exception as e:
+        print(f"Collection error: {e}", file=sys.stderr)
+        sys.exit(1)
+
+
+if __name__ == "__main__":
+    main()