fix(firecracker): make cleanup reap orphans only, never live VMs

`cleanup()` enumerated *every* firecracker process under the run root and
*every* run dir, so `./cli.py cleanup` would kill running bottles and
delete their rootfs — despite being described as orphan cleanup. The
backend's `enumerate_active` registry is still a stub (#354), so there
was no live/dead signal.

Use the running firecracker processes themselves as that signal: a run
dir with a live VM is protected (never killed, never removed); only run
dirs with no live process (leaked by a hard-killed launch) are reaped,
plus firecracker pids whose run dir is already gone (lingering VMMs).

This makes cleanup safe to run alongside live bottles and lets it back a
periodic GC for the run-dir leak the launch.py teardown fix closes on the
clean-exit path.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

bot_bottle/backend/firecracker/cleanup.py

@@ -1,8 +1,23 @@
 """Cleanup for the Firecracker backend.
 
-Orphans are: firecracker VMM processes whose config lives under our run
-dir, and the per-bottle run dirs. TAP slots free themselves (the flock
-drops when the launcher exits), so there is nothing to reclaim there.
+Reaps *orphans* only — resources with no live VM behind them:
+
+  * orphan run dirs: a per-bottle run dir (holding the ~1G rootfs.ext4)
+    whose firecracker process has exited. These leak when a launch is
+    hard-killed before its teardown runs (host OOM/crash, a cancelled CI
+    job, `kill -9`); the clean-exit path already removes its own dir in
+    launch.py.
+  * orphan VM pids: a firecracker process whose run dir is already gone
+    — a VMM left lingering after its dir was removed.
+
+A run dir with a *live* firecracker process is a running bottle and is
+left strictly alone: it is neither killed nor removed. (The backend's
+`enumerate_active` registry is still a stub — #354 — so a live process
+is the only reliable "this bottle is in use" signal we have. Once the
+registry lands, registry-orphaned-but-running VMs can be reaped too.)
+
+TAP slots free themselves (the flock drops when the launcher exits), so
+there is nothing to reclaim there.
 """
 
 from __future__ import annotations
@@ -22,38 +37,73 @@ def _run_root() -> Path:
     return util.cache_dir() / "run"
 
 
-def _orphan_vm_pids() -> list[int]:
-    """firecracker processes whose --config-file is under our run dir."""
-    run_root = str(_run_root())
+def _run_dir_of(cmd: str, run_root: Path) -> Path | None:
+    """The bottle run dir a firecracker cmdline belongs to, or None.
+
+    A bottle VM is launched with `--config-file <run_root>/<slug>/config.json`,
+    so the run dir is the config file's parent when it sits directly under
+    the run root. Anything else (a builder VM, the infra VM elsewhere) is
+    not ours to reap here.
+    """
+    toks = cmd.split()
+    for i, tok in enumerate(toks):
+        if tok == "--config-file" and i + 1 < len(toks):
+            parent = Path(toks[i + 1]).parent
+            if parent.parent == run_root:
+                return parent
+    return None
+
+
+def _scan_processes(run_root: Path) -> tuple[set[str], list[int]]:
+    """Inspect running firecracker VMs under ``run_root``.
+
+    Returns ``(live_run_dirs, orphan_pids)``:
+      * ``live_run_dirs`` — run dirs backed by a running VM (never reaped);
+      * ``orphan_pids`` — firecracker pids whose run dir no longer exists
+        (a lingering VMM to kill).
+    """
     result = subprocess.run(
         ["pgrep", "-a", "firecracker"],
         capture_output=True, text=True, check=False,
     )
     if result.returncode != 0:
-        return []
-    pids: list[int] = []
+        return set(), []
+    live: set[str] = set()
+    orphan_pids: list[int] = []
     for line in result.stdout.splitlines():
         parts = line.split(None, 1)
-        if len(parts) != 2 or run_root not in parts[1]:
+        if len(parts) != 2:
             continue
         try:
-            pids.append(int(parts[0]))
+            pid = int(parts[0])
         except ValueError:
             continue
-    return pids
+        run_dir = _run_dir_of(parts[1], run_root)
+        if run_dir is None:
+            continue
+        if run_dir.is_dir():
+            live.add(str(run_dir))
+        else:
+            orphan_pids.append(pid)
+    return live, orphan_pids
 
 
-def _run_dirs() -> list[str]:
-    run_root = _run_root()
+def _orphan_run_dirs(run_root: Path, live: set[str]) -> list[str]:
+    """Run dirs with no live VM behind them — the leaked ones to remove."""
     if not run_root.is_dir():
         return []
-    return sorted(str(p) for p in run_root.iterdir() if p.is_dir())
+    return sorted(
+        str(p) for p in run_root.iterdir()
+        if p.is_dir() and str(p) not in live
+    )
 
 
 def prepare_cleanup() -> FirecrackerBottleCleanupPlan:
+    run_root = _run_root()
+    live, orphan_pids = _scan_processes(run_root)
     return FirecrackerBottleCleanupPlan(
-        vm_pids=tuple(_orphan_vm_pids()),
-        run_dirs=tuple(_run_dirs()),
+        vm_pids=tuple(orphan_pids),
+        run_dirs=tuple(_orphan_run_dirs(run_root, live)),
     )