fix(snmp): close SnmpEngine after each poll to stop fd leak
CI / lint (push) Successful in 2s
CI / unit (push) Successful in 45s
CI / integration (push) Successful in 2m20s
CI / publish (push) Successful in 1m2s

A fresh SnmpEngine() was created on every poll_device() call and never
closed. pysnmp opens a UDP transport socket per engine and doesn't release
it on GC, so each scheduler tick (default 60s, per device) leaked a file
descriptor. Over hours of polling the process hit its fd ceiling and the
listening socket could no longer accept connections — OSError: [Errno 24]
Too many open files on socket.accept(), locking up the app.

Wrap the engine in try/finally and release its transport socket via a new
_close_engine() helper that probes both pysnmp API shapes (6.2.x lextudio
camelCase, canonical 7.x snake_case); all close paths are best-effort so a
failed close never breaks the poll loop. Regression tests cover both shapes
and the never-raises contract.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
2026-06-24 09:10:30 -04:00
parent b1aabb7dfa
commit 2df5fc94a3
2 changed files with 125 additions and 39 deletions
+44
View File
@@ -30,3 +30,47 @@ def test_poll_device_without_pysnmp_returns_empty(monkeypatch):
poller.poll_device("192.0.2.1", 161, "public", "2c", [{"oid": "1.3.6.1.2.1.1.3.0"}])
)
assert out == {}
# --- _close_engine: the fd-leak guard ---------------------------------------
# A fresh SnmpEngine per poll leaks its UDP socket unless closed; _close_engine
# must release it across both pysnmp API shapes. These fakes stand in for the
# engine since the unit lane has no pysnmp.
def test_close_engine_uses_engine_level_close_7x():
# canonical pysnmp 7.x: close method lives directly on the engine.
calls = []
class Engine:
def close_dispatcher(self):
calls.append("engine")
poller._close_engine(Engine())
assert calls == ["engine"]
def test_close_engine_falls_back_to_dispatcher_6x():
# pysnmp-lextudio 6.2.x: no engine-level close; go through the dispatcher.
calls = []
class Dispatcher:
def closeDispatcher(self):
calls.append("dispatcher")
class Engine:
transportDispatcher = Dispatcher()
poller._close_engine(Engine())
assert calls == ["dispatcher"]
def test_close_engine_never_raises():
# A failing close must not break the poll loop.
class Engine:
def close_dispatcher(self):
raise OSError("boom")
poller._close_engine(Engine()) # no exception
# No close path at all (defensive) is also fine.
poller._close_engine(object())