mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-11-17 04:11:35 +00:00
The session-hygiene compressor runs on the loop's default executor (run_in_executor(None, ...)), which the self._executor quiesce never joins. It was only registered with _track_deferred_agent_worker once a timeout, turn-hold or unwind deferred it, so a gateway stop() that reached the SessionDB close while the awaiting turn was still waiting saw zero deferred workers and closed/checkpointed state.db under the worker's late write (#101064 shape). Track the future right after run_in_executor, mirroring run_codex_hygiene_compaction. The existing close guard (#102198) now skips close:session_db while the worker is live, and shutdown's interrupt pass reaches the hygiene agent. _defer_agent_cleanup_until_future_done is kept for cleanup; re-tracking the same future is idempotent (dict key). The default executor is deliberately NOT shut down: stop() still uses asyncio.to_thread afterwards (terminal runtime-status flush). Regression test adapted from @pmaho's test_default_executor_worker_is_seen_by_the_close_guard (#121360), now driving the real _hmwa_hygiene_detached_attempt submission site. Co-authored-by: pmaho <[email protected]>
437 lines
16 KiB
Python
437 lines
16 KiB
Python
"""Gateway shutdown quiesces its thread pool before closing state.db (#101093).
|
|
|
|
``_shutdown_executor()`` used to run *after* the SessionDB close block in
|
|
``_stop_impl``, and it never waited: ``cancel_futures`` only drops work that has
|
|
not started, and cancelling the awaiting task does not stop the worker thread
|
|
behind a ``run_in_executor`` future. So blocking DB work could still be running
|
|
when ``SessionDB.close()`` checkpointed the WAL and let SQLite unlink the
|
|
sidecar. The late write then reopens the handle (#94736) and mints a fresh WAL
|
|
generation behind that checkpoint, leaving teardown to checkpoint the same file
|
|
a second time from a connection the shutdown log never accounts for -- the
|
|
close-time page-write damage in #101093 and the split WAL generation in #101064.
|
|
|
|
The order is now: quiesce (bounded) -> close.
|
|
"""
|
|
|
|
import asyncio
|
|
import concurrent.futures
|
|
import threading
|
|
import time
|
|
from collections import OrderedDict
|
|
from contextlib import suppress
|
|
|
|
import pytest
|
|
|
|
import gateway.run as gw_mod
|
|
|
|
class _FakeSessionDB:
|
|
"""Records when the gateway closed it, on a shared event log."""
|
|
|
|
def __init__(self, events, name):
|
|
self._events = events
|
|
self._name = name
|
|
|
|
def close(self):
|
|
self._events.append(f"close:{self._name}")
|
|
|
|
class _FakeGateway:
|
|
"""Minimal stand-in with just enough state for ``stop()`` to run."""
|
|
|
|
def __init__(self, events):
|
|
self._events = events
|
|
self._running = True
|
|
self._draining = False
|
|
self._restart_requested = False
|
|
self._restart_detached = False
|
|
self._restart_via_service = False
|
|
self._stop_task = None
|
|
self._exit_cleanly = False
|
|
self._exit_with_failure = False
|
|
self._exit_reason = None
|
|
self._exit_code = None
|
|
self._restart_drain_timeout = 0.01
|
|
self._running_agents = {}
|
|
self._running_agents_ts = {}
|
|
self._agent_cache = OrderedDict()
|
|
self._agent_cache_lock = threading.Lock()
|
|
self.adapters = {}
|
|
self._background_tasks = set()
|
|
self._failed_platforms = []
|
|
self._shutdown_event = asyncio.Event()
|
|
self._pending_messages = {}
|
|
self._pending_approvals = {}
|
|
self._busy_ack_ts = {}
|
|
self._executor_lock = threading.Lock()
|
|
self._executor_closing = False
|
|
self._executor = concurrent.futures.ThreadPoolExecutor(
|
|
max_workers=2, thread_name_prefix="quiesce-test"
|
|
)
|
|
self._session_db = _FakeSessionDB(events, "session_db")
|
|
self.session_store = None
|
|
|
|
# -- shutdown collaborators the real stop() reaches into ---------------
|
|
|
|
def _running_agent_count(self):
|
|
return len(self._running_agents)
|
|
|
|
def _active_cron_job_count(self):
|
|
return 0
|
|
|
|
# Real hook + counter: 0 while ``adapters`` is empty, the API-server count once a fake adapter is in.
|
|
_api_server_hook = gw_mod.GatewayShutdownMixin._api_server_hook
|
|
_mark_api_runs_shutdown_requested = gw_mod.GatewayShutdownMixin._mark_api_runs_shutdown_requested
|
|
_active_api_run_count = gw_mod.GatewayShutdownMixin._active_api_run_count
|
|
_active_api_worker_count = gw_mod.GatewayShutdownMixin._active_api_worker_count
|
|
_active_deferred_agent_worker_count = gw_mod.GatewayShutdownMixin._active_deferred_agent_worker_count
|
|
|
|
def _update_runtime_status(self, *_a, **_kw):
|
|
pass
|
|
|
|
def _clear_plugin_message_injector(self):
|
|
pass
|
|
|
|
async def _run_in_executor_with_context(self, func, *args):
|
|
return func(*args)
|
|
|
|
async def _cleanup_agent_resources_off_loop(self, agent, *, context=""):
|
|
self._cleanup_agent_resources(agent)
|
|
|
|
async def _notify_active_sessions_of_shutdown(self):
|
|
pass
|
|
|
|
async def _cancel_secondary_profile_reconnect_tasks(self):
|
|
pass
|
|
|
|
async def _drain_active_agents(self, timeout, cron_timeout=None):
|
|
return {}, False
|
|
|
|
async def _finalize_shutdown_agents(self, agents):
|
|
pass
|
|
|
|
def _cleanup_agent_resources(self, agent):
|
|
pass
|
|
|
|
def _evict_cached_agent(self, key):
|
|
pass
|
|
|
|
def _release_running_agent_state(self, session_key, **_kwargs):
|
|
self._running_agents.pop(session_key, None)
|
|
self._running_agents_ts.pop(session_key, None)
|
|
return False
|
|
|
|
def close_all_session_db_handles(self):
|
|
pass
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_running_executor_work_finishes_before_session_db_close():
|
|
"""A future already running when stop() begins writes before the close."""
|
|
events = []
|
|
gw = _FakeGateway(events)
|
|
started = threading.Event()
|
|
|
|
def _blocking_db_write():
|
|
started.set()
|
|
# Longer than the rest of the shutdown tail (~0.4s), shorter than the
|
|
# 2s quiesce ceiling: without the wait the close lands first.
|
|
time.sleep(1.0)
|
|
events.append("worker_write")
|
|
|
|
future = gw._executor.submit(_blocking_db_write)
|
|
assert started.wait(2.0), "worker never started"
|
|
|
|
await gw_mod.GatewayRunner.stop(gw)
|
|
future.result(timeout=5)
|
|
|
|
assert "worker_write" in events, "worker never ran"
|
|
assert "close:session_db" in events, "SessionDB was never closed"
|
|
assert events.index("worker_write") < events.index("close:session_db"), (
|
|
f"state.db was closed while a worker was still writing: {events}"
|
|
)
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_executor_refuses_new_work_before_session_db_close():
|
|
"""``_executor_closing`` is set before the close, so no fresh pool is minted."""
|
|
events = []
|
|
gw = _FakeGateway(events)
|
|
|
|
real_close = gw._session_db.close
|
|
|
|
def _close_and_probe():
|
|
# The flag must already be set by the time the DB is closed, or a
|
|
# coroutine reaching _get_executor() here would spin up a new pool and
|
|
# run more blocking DB work against the handle being torn down.
|
|
events.append(f"closing_flag:{gw._executor_closing}")
|
|
real_close()
|
|
|
|
gw._session_db.close = _close_and_probe
|
|
|
|
await gw_mod.GatewayRunner.stop(gw)
|
|
|
|
assert "closing_flag:True" in events, events
|
|
with pytest.raises(RuntimeError):
|
|
gw_mod.GatewayRunner._get_executor(gw)
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_stuck_worker_skips_the_session_db_close():
|
|
"""A worker that outlives the quiesce budget must not be raced by close().
|
|
|
|
Reporting the live worker with a "may reopen state.db" warning is not
|
|
enough: the close()/checkpoint itself is the operation that raced the
|
|
late write and produced the wrong-page-number corruption in #101093,
|
|
so the close path has to be skipped whenever a worker survives the
|
|
budget, not merely logged around.
|
|
"""
|
|
events = []
|
|
gw = _FakeGateway(events)
|
|
release = threading.Event()
|
|
started = threading.Event()
|
|
|
|
def _stuck():
|
|
started.set()
|
|
release.wait(5.0)
|
|
events.append("worker_write")
|
|
|
|
future = gw._executor.submit(_stuck)
|
|
assert started.wait(2.0), "worker never started"
|
|
|
|
# Force the quiesce budget to 0 so the worker is deterministically still
|
|
# alive when `_shutdown_executor` returns, without sleeping through the
|
|
# real 2s ceiling.
|
|
original_timeout = gw_mod._EXECUTOR_QUIESCE_TIMEOUT
|
|
gw_mod._EXECUTOR_QUIESCE_TIMEOUT = 0.0
|
|
try:
|
|
await gw_mod.GatewayRunner.stop(gw)
|
|
finally:
|
|
gw_mod._EXECUTOR_QUIESCE_TIMEOUT = original_timeout
|
|
|
|
assert "close:session_db" not in events, (
|
|
f"SessionDB was closed/checkpointed while a worker was still alive: {events}"
|
|
)
|
|
|
|
release.set()
|
|
future.result(timeout=5)
|
|
assert "worker_write" in events, "worker never finished"
|
|
|
|
def _arm_cron(gw):
|
|
gw._active_cron_job_count = lambda: 1
|
|
|
|
def _arm_api(gw):
|
|
# Through the real hook: the adapter map is cleared one phase before the close gate, so the
|
|
# gate must use the count taken before the clear, not a live lookup.
|
|
from gateway.config import Platform
|
|
|
|
class _ApiAdapter:
|
|
def active_agent_work_count(self):
|
|
return 1
|
|
|
|
async def _teardown(adapter, platform, *, profile=None):
|
|
pass # the run keeps going on the default executor after the transport is torn down
|
|
|
|
gw.adapters[Platform.API_SERVER] = _ApiAdapter()
|
|
gw._bounded_adapter_teardown = _teardown
|
|
|
|
def _arm_deferred(gw):
|
|
# A hygiene worker on the loop's default executor, never finished.
|
|
gw._deferred_agent_workers = {asyncio.get_event_loop().create_future(): object()}
|
|
|
|
@pytest.mark.asyncio
|
|
@pytest.mark.parametrize("arm", [_arm_cron, _arm_api, _arm_deferred], ids=["cron", "api", "deferred"])
|
|
async def test_live_writer_outside_the_executor_skips_the_session_db_close(monkeypatch, arm):
|
|
"""A cron job, API-server run or deferred worker that outlived the drain must not have state.db
|
|
closed under it (#102198).
|
|
|
|
None of them run on ``self._executor`` (scheduler pool / loop default executor), so the executor
|
|
join above the close block never sees them; the close has to consult their counters too.
|
|
The executor must still be sealed on this path (#101118).
|
|
"""
|
|
import hermes_state_registry
|
|
|
|
events = []
|
|
gw = _FakeGateway(events)
|
|
arm(gw)
|
|
monkeypatch.setattr(
|
|
hermes_state_registry, "close_all", lambda: events.append("close_all") or 0
|
|
)
|
|
|
|
await gw_mod.GatewayRunner.stop(gw)
|
|
|
|
assert "close:session_db" not in events and "close_all" not in events, (
|
|
f"SessionDB closed despite a live {arm.__name__[5:]} writer: {events}"
|
|
)
|
|
assert gw._executor_closing is True, "executor left unsealed on the outside-writer path"
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_cancelled_api_handler_worker_still_blocks_session_db_close(monkeypatch):
|
|
"""A cancelled request handler must not let state.db be closed under its live worker (#116535).
|
|
|
|
The handler-side ``_inflight_agent_runs`` count drops in the handler's ``finally`` when the
|
|
handler task is cancelled, while the executor thread behind ``run_in_executor`` is still
|
|
blocked in the turn. Only the worker-scoped count still sees that thread, so the SessionDB
|
|
close gate must consult it alongside the handler snapshot -- and must not close until the
|
|
worker itself exits.
|
|
"""
|
|
import hermes_state_registry
|
|
from gateway.platforms import api_server_runs as api_runs
|
|
|
|
events = []
|
|
gw = _FakeGateway(events)
|
|
monkeypatch.setattr(
|
|
hermes_state_registry, "close_all", lambda: events.append("close_all") or 0
|
|
)
|
|
|
|
release = threading.Event()
|
|
worker_started = threading.Event()
|
|
|
|
def _blocked_turn():
|
|
worker_started.set()
|
|
assert release.wait(5.0), "test tore down while the worker was still blocked"
|
|
events.append("worker_done")
|
|
|
|
loop = asyncio.get_running_loop()
|
|
|
|
async def _handler():
|
|
# Same shape as the api_server call sites: the worker-scoped count is taken before
|
|
# run_in_executor and released in the worker's own finally, so cancelling this task
|
|
# drops the handler side while the thread keeps holding the worker side.
|
|
return await api_runs._submit_api_worker(loop, _blocked_turn)
|
|
|
|
task = asyncio.ensure_future(_handler())
|
|
assert await loop.run_in_executor(None, worker_started.wait, 5.0), "worker never started"
|
|
task.cancel()
|
|
with suppress(asyncio.CancelledError):
|
|
await task
|
|
assert api_runs.api_worker_live_count() == 1, (
|
|
"cancelled handler took the worker count with it"
|
|
)
|
|
|
|
await gw_mod.GatewayRunner.stop(gw)
|
|
|
|
assert "close:session_db" not in events and "close_all" not in events, (
|
|
f"SessionDB closed while the API worker was still alive: {events}"
|
|
)
|
|
|
|
release.set()
|
|
for _ in range(100):
|
|
if "worker_done" in events:
|
|
break
|
|
await asyncio.sleep(0.05)
|
|
assert "worker_done" in events, "worker never finished"
|
|
assert api_runs.api_worker_live_count() == 0, "worker exit did not release the count"
|
|
|
|
def test_shutdown_executor_defaults_to_no_wait():
|
|
"""The no-argument call keeps the historical fire-and-forget contract."""
|
|
gw = _FakeGateway([])
|
|
release = threading.Event()
|
|
started = threading.Event()
|
|
|
|
def _slow():
|
|
started.set()
|
|
release.wait(5.0)
|
|
|
|
future = gw._executor.submit(_slow)
|
|
assert started.wait(2.0)
|
|
|
|
began = time.monotonic()
|
|
still_live = gw_mod.GatewayRunner._shutdown_executor(gw)
|
|
elapsed = time.monotonic() - began
|
|
|
|
assert elapsed < 0.5, f"default call waited {elapsed:.2f}s"
|
|
assert still_live == 1
|
|
release.set()
|
|
future.result(timeout=5)
|
|
|
|
def test_shutdown_executor_reports_a_stuck_worker():
|
|
"""A worker that outlives the budget is reported, not waited on forever."""
|
|
gw = _FakeGateway([])
|
|
release = threading.Event()
|
|
started = threading.Event()
|
|
|
|
def _stuck():
|
|
started.set()
|
|
release.wait(5.0)
|
|
|
|
future = gw._executor.submit(_stuck)
|
|
assert started.wait(2.0)
|
|
|
|
began = time.monotonic()
|
|
still_live = gw_mod.GatewayRunner._shutdown_executor(gw, drain_timeout=0.2)
|
|
elapsed = time.monotonic() - began
|
|
|
|
assert still_live == 1
|
|
assert 0.15 <= elapsed < 2.0, f"budget not honoured: {elapsed:.2f}s"
|
|
release.set()
|
|
future.result(timeout=5)
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_default_executor_worker_is_seen_by_the_close_guard():
|
|
"""The detached hygiene compressor runs on the loop's DEFAULT executor, which the
|
|
``self._executor`` quiesce never joins. It must be visible to the SessionDB-close guard
|
|
from the moment it is submitted -- not only once a timeout/turn-hold/unwind defers it --
|
|
or stop() closes/checkpoints state.db under the worker's late write (#101064 shape;
|
|
reported by @pmaho in #121360).
|
|
|
|
Drives the real ``_hmwa_hygiene_detached_attempt`` submission site, and stops the gateway
|
|
while the awaiting turn has NOT been unwound yet (the window where nothing else tracks it).
|
|
"""
|
|
from types import MethodType, SimpleNamespace
|
|
|
|
events = []
|
|
gw = _FakeGateway(events)
|
|
for name in ("_hmwa_hygiene_detached_attempt", "_track_deferred_agent_worker"):
|
|
setattr(gw, name, MethodType(getattr(gw_mod.GatewayRunner, name), gw))
|
|
|
|
release = threading.Event()
|
|
started = threading.Event()
|
|
turn_waiting = asyncio.Event()
|
|
|
|
class _HygieneAgent:
|
|
context_compressor = None
|
|
|
|
def _compress_context(self, msgs, *_a, **_kw):
|
|
started.set()
|
|
release.wait(30.0)
|
|
events.append("hygiene_worker_write")
|
|
return msgs, None
|
|
|
|
agent = _HygieneAgent()
|
|
|
|
async def _build_agent(*_a):
|
|
return agent, None
|
|
|
|
async def _wait_for_summary(attempt, *_a):
|
|
turn_waiting.set()
|
|
await asyncio.shield(attempt.future) # the live turn is still waiting at stop()
|
|
return None
|
|
|
|
async def _apply_result(*_a, **_kw):
|
|
pass
|
|
|
|
gw._hmwa_hygiene_build_agent = _build_agent
|
|
gw._hmwa_hygiene_wait_for_summary = _wait_for_summary
|
|
gw._hmwa_hygiene_apply_result = _apply_result
|
|
|
|
attempt = gw_mod.GatewayRunner._HygieneAttempt(agent=None, meta=None)
|
|
hs = SimpleNamespace(total_ceiling_seconds=60.0)
|
|
plan = SimpleNamespace(approx_tokens=100)
|
|
turn = asyncio.create_task(gw._hmwa_hygiene_detached_attempt(
|
|
attempt, hs, plan, [], [{"role": "user", "content": "x"}] * 4, "m", {},
|
|
None, SimpleNamespace(session_id="sess-1"), "sk", "qk", 0,
|
|
))
|
|
await asyncio.wait_for(turn_waiting.wait(), 2.0)
|
|
assert await asyncio.to_thread(started.wait, 2.0), "hygiene worker never started"
|
|
|
|
original_timeout = gw_mod._EXECUTOR_QUIESCE_TIMEOUT
|
|
gw_mod._EXECUTOR_QUIESCE_TIMEOUT = 0.0
|
|
try:
|
|
await gw_mod.GatewayRunner.stop(gw)
|
|
finally:
|
|
gw_mod._EXECUTOR_QUIESCE_TIMEOUT = original_timeout
|
|
release.set()
|
|
await asyncio.wait_for(turn, 5.0)
|
|
|
|
assert "hygiene_worker_write" in events, "worker never finished"
|
|
assert "close:session_db" not in events, (
|
|
f"SessionDB was closed/checkpointed while the detached hygiene worker was still alive: {events}"
|
|
)
|
|
assert gw._active_deferred_agent_worker_count() == 0, "finished worker left registered"
|