Issue #3 - Unbounded Query Results (OOM): - get_all_archived_history() now uses keyset pagination with bounded max_rows (50k default) - Added 'id' field to records from get_archived_history() and get_archived_history_keyset() - Protocol signature updated with page_size, max_rows, last_ban_id params Issue #7 - Docker Health Check Fails: - Added curl to Dockerfile.backend runtime image - HEALTHCHECK now uses 'curl -f http://localhost:8000/api/health' - compose.prod.yml: increased start_period to 40s, timeout to 10s - Frontend healthcheck proxies to backend /api/health Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
533 lines
18 KiB
Python
533 lines
18 KiB
Python
"""Tests for the scheduler lock mechanism.
|
|
|
|
These tests verify that the database-backed scheduler lock correctly enforces
|
|
single-executor safety across multiple startup attempts, including stale lock
|
|
cleanup, heartbeat updates, and multi-process race condition prevention.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
import os
|
|
import time
|
|
from typing import Any
|
|
|
|
import aiosqlite
|
|
import pytest
|
|
|
|
from app.utils.scheduler_lock import (
|
|
SCHEDULER_LOCK_HEARTBEAT_INTERVAL_SECONDS,
|
|
SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS,
|
|
SCHEDULER_LOCK_TTL_SECONDS,
|
|
acquire_scheduler_lock,
|
|
get_lock_health,
|
|
get_scheduler_lock_info,
|
|
is_lock_stale,
|
|
release_scheduler_lock,
|
|
update_scheduler_lock_heartbeat,
|
|
)
|
|
|
|
|
|
@pytest.fixture
|
|
async def lock_db(tmp_path: Any) -> aiosqlite.Connection:
|
|
"""Create a temporary database with scheduler_lock table."""
|
|
db_path = tmp_path / "test.db"
|
|
db = await aiosqlite.connect(str(db_path))
|
|
await db.execute(
|
|
f"""
|
|
CREATE TABLE scheduler_lock (
|
|
id INTEGER PRIMARY KEY CHECK (id = 1),
|
|
pid INTEGER NOT NULL,
|
|
hostname TEXT NOT NULL,
|
|
created_at REAL NOT NULL,
|
|
heartbeat_at REAL NOT NULL,
|
|
heartbeat_timeout REAL NOT NULL DEFAULT {SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS}
|
|
);
|
|
"""
|
|
)
|
|
await db.commit()
|
|
yield db
|
|
await db.close()
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_acquire_scheduler_lock_success(lock_db: aiosqlite.Connection) -> None:
|
|
"""Test successful lock acquisition."""
|
|
result = await acquire_scheduler_lock(lock_db)
|
|
assert result is True
|
|
|
|
# Verify the lock is in the database
|
|
cursor = await lock_db.execute("SELECT COUNT(*) FROM scheduler_lock")
|
|
count = await cursor.fetchone()
|
|
assert count[0] == 1
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_acquire_scheduler_lock_fails_when_held(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test that lock acquisition fails if already held by another process.
|
|
|
|
Note: Same-PID re-acquire is allowed (refresh). Use separate connection
|
|
with different PID to test rejection.
|
|
"""
|
|
# First instance acquires the lock
|
|
result1 = await acquire_scheduler_lock(lock_db)
|
|
assert result1 is True
|
|
|
|
# Second instance (same process, same PID) - re-acquire succeeds (refresh)
|
|
result_same_pid = await acquire_scheduler_lock(lock_db)
|
|
assert result_same_pid is True
|
|
|
|
# To test rejection, create a separate database with a conflicting lock
|
|
# Simulate a different process holding the lock by inserting directly
|
|
import tempfile
|
|
import os
|
|
|
|
# Create a new in-memory database with pre-existing lock from "another process"
|
|
db_other = await aiosqlite.connect(":memory:")
|
|
await db_other.execute(
|
|
f"""
|
|
CREATE TABLE scheduler_lock (
|
|
id INTEGER PRIMARY KEY CHECK (id = 1),
|
|
pid INTEGER NOT NULL,
|
|
hostname TEXT NOT NULL,
|
|
created_at REAL NOT NULL,
|
|
heartbeat_at REAL NOT NULL,
|
|
heartbeat_timeout REAL NOT NULL DEFAULT {SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS}
|
|
)
|
|
"""
|
|
)
|
|
# Insert lock with PID=-1 (simulating another active process with recent heartbeat)
|
|
now = time.time()
|
|
await db_other.execute(
|
|
f"""
|
|
INSERT INTO scheduler_lock (id, pid, hostname, created_at, heartbeat_at, heartbeat_timeout)
|
|
VALUES (1, -1, 'other-host', ?, ?, {SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS})
|
|
""",
|
|
(now, now),
|
|
)
|
|
await db_other.commit()
|
|
|
|
# Now test that acquire fails when lock is held by another process
|
|
result_other = await acquire_scheduler_lock(db_other)
|
|
assert result_other is False
|
|
|
|
await db_other.close()
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_acquire_scheduler_lock_cleans_stale_locks(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test that stale locks are automatically cleaned up."""
|
|
# Insert a stale lock manually (old heartbeat)
|
|
now = time.time()
|
|
stale_heartbeat = now - SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS - 10
|
|
await lock_db.execute(
|
|
"""
|
|
INSERT INTO scheduler_lock (id, pid, hostname, created_at, heartbeat_at, heartbeat_timeout)
|
|
VALUES (1, 9999, 'stale-host', ?, ?, ?)
|
|
""",
|
|
(now - 100, stale_heartbeat, SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS),
|
|
)
|
|
await lock_db.commit()
|
|
|
|
# New instance should clean up the stale lock and acquire
|
|
result = await acquire_scheduler_lock(lock_db)
|
|
assert result is True
|
|
|
|
# Verify the old lock is gone and new one is in place
|
|
cursor = await lock_db.execute(
|
|
"SELECT pid, hostname FROM scheduler_lock WHERE id = 1"
|
|
)
|
|
row = await cursor.fetchone()
|
|
assert row is not None
|
|
pid, hostname = row
|
|
assert pid == os.getpid()
|
|
assert hostname is not None
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_acquire_scheduler_lock_cleans_stale_locks_with_new_schema(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test that stale locks are automatically cleaned up with new timeout field."""
|
|
# Insert a stale lock manually (heartbeat past timeout)
|
|
now = time.time()
|
|
stale_heartbeat = now - SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS - 10
|
|
await lock_db.execute(
|
|
"""
|
|
INSERT INTO scheduler_lock (id, pid, hostname, created_at, heartbeat_at, heartbeat_timeout)
|
|
VALUES (1, 9999, 'stale-host', ?, ?, ?)
|
|
""",
|
|
(now - 100, stale_heartbeat, SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS),
|
|
)
|
|
await lock_db.commit()
|
|
|
|
# New instance should steal the stale lock and acquire
|
|
result = await acquire_scheduler_lock(lock_db)
|
|
assert result is True
|
|
|
|
# Verify the old lock is gone and new one is in place
|
|
cursor = await lock_db.execute(
|
|
"SELECT pid, hostname, heartbeat_timeout FROM scheduler_lock WHERE id = 1"
|
|
)
|
|
row = await cursor.fetchone()
|
|
assert row is not None
|
|
pid, hostname, timeout = row
|
|
assert pid == os.getpid()
|
|
assert hostname is not None
|
|
assert timeout == SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_release_scheduler_lock_success(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test successful lock release."""
|
|
# Acquire the lock
|
|
await acquire_scheduler_lock(lock_db)
|
|
|
|
# Release it
|
|
await release_scheduler_lock(lock_db)
|
|
|
|
# Verify the lock is gone
|
|
cursor = await lock_db.execute("SELECT COUNT(*) FROM scheduler_lock")
|
|
count = await cursor.fetchone()
|
|
assert count[0] == 0
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_release_scheduler_lock_not_held(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test that releasing a lock we don't hold is safe."""
|
|
# Try to release without acquiring — should not crash
|
|
await release_scheduler_lock(lock_db)
|
|
|
|
# Verify the lock is still empty
|
|
cursor = await lock_db.execute("SELECT COUNT(*) FROM scheduler_lock")
|
|
count = await cursor.fetchone()
|
|
assert count[0] == 0
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_update_scheduler_lock_heartbeat_success(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test successful heartbeat update."""
|
|
# Acquire the lock
|
|
await acquire_scheduler_lock(lock_db)
|
|
|
|
# Get the original heartbeat
|
|
cursor = await lock_db.execute("SELECT heartbeat_at FROM scheduler_lock WHERE id = 1")
|
|
original_row = await cursor.fetchone()
|
|
original_heartbeat = original_row[0]
|
|
|
|
# Wait a moment and update the heartbeat
|
|
time.sleep(0.01)
|
|
result = await update_scheduler_lock_heartbeat(lock_db)
|
|
assert result is True
|
|
|
|
# Verify the heartbeat was updated
|
|
cursor = await lock_db.execute("SELECT heartbeat_at FROM scheduler_lock WHERE id = 1")
|
|
new_row = await cursor.fetchone()
|
|
new_heartbeat = new_row[0]
|
|
assert new_heartbeat > original_heartbeat
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_update_scheduler_lock_heartbeat_fails_if_not_held(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test that heartbeat update fails if we don't hold the lock."""
|
|
result = await update_scheduler_lock_heartbeat(lock_db)
|
|
assert result is False
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_get_scheduler_lock_info_returns_details(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test that lock info includes all relevant fields."""
|
|
await acquire_scheduler_lock(lock_db)
|
|
|
|
info = await get_scheduler_lock_info(lock_db)
|
|
assert info is not None
|
|
assert "pid" in info
|
|
assert "hostname" in info
|
|
assert "created_at" in info
|
|
assert "heartbeat_at" in info
|
|
assert info["pid"] == os.getpid()
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_get_scheduler_lock_info_returns_none_when_empty(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test that lock info returns None when no lock is held."""
|
|
info = await get_scheduler_lock_info(lock_db)
|
|
assert info is None
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_scheduler_lock_full_lifecycle(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test the full lifecycle: acquire, update heartbeat, release."""
|
|
# Initially no lock
|
|
info = await get_scheduler_lock_info(lock_db)
|
|
assert info is None
|
|
|
|
# Acquire the lock
|
|
result = await acquire_scheduler_lock(lock_db)
|
|
assert result is True
|
|
info = await get_scheduler_lock_info(lock_db)
|
|
assert info is not None
|
|
initial_heartbeat = info["heartbeat_at"]
|
|
|
|
# Update heartbeat multiple times
|
|
time.sleep(0.01)
|
|
result = await update_scheduler_lock_heartbeat(lock_db)
|
|
assert result is True
|
|
info = await get_scheduler_lock_info(lock_db)
|
|
updated_heartbeat = info["heartbeat_at"]
|
|
assert updated_heartbeat > initial_heartbeat
|
|
|
|
# Release the lock
|
|
await release_scheduler_lock(lock_db)
|
|
info = await get_scheduler_lock_info(lock_db)
|
|
assert info is None
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_scheduler_lock_heartbeat_interval_sanity(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Verify heartbeat interval is less than TTL to prevent premature expiry.
|
|
|
|
With a 5-second heartbeat interval and 60-second TTL, the lock can survive
|
|
~12 missed heartbeats before expiring. This provides robust protection against
|
|
temporary delays or high load that could cause a single missed heartbeat.
|
|
"""
|
|
# Verify the configuration ratio is safe (interval < TTL)
|
|
assert SCHEDULER_LOCK_HEARTBEAT_INTERVAL_SECONDS < SCHEDULER_LOCK_TTL_SECONDS
|
|
|
|
# With this ratio, the lock can survive at least 12 missed heartbeats
|
|
# (60s TTL / 5s interval = 12 intervals between heartbeats before expiry)
|
|
safe_ratio = SCHEDULER_LOCK_TTL_SECONDS / SCHEDULER_LOCK_HEARTBEAT_INTERVAL_SECONDS
|
|
assert safe_ratio >= 12, (
|
|
f"Heartbeat interval too long: lock can only survive {safe_ratio:.1f} missed heartbeats. "
|
|
f"Should be at least 12 for safety."
|
|
)
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_scheduler_lock_two_instances_cannot_both_hold(
|
|
tmp_path: Any,
|
|
) -> None:
|
|
"""Test that two different processes cannot both hold the lock.
|
|
|
|
This simulates two instances trying to acquire the lock. The second
|
|
instance should fail to acquire while the first holds a valid lock.
|
|
|
|
Note: Same-PID re-acquire is allowed (refresh). To test rejection,
|
|
we insert a lock with a different PID before testing.
|
|
"""
|
|
db_path = tmp_path / "test.db"
|
|
|
|
# Instance A connects and acquires the lock
|
|
db_a = await aiosqlite.connect(str(db_path))
|
|
await db_a.execute(
|
|
f"""
|
|
CREATE TABLE scheduler_lock (
|
|
id INTEGER PRIMARY KEY CHECK (id = 1),
|
|
pid INTEGER NOT NULL,
|
|
hostname TEXT NOT NULL,
|
|
created_at REAL NOT NULL,
|
|
heartbeat_at REAL NOT NULL,
|
|
heartbeat_timeout REAL NOT NULL DEFAULT {SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS}
|
|
);
|
|
"""
|
|
)
|
|
await db_a.commit()
|
|
|
|
result_a = await acquire_scheduler_lock(db_a)
|
|
assert result_a is True
|
|
|
|
# Same-PID re-acquire succeeds (refresh behavior)
|
|
result_a_refresh = await acquire_scheduler_lock(db_a)
|
|
assert result_a_refresh is True
|
|
|
|
# Simulate another process holding the lock by inserting with a different PID
|
|
# (this is the "conflicting" lock we want to reject)
|
|
await db_a.execute(
|
|
f"""
|
|
INSERT OR REPLACE INTO scheduler_lock (id, pid, hostname, created_at, heartbeat_at, heartbeat_timeout)
|
|
VALUES (1, -999, 'other-host', {time.time()}, {time.time()}, {SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS})
|
|
"""
|
|
)
|
|
await db_a.commit()
|
|
|
|
# Instance B (different connection, same PID in test) tries to acquire
|
|
# Should fail because different PID (-999) holds the lock
|
|
db_b = await aiosqlite.connect(str(db_path))
|
|
result_b = await acquire_scheduler_lock(db_b)
|
|
assert result_b is False
|
|
|
|
# Clear the conflicting lock directly (simulating other process dying)
|
|
await db_a.execute("DELETE FROM scheduler_lock")
|
|
await db_a.commit()
|
|
|
|
# Now Instance B can acquire
|
|
result_b3 = await acquire_scheduler_lock(db_b)
|
|
assert result_b3 is True
|
|
|
|
await db_a.close()
|
|
await db_b.close()
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_acquire_scheduler_lock_steals_stale_lock(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test that a stale lock can be stolen by another instance.
|
|
|
|
Scenario: Process A acquires the lock but crashes (never releases it).
|
|
Process B starts up and sees the lock has stale heartbeat (past timeout).
|
|
Process B should be able to steal the lock.
|
|
|
|
This is the key fix for the race condition issue: orphaned locks no longer
|
|
permanently block the scheduler.
|
|
"""
|
|
# Simulate Process A acquiring the lock
|
|
result_a = await acquire_scheduler_lock(lock_db)
|
|
assert result_a is True
|
|
|
|
# Get lock info to see heartbeat timeout
|
|
info = await get_scheduler_lock_info(lock_db)
|
|
assert info is not None
|
|
heartbeat_timeout = info["heartbeat_timeout"]
|
|
|
|
# Simulate stale lock: manually set heartbeat to far in the past
|
|
now = time.time()
|
|
stale_heartbeat = now - heartbeat_timeout - 10
|
|
await lock_db.execute(
|
|
"UPDATE scheduler_lock SET heartbeat_at = ? WHERE id = 1",
|
|
(stale_heartbeat,),
|
|
)
|
|
await lock_db.commit()
|
|
|
|
# Process B should now be able to acquire (steal) the stale lock
|
|
result_b = await acquire_scheduler_lock(lock_db)
|
|
assert result_b is True
|
|
|
|
# Verify Process B now holds the lock
|
|
info_b = await get_scheduler_lock_info(lock_db)
|
|
assert info_b is not None
|
|
assert info_b["pid"] == os.getpid()
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_is_lock_stale_function() -> None:
|
|
"""Test the is_lock_stale helper function."""
|
|
now = time.time()
|
|
timeout = 300.0
|
|
|
|
# Fresh lock is not stale
|
|
heartbeat_at = now - 10
|
|
assert await is_lock_stale(heartbeat_at, timeout, now) is False
|
|
|
|
# Lock past timeout is stale
|
|
heartbeat_at = now - 400
|
|
assert await is_lock_stale(heartbeat_at, timeout, now) is True
|
|
|
|
# Exactly at timeout is not stale (boundary condition)
|
|
heartbeat_at = now - 300
|
|
assert await is_lock_stale(heartbeat_at, timeout, now) is False
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_get_lock_health_no_lock(lock_db: aiosqlite.Connection) -> None:
|
|
"""Test get_lock_health when no lock exists."""
|
|
health = await get_lock_health(lock_db)
|
|
assert health["has_lock"] is False
|
|
assert health["is_stale"] is False
|
|
assert health["pid"] is None
|
|
assert health["stale_reason"] is None
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_get_lock_health_active_lock(lock_db: aiosqlite.Connection) -> None:
|
|
"""Test get_lock_health with an active, healthy lock."""
|
|
await acquire_scheduler_lock(lock_db)
|
|
|
|
health = await get_lock_health(lock_db)
|
|
assert health["has_lock"] is True
|
|
assert health["is_stale"] is False
|
|
assert health["pid"] == os.getpid()
|
|
assert health["hostname"] is not None
|
|
assert health["heartbeat_timeout"] == SCHEDULER_LOCK_HEARTBEAT_TIMEOUT_SECONDS
|
|
assert health["stale_reason"] is None
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_get_lock_health_stale_lock(lock_db: aiosqlite.Connection) -> None:
|
|
"""Test get_lock_health with a stale lock."""
|
|
await acquire_scheduler_lock(lock_db)
|
|
|
|
# Manually make the lock stale
|
|
now = time.time()
|
|
info = await get_scheduler_lock_info(lock_db)
|
|
stale_heartbeat = now - info["heartbeat_timeout"] - 10
|
|
await lock_db.execute(
|
|
"UPDATE scheduler_lock SET heartbeat_at = ? WHERE id = 1",
|
|
(stale_heartbeat,),
|
|
)
|
|
await lock_db.commit()
|
|
|
|
health = await get_lock_health(lock_db)
|
|
assert health["has_lock"] is True
|
|
assert health["is_stale"] is True
|
|
assert health["stale_reason"] is not None
|
|
assert "heartbeat_age" in health["stale_reason"]
|
|
assert "timeout" in health["stale_reason"]
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_heartbeat_update_error_returns_false(
|
|
lock_db: aiosqlite.Connection,
|
|
) -> None:
|
|
"""Test that heartbeat update errors return False instead of raising."""
|
|
# Try to update heartbeat without acquiring lock first
|
|
result = await update_scheduler_lock_heartbeat(lock_db)
|
|
assert result is False
|
|
|
|
# Acquire lock
|
|
await acquire_scheduler_lock(lock_db)
|
|
|
|
# Heartbeat should work
|
|
result = await update_scheduler_lock_heartbeat(lock_db)
|
|
assert result is True
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_concurrent_acquire_from_same_process(lock_db: aiosqlite.Connection) -> None:
|
|
"""Test that concurrent acquire attempts from same process re-acquires (refreshes)."""
|
|
# First acquisition should succeed
|
|
result1 = await acquire_scheduler_lock(lock_db)
|
|
assert result1 is True
|
|
|
|
# Second acquisition from same process should succeed (re-acquire/refresh)
|
|
result2 = await acquire_scheduler_lock(lock_db)
|
|
assert result2 is True
|
|
|
|
# Heartbeat should be updated
|
|
info = await get_scheduler_lock_info(lock_db)
|
|
assert info is not None
|
|
|
|
# Release and re-acquire should work
|
|
await release_scheduler_lock(lock_db)
|
|
result3 = await acquire_scheduler_lock(lock_db)
|
|
assert result3 is True
|