Files
sglang/test/registered/scheduler/test_abort.py
2026-03-23 00:18:45 -07:00

366 lines
12 KiB
Python

import multiprocessing
import threading
import time
import unittest
from concurrent.futures import ThreadPoolExecutor, as_completed
import requests
from sglang.srt.environ import envs
from sglang.srt.utils import kill_process_tree
from sglang.test.ci.ci_register import register_amd_ci, register_cuda_ci
from sglang.test.kits.abort_timeout_kit import AbortAllMixin, WaitingTimeoutMixin
from sglang.test.test_utils import (
DEFAULT_MODEL_NAME_FOR_TEST,
DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
DEFAULT_URL_FOR_TEST,
CustomTestCase,
popen_launch_server,
run_and_check_memory_leak,
)
register_cuda_ci(est_time=131, suite="stage-b-test-1-gpu-small")
register_amd_ci(est_time=300, suite="stage-b-test-1-gpu-small-amd")
class TestAbort(CustomTestCase):
def workload_func(self, base_url, model):
def process_func():
def run_one(_):
prompt = """
System: You are a helpful assistant.
User: What is the capital of France?
Assistant: The capital of France is
"""
response = requests.post(
f"{base_url}/generate",
json={
"text": prompt,
"sampling_params": {
"temperature": 0,
"max_new_tokens": 2048,
},
},
)
ret = response.json()
with ThreadPoolExecutor(16) as executor:
list(executor.map(run_one, list(range(16))))
p = multiprocessing.Process(target=process_func)
p.start()
time.sleep(0.5)
p.terminate()
time.sleep(10)
def test_memory_leak(self):
run_and_check_memory_leak(
self.workload_func,
disable_radix_cache=False,
enable_mixed_chunk=False,
disable_overlap=False,
chunked_prefill_size=8192,
assert_has_abort=True,
)
class TestAbortWithApiKey(CustomTestCase):
def workload_func(self, base_url, model, api_key: str):
def process_func():
def run_one(_):
prompt = """
System: You are a helpful assistant.
User: What is the capital of France?
Assistant: The capital of France is
"""
response = requests.post(
f"{base_url}/generate",
json={
"text": prompt,
"sampling_params": {
"temperature": 0,
"max_new_tokens": 2048,
},
},
headers={"Authorization": f"Bearer {api_key}"},
)
response.json()
with ThreadPoolExecutor(16) as executor:
list(executor.map(run_one, list(range(16))))
p = multiprocessing.Process(target=process_func)
p.start()
time.sleep(0.5)
p.terminate()
time.sleep(10)
def test_memory_leak_with_api_key(self):
api_key = "test-api-key"
run_and_check_memory_leak(
lambda base_url, model: self.workload_func(base_url, model, api_key),
disable_radix_cache=False,
enable_mixed_chunk=False,
disable_overlap=False,
chunked_prefill_size=8192,
assert_has_abort=True,
api_key=api_key,
)
class TestAbortAll(AbortAllMixin, CustomTestCase):
@classmethod
def setUpClass(cls):
cls.model = DEFAULT_MODEL_NAME_FOR_TEST
cls.base_url = DEFAULT_URL_FOR_TEST
cls.process = popen_launch_server(
cls.model,
cls.base_url,
timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
other_args=["--max-running-requests", 8],
)
@classmethod
def tearDownClass(cls):
kill_process_tree(cls.process.pid)
def _generate_with_rid(self, rid, max_new_tokens=8):
return requests.post(
f"{self.base_url}/generate",
json={
"text": "The capital of France is",
"sampling_params": {
"temperature": 0,
"max_new_tokens": max_new_tokens,
},
"rid": rid,
},
timeout=30,
)
def test_duplicate_rid_sequential_ok(self):
rid = "dup-rid-test-sequential"
resp1 = self._generate_with_rid(rid)
self.assertEqual(resp1.status_code, 200)
self.assertNotIn("error", resp1.json())
resp2 = self._generate_with_rid(rid)
self.assertEqual(resp2.status_code, 200)
self.assertNotIn("error", resp2.json())
def test_duplicate_rid_concurrent_rejected(self):
rid = "dup-rid-test-concurrent"
results = {}
def send(key, max_tokens):
results[key] = self._generate_with_rid(rid, max_new_tokens=max_tokens)
t1 = threading.Thread(target=send, args=("first", 512))
t2 = threading.Thread(target=send, args=("second", 8))
t1.start()
time.sleep(0.1)
t2.start()
t1.join(timeout=30)
t2.join(timeout=30)
r1, r2 = results["first"], results["second"]
self.assertTrue(
r1.status_code == 400 or r2.status_code == 400,
"One of the concurrent duplicate-rid requests should be rejected",
)
rejected = r2 if r2.status_code == 400 else r1
self.assertIn("Duplicate request ID", rejected.json()["error"]["message"])
def test_duplicate_rid_in_batch(self):
rid = "dup-rid-batch"
response = requests.post(
f"{self.base_url}/generate",
json={
"text": ["Hello", "World"],
"sampling_params": {"temperature": 0, "max_new_tokens": 8},
"rid": [rid, rid],
},
timeout=30,
)
self.assertEqual(response.status_code, 400)
self.assertIn("Duplicate request ID", response.json()["error"]["message"])
def test_server_healthy_after_duplicate_rid(self):
requests.post(
f"{self.base_url}/generate",
json={
"text": ["Hello", "World"],
"sampling_params": {"temperature": 0, "max_new_tokens": 8},
"rid": ["dup-health", "dup-health"],
},
timeout=30,
)
resp = requests.get(f"{self.base_url}/health", timeout=5)
self.assertEqual(resp.status_code, 200)
resp = self._generate_with_rid("after-dup-health")
self.assertEqual(resp.status_code, 200)
self.assertIn("text", resp.json())
class TestAbortAllWithRetraction(CustomTestCase):
@classmethod
def setUpClass(cls):
cls.model = DEFAULT_MODEL_NAME_FOR_TEST
cls.base_url = DEFAULT_URL_FOR_TEST
# Here's a small trick: in scheduler.py, when SGLANG_TEST_RETRACT is enabled,
# retraction is triggered when the batch size reaches 10.
# However, since SGLANG_TEST_RETRACT_NO_PREFILL_BS is set to 6, the remaining 4
# requests will stay in the waiting queue.
with (
envs.SGLANG_TEST_RETRACT.override(True),
envs.SGLANG_TEST_RETRACT_NO_PREFILL_BS.override(6),
):
cls.process = popen_launch_server(
cls.model,
cls.base_url,
timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
other_args=[
"--max-running-requests",
16,
"--schedule-policy",
"random",
],
)
@classmethod
def tearDownClass(cls):
kill_process_tree(cls.process.pid)
def _run_decode(self):
response = requests.post(
self.base_url + "/generate",
json={
"text": "The capital of France is",
"sampling_params": {
"temperature": 0,
"max_new_tokens": 4000,
"ignore_eos": True,
},
"return_logprob": True,
"top_logprobs_num": 3,
},
)
return response.json()
def test_abort_all_with_retraction(self):
num_requests = 32
with ThreadPoolExecutor(num_requests) as executor:
futures = [executor.submit(self._run_decode) for _ in range(num_requests)]
# ensure the decode has been started and retractions happen.
time.sleep(8)
requests.post(
self.base_url + "/abort_request",
json={
"abort_all": True,
},
)
abort_in_queue_count = 0
abort_in_queue_with_partial_gen = 0
for future in as_completed(futures):
result = future.result()
meta_info = result["meta_info"]
finish_reason = meta_info.get("finish_reason", {})
self.assertEqual(finish_reason.get("type"), "abort")
if finish_reason.get("message") == "Abort in waiting queue":
abort_in_queue_count += 1
output_ids = result.get("output_ids", [])
if len(output_ids) > 0:
abort_in_queue_with_partial_gen += 1
self.assertEqual(
meta_info.get("completion_tokens"), len(output_ids)
)
self.assertGreater(len(result.get("text", "")), 0)
self.assertIsNotNone(meta_info.get("weight_version"))
self.assertGreater(meta_info.get("e2e_latency"), 0)
for logprob_key in [
"output_token_logprobs",
"output_top_logprobs",
]:
self.assertEqual(
len(meta_info.get(logprob_key, [])),
len(output_ids),
f"Length of '{logprob_key}' should match output_ids length",
)
self.assertGreater(abort_in_queue_count, 0)
self.assertGreater(abort_in_queue_with_partial_gen, 0)
print("Finished test_abort_all_with_retraction")
class TestAbortWithWaitingTimeout(WaitingTimeoutMixin, CustomTestCase):
@classmethod
def setUpClass(cls):
cls.model = DEFAULT_MODEL_NAME_FOR_TEST
cls.base_url = DEFAULT_URL_FOR_TEST
with envs.SGLANG_REQ_WAITING_TIMEOUT.override(0.001):
cls.process = popen_launch_server(
cls.model,
cls.base_url,
timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
other_args=[
"--max-running-requests=1",
],
)
@classmethod
def tearDownClass(cls):
kill_process_tree(cls.process.pid)
class TestAbortWithRunningTimeout(CustomTestCase):
@classmethod
def setUpClass(cls):
cls.model = DEFAULT_MODEL_NAME_FOR_TEST
cls.base_url = DEFAULT_URL_FOR_TEST
with envs.SGLANG_REQ_RUNNING_TIMEOUT.override(
0.001
), envs.SGLANG_ENABLE_HEALTH_ENDPOINT_GENERATION.override(False):
cls.process = popen_launch_server(
cls.model,
cls.base_url,
timeout=DEFAULT_TIMEOUT_FOR_SERVER_LAUNCH,
other_args=["--skip-server-warmup"],
)
@classmethod
def tearDownClass(cls):
kill_process_tree(cls.process.pid)
def test_running_timeout(self):
response = requests.post(
self.base_url + "/generate",
json={
"text": "Today is ",
"sampling_params": {
"temperature": 0,
"max_new_tokens": 512,
"ignore_eos": True,
},
},
)
result = response.json()
self.assertEqual(result["object"], "error")
self.assertEqual(result["code"], 503)
if __name__ == "__main__":
unittest.main()