mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2026-04-23 08:21:53 +08:00
[PD Disaggregation] Add unittest for splitwise deployment with using rdma (#5189)
CE Compile Job / ce_job_pre_check (push) Has been cancelled
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
CE Compile Job / Show Code Archive Output (push) Has been cancelled
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Deploy GitHub Pages / deploy (push) Has been cancelled
CE Compile Job / ce_job_pre_check (push) Has been cancelled
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
CE Compile Job / Show Code Archive Output (push) Has been cancelled
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Deploy GitHub Pages / deploy (push) Has been cancelled
* Add splitwise deployment with using rdma * clean cuda
This commit is contained in:
@@ -30,7 +30,7 @@ from utils.serving_utils import (
|
||||
FD_CACHE_QUEUE_PORT,
|
||||
FD_ENGINE_QUEUE_PORT,
|
||||
FD_METRICS_PORT,
|
||||
clean_ports,
|
||||
clean,
|
||||
get_registered_number,
|
||||
)
|
||||
|
||||
@@ -64,7 +64,7 @@ def setup_and_run_server():
|
||||
- Tears down server after all tests finish
|
||||
"""
|
||||
print("Pre-test port cleanup...")
|
||||
clean_ports(PORTS_TO_CLEAN)
|
||||
clean(PORTS_TO_CLEAN)
|
||||
|
||||
print("log dir clean ")
|
||||
if os.path.exists("log_router") and os.path.isdir("log_router"):
|
||||
@@ -111,7 +111,7 @@ def setup_and_run_server():
|
||||
env_prefill["CUDA_VISIBLE_DEVICES"] = "0"
|
||||
env_prefill["ENABLE_V1_KVCACHE_SCHEDULER"] = "0"
|
||||
env_prefill["FD_LOG_DIR"] = "log_prefill"
|
||||
prefill_log_path = "server.log"
|
||||
prefill_log_path = "server_prefill.log"
|
||||
prefill_cmd = [
|
||||
sys.executable,
|
||||
"-m",
|
||||
@@ -161,7 +161,7 @@ def setup_and_run_server():
|
||||
env_decode["CUDA_VISIBLE_DEVICES"] = "1"
|
||||
env_decode["ENABLE_V1_KVCACHE_SCHEDULER"] = "0"
|
||||
env_decode["FD_LOG_DIR"] = "log_decode"
|
||||
decode_log_path = "decode_server.log"
|
||||
decode_log_path = "server_decode.log"
|
||||
decode_cmd = [
|
||||
sys.executable,
|
||||
"-m",
|
||||
@@ -216,7 +216,7 @@ def setup_and_run_server():
|
||||
try:
|
||||
os.killpg(process_prefill.pid, signal.SIGTERM)
|
||||
os.killpg(process_decode.pid, signal.SIGTERM)
|
||||
clean_ports()
|
||||
clean()
|
||||
except Exception as e:
|
||||
print(f"Failed to kill process group: {e}")
|
||||
raise RuntimeError(f"API server did not start on port {FD_API_PORT}")
|
||||
@@ -228,7 +228,7 @@ def setup_and_run_server():
|
||||
os.killpg(process_router.pid, signal.SIGTERM)
|
||||
os.killpg(process_prefill.pid, signal.SIGTERM)
|
||||
os.killpg(process_decode.pid, signal.SIGTERM)
|
||||
clean_ports(PORTS_TO_CLEAN)
|
||||
clean(PORTS_TO_CLEAN)
|
||||
print(f"Prefill server (pid={process_prefill.pid}) terminated")
|
||||
print(f"Decode server (pid={process_decode.pid}) terminated")
|
||||
except Exception as e:
|
||||
|
||||
Reference in New Issue
Block a user