A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 16, 17:00 UTC
[CI/Build] Increasing timeout for multiproc worker tests (#8203)
Alexey Kondratiev(AMD) committed
Sep 6, 2024 at 14:51 UTC
1447c97e753919709b613590d7267c93d07d9382
1 file changed
+3
-3
tests/engine/test_multiproc_workers.py
+3
-3
@@ -83,7 +83,7 @@ def test_local_workers() -> None:
83
workers[3].process.kill()
84
85
# Other workers should get shut down here
86
- worker_monitor.join(2)
86
+ worker_monitor.join(20)
87
88
# Ensure everything is stopped
89
assert not worker_monitor.is_alive()
@@ -108,7 +108,7 @@ def test_local_workers_clean_shutdown() -> None:
108
# Clean shutdown
109
worker_monitor.close()
110
111
- worker_monitor.join(5)
111
+ worker_monitor.join(20)
112
113
# Ensure everything is stopped
114
assert not worker_monitor.is_alive()
@@ -161,7 +161,7 @@ async def test_local_workers_async() -> None:
161
workers[3].process.kill()
162
163
# Other workers should get shut down here
164
- worker_monitor.join(2)
164
+ worker_monitor.join(20)
165
166
# Ensure everything is stopped
167
assert not worker_monitor.is_alive()