[ci][distributed] shorten wait time if server hangs (vllm-project#7098)

ruisearch42 · Aug 3, 2024 · 69ea15e · 69ea15e
1 parent ed812a7
commit 69ea15e
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/tests/utils.py b/tests/utils.py
@@ -50,7 +50,7 @@ def _nvml():
 
 class RemoteOpenAIServer:
     DUMMY_API_KEY = "token-abc123"  # vLLM's OpenAI server does not need API key
-    MAX_SERVER_START_WAIT_S = 600  # wait for server to start for 60 seconds
+    MAX_SERVER_START_WAIT_S = 120  # wait for server to start for 120 seconds
 
     def __init__(
         self,