[CI][Bugfix] Rerun test_engine_log_metrics_ray on Ray GCS startup timeout (#47208)

Signed-off-by: pei.zhang <pei.zhang@amd.com>
Co-authored-by: Claude <noreply@anthropic.com>
This commit is contained in:
peizhang56
2026-07-01 21:32:09 -05:00
committed by GitHub
co-authored by Claude
parent 2b753ad200
commit 7fe7fa9cda
+13
View File
@@ -7,6 +7,7 @@ import pytest
import ray
from vllm.config.model import ModelDType
from vllm.platforms import current_platform
from vllm.sampling_params import SamplingParams
from vllm.v1.engine.async_llm import AsyncEngineArgs, AsyncLLM
from vllm.v1.metrics.ray_wrappers import (
@@ -22,6 +23,18 @@ MODELS = [
]
# The first .remote() call starts a local Ray cluster via ray.init(), whose
# GCS server occasionally fails to start within Ray's fixed 30s bootstrap
# window on ROCm CI (RuntimeError: "Timed out waiting for file
# .../gcs_server_port_..."). That timeout is not configurable, so retry the
# whole test: a fresh ray.init() gets a new GCS process. Scoped to that error
# so real failures still fail immediately.
@pytest.mark.flaky(
reruns=2,
reruns_delay=5,
only_rerun="Timed out waiting for file",
condition=current_platform.is_rocm(),
)
@pytest.mark.parametrize("model", MODELS)
@pytest.mark.parametrize("dtype", ["half"])
@pytest.mark.parametrize("max_tokens", [16])