[Misc] Deprecation Warning when setting --engine-use-ray (#7424)
Signed-off-by: Wallas Santos <wallashss@ibm.com> Co-authored-by: youkaichao <youkaichao@gmail.com> Co-authored-by: Nick Hill <nickhill@us.ibm.com> Co-authored-by: youkaichao <youkaichao@126.com>
This commit is contained in:
co-authored by
youkaichao
Nick Hill
youkaichao
parent
67d115db08
commit
70b746efcf
@@ -923,7 +923,13 @@ class AsyncEngineArgs(EngineArgs):
|
||||
parser.add_argument('--engine-use-ray',
|
||||
action='store_true',
|
||||
help='Use Ray to start the LLM engine in a '
|
||||
'separate process as the server process.')
|
||||
'separate process as the server process.'
|
||||
'(DEPRECATED. This argument is deprecated '
|
||||
'and will be removed in a future update. '
|
||||
'Set `VLLM_ALLOW_ENGINE_USE_RAY=1` to force '
|
||||
'use it. See '
|
||||
'https://github.com/vllm-project/vllm/issues/7045.'
|
||||
')')
|
||||
parser.add_argument('--disable-log-requests',
|
||||
action='store_true',
|
||||
help='Disable logging requests.')
|
||||
|
||||
@@ -29,6 +29,7 @@ from vllm.prompt_adapter.request import PromptAdapterRequest
|
||||
from vllm.sampling_params import SamplingParams
|
||||
from vllm.sequence import ExecuteModelRequest, SamplerOutput
|
||||
from vllm.usage.usage_lib import UsageContext
|
||||
from vllm.utils import print_warning_once
|
||||
|
||||
logger = init_logger(__name__)
|
||||
ENGINE_ITERATION_TIMEOUT_S = envs.VLLM_ENGINE_ITERATION_TIMEOUT_S
|
||||
@@ -510,6 +511,20 @@ class AsyncLLMEngine:
|
||||
self.log_requests = log_requests
|
||||
self.engine = self._init_engine(*args, **kwargs)
|
||||
|
||||
if self.engine_use_ray:
|
||||
print_warning_once(
|
||||
"DEPRECATED. `--engine-use-ray` is deprecated and will "
|
||||
"be removed in a future update. "
|
||||
"See https://github.com/vllm-project/vllm/issues/7045.")
|
||||
|
||||
if envs.VLLM_ALLOW_ENGINE_USE_RAY:
|
||||
print_warning_once(
|
||||
"VLLM_ALLOW_ENGINE_USE_RAY is set, force engine use Ray")
|
||||
else:
|
||||
raise ValueError("`--engine-use-ray` is deprecated. "
|
||||
"Set `VLLM_ALLOW_ENGINE_USE_RAY=1` to "
|
||||
"force use it")
|
||||
|
||||
self.background_loop: Optional[asyncio.Future] = None
|
||||
# We need to keep a reference to unshielded
|
||||
# task as well to prevent it from being garbage
|
||||
|
||||
@@ -55,6 +55,7 @@ if TYPE_CHECKING:
|
||||
VERBOSE: bool = False
|
||||
VLLM_ALLOW_LONG_MAX_MODEL_LEN: bool = False
|
||||
VLLM_TEST_FORCE_FP8_MARLIN: bool = False
|
||||
VLLM_ALLOW_ENGINE_USE_RAY: bool = False
|
||||
VLLM_PLUGINS: Optional[List[str]] = None
|
||||
|
||||
|
||||
@@ -364,6 +365,14 @@ environment_variables: Dict[str, Callable[[], Any]] = {
|
||||
(os.environ.get("VLLM_TEST_FORCE_FP8_MARLIN", "0").strip().lower() in
|
||||
("1", "true")),
|
||||
|
||||
# If set, allow running the engine as a separate ray actor,
|
||||
# which is a deprecated feature soon to be removed.
|
||||
# See https://github.com/vllm-project/vllm/issues/7045
|
||||
"VLLM_ALLOW_ENGINE_USE_RAY":
|
||||
lambda:
|
||||
(os.environ.get("VLLM_ALLOW_ENGINE_USE_RAY", "0").strip().lower() in
|
||||
("1", "true")),
|
||||
|
||||
# a list of plugin names to load, separated by commas.
|
||||
# if this is not set, it means all plugins will be loaded
|
||||
# if this is set to an empty string, no plugins will be loaded
|
||||
|
||||
Reference in New Issue
Block a user