MCPcopy Create free account
hub / github.com/NVIDIA/TensorRT-LLM / pre_shutdown

Method pre_shutdown

tensorrt_llm/executor/proxy.py:276–290  ·  view source on GitHub ↗
(self)

Source from the content-addressed store, hash-verified

274 result.abort()
275
276 def pre_shutdown(self):
277 if not self.workers_started:
278 return
279 logger_debug('Proxy.pre_shutdown...\n', "yellow")
280
281 if self.doing_shutdown:
282 return
283 else:
284 self.doing_shutdown = True
285
286 self._abort_all_requests()
287
288 # notify the workers to quit
289 if all(not f.done() for f in self.mpi_futures):
290 self.request_queue.put_noblock(None, retry=4)
291
292 def shutdown(self):
293 if not self.workers_started:

Callers 1

shutdownMethod · 0.95

Calls 3

_abort_all_requestsMethod · 0.95
logger_debugFunction · 0.85
put_noblockMethod · 0.80

Tested by

no test coverage detected