apextoaster
/
onnx-web
1
0
Fork 0
Code Actions

fix(api): track and repeatedly attempt to recycle leaking workers (#219)

This commit is contained in:
Sean Sube 2023-03-05 18:58:13 -06:00
parent ccd15c1921
commit 7a3a81a4ef
Signed by: ssube
GPG Key ID: 3EED7B957D362AF1
1 changed files with 22 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
api/onnx_web/worker/pool.py
View File

@ -21,6 +21,7 @@ class DevicePoolExecutor:
max_pending_per_worker: int
join_timeout: float
leaking: List[Tuple[str, Process]]
context: Dict[str, WorkerContext] # Device -> Context
pending: Dict[str, "Queue[Tuple[str, Callable[..., None], Any, Any]]"]
threads: Dict[str, Thread]
@ -49,6 +50,7 @@ class DevicePoolExecutor:
self.max_pending_per_worker = max_pending_per_worker
self.join_timeout = join_timeout
self.leaking = []
self.context = {}
self.pending = {}
self.threads = {}
@ -133,7 +135,7 @@ class DevicePoolExecutor:
while True:
try:
job, device, value = progress.get(timeout=(self.join_timeout / 2))
logger.info("progress update for job: %s to %s", job, value)
logger.debug("progress update for job: %s to %s", job, value)
self.active_jobs[job] = (device, value)
if job in self.cancelled_jobs:
logger.debug(
@ -270,6 +272,7 @@ class DevicePoolExecutor:
logger.warning(
"worker for device %s could not be stopped in time", device
)
self.leaking.append((device, worker))
else:
logger.debug("worker for device %s has died", device)
@ -281,25 +284,38 @@ class DevicePoolExecutor:
def recycle(self):
logger.debug("recycling worker pool")
if len(self.leaking) > 0:
logger.warning("cleaning up %s leaking workers", len(self.leaking))
for device, worker in self.leaking:
logger.debug("shutting down worker for device %s", device)
worker.join(self.join_timeout)
if worker.is_alive():
logger.error("leaking worker for device %s could not be shut down", device)
self.leaking[:] = [dw for dw in self.leaking if dw[1].is_alive()]
needs_restart = []
for device, proc in self.workers.items():
for device, worker in self.workers.items():
jobs = self.total_jobs.get(device, 0)
if not proc.is_alive():
if not worker.is_alive():
logger.warning("worker for device %s has died", device)
needs_restart.append(device)
elif jobs > self.max_jobs_per_worker:
logger.info(
"shutting down worker for device %s after %s jobs", device, jobs
)
proc.join(self.join_timeout)
if proc.is_alive():
worker.join(self.join_timeout)
if worker.is_alive():
logger.warning(
"worker for device %s could not be recycled in time", device
)
self.leaking.append((device, worker))
else:
del worker
self.workers[device] = None
del proc
needs_restart.append(device)
else:
logger.debug(