Files
awx/awx/main/dispatch/periodic.py
Ryan Petrello 5364e78397 switch the periodic scheduler to a child process (instead of a thread)
I have a hunch that our usage of a daemon thread is causing import lock
contention related to https://github.com/ansible/awx/issues/5617
We've encountered similar issues before with threads across dispatcher
processes at fork time, and cpython has had bugs like this in recent
history:

https://bugs.python.org/issue38884

My gut tells me this might be related.

The prior implementation - based on celerybeat - ran its code in
a process (not a thread), and the timing of that merge matches the
period of time we started noticing issues.

Currently testing it to see if it resolves some of the issues we're
seeing.
2020-02-27 12:15:15 -05:00

57 lines
1.8 KiB
Python

import logging
import os
import time
from multiprocessing import Process
from django.conf import settings
from django.db import connections
from schedule import Scheduler
from awx.main.dispatch.worker import TaskWorker
logger = logging.getLogger('awx.main.dispatch.periodic')
class Scheduler(Scheduler):
def run_continuously(self):
idle_seconds = max(
1,
min(self.jobs).period.total_seconds() / 2
)
def run():
ppid = os.getppid()
logger.warn(f'periodic beat started')
while True:
if os.getppid() != ppid:
# if the parent PID changes, this process has been orphaned
# via e.g., segfault or sigkill, we should exit too
pid = os.getpid()
logger.warn(f'periodic beat exiting gracefully pid:{pid}')
raise SystemExit()
try:
for conn in connections.all():
# If the database connection has a hiccup, re-establish a new
# connection
conn.close_if_unusable_or_obsolete()
self.run_pending()
except Exception:
logger.exception(
'encountered an error while scheduling periodic tasks'
)
time.sleep(idle_seconds)
process = Process(target=run)
process.daemon = True
process.start()
def run_continuously():
scheduler = Scheduler()
for task in settings.CELERYBEAT_SCHEDULE.values():
apply_async = TaskWorker.resolve_callable(task['task']).apply_async
total_seconds = task['schedule'].total_seconds()
scheduler.every(total_seconds).seconds.do(apply_async)
scheduler.run_continuously()