Error with database cron job locks

[akjackson67]

Occasionally this error is being logged when running cron loop on multiple containers with PostgreSQL database.

Traceback (most recent call last):
  File "/opt/app-root/lib64/python3.9/site-packages/django_cron/management/commands/runcrons.py", line 82, in run_cron_with_cache_check
    manager.run(force)
  File "/opt/app-root/lib64/python3.9/site-packages/django_cron/__init__.py", line 292, in run
    self.stdout.write(u"[ ] {0}\n".format(self.cron_job.code))
  File "/opt/app-root/lib64/python3.9/site-packages/django_cron/backends/lock/base.py", line 66, in __exit__
    self.release()
  File "/usr/lib64/python3.9/contextlib.py", line 79, in inner
    return func(*args, **kwds)
  File "/opt/app-root/lib64/python3.9/site-packages/django_cron/backends/lock/database.py", line 25, in release
    lock.locked = False
AttributeError: 'NoneType' object has no attribute 'locked'

[akjackson67]

I believe this problem is caused because the lock method of database.py allows two or more threads to set the locked attribute to true. If both threads select the record before it is locked, then both threads will believe that it has obtained an exclusive lock on the cron job. This could be corrected by adding a select_for_update to the get_or_create. Like follows:

        lock, created = CronJobLock.objects.select_for_update().get_or_create(job_name=self.job_name)

That will force the second thread to wait until update has been saved.

[guillempuey]

+1

Please, fix this with select_for_update()