salt-master defunct processes

Description of Issue

I have a number of defunct processes which increment and eventually cause the host to run out of memory, requiring a reboot. There is a scheduled runner which executes every 30 seconds, if I increase that interval, I've noticed the rate of defunct processes reduces too.

I do not see any references to an exception in the debug file.

Setup

Salt master is running in an Alpine docker container.

Steps to Reproduce Issue

Schedule -

schedule:
  runner_queue:
    function: queue.process_runner
    seconds: 30
    kwargs:
      queue: runners
      quantity: 1
      backend: sqlite

: ps aux | grep defunct | more
root       479  0.0  0.0      0     0 ?        Zs   06:55   0:00 [salt-master] <defunct>
root       480  0.0  0.0      0     0 ?        Z    06:55   0:00 [salt-master] <defunct>
root      2908  0.0  0.0      0     0 ?        Zs   06:56   0:00 [salt-master] <defunct>
root      2909  0.0  0.0      0     0 ?        Z    06:56   0:00 [salt-master] <defunct>
root      3134  0.0  0.0      0     0 ?        Zs   06:56   0:00 [salt-master] <defunct>
root      3135  0.2  0.0      0     0 ?        Z    06:56   0:00 [salt-master] <defunct>
root     19384  0.0  0.0      0     0 ?        Zs   06:36   0:00 [salt-master] <defunct>
root     19385  0.0  0.0      0     0 ?        Z    06:36   0:00 [salt-master] <defunct>
root     19569  0.0  0.0      0     0 ?        Zs   06:36   0:00 [salt-master] <defunct>
root     19570  0.0  0.0      0     0 ?        Z    06:36   0:00 [salt-master] <defunct>
root     20046  0.0  0.0      0     0 ?        Zs   06:37   0:00 [salt-master] <defunct>

: ps aux | grep defunct | wc -l
85

Versions Report

/var/log/salt # salt --versions-report
Salt Version:
           Salt: 2019.2.0

Dependency Versions:
           cffi: 1.12.3
       cherrypy: Not Installed
       dateutil: Not Installed
      docker-py: Not Installed
          gitdb: Not Installed
      gitpython: Not Installed
          ioflo: Not Installed
         Jinja2: 2.10
        libgit2: Not Installed
        libnacl: Not Installed
       M2Crypto: Not Installed
           Mako: Not Installed
   msgpack-pure: Not Installed
 msgpack-python: 0.6.1
   mysql-python: Not Installed
      pycparser: 2.19
       pycrypto: 2.6.1
   pycryptodome: Not Installed
         pygit2: Not Installed
         Python: 3.6.8 (default, Apr  8 2019, 18:17:52)
   python-gnupg: Not Installed
         PyYAML: 5.1
          PyZMQ: 18.0.1
           RAET: Not Installed
          smmap: Not Installed
        timelib: Not Installed
        Tornado: 4.5.3
            ZMQ: 4.3.1

System Versions:
           dist:
         locale: UTF-8
        machine: x86_64
        release: 3.10.0-693.el7.x86_64
         system: Linux
        version: Not Installed

Start by reproducing the scheduled runner_queue configuration with queue.process_runner in the reported Alpine Docker environment, then inspect process creation and cleanup around that runner. Compare the defunct-process count over repeated 30-second runs and after increasing the interval; done means the schedule no longer accumulates salt-master zombie processes.