Celery: Systemd: recarregar o aipo desliga-o de forma limpa, mas não inicia novamente

Lista de controle

• [x] Eu verifiquei que o problema existe no ramo master do aipo.
• [x] Incluí o resultado de celery -A proj report na edição:

software -> celery:4.1.0 (latentcall) kombu:4.1.0 py:3.5.2
            billiard:3.5.0.3 redis:2.10.6
platform -> system:Linux arch:64bit, ELF imp:CPython
loader   -> celery.loaders.app.AppLoader
settings -> transport:redis results:disabled

CACHES: {
    'default': {   'BACKEND': 'django_redis.cache.RedisCache',
                   'LOCATION': 'redis://localhost:6379/1',
                   'TIMEOUT': 3600}}

CELERY_TASK_COMPRESSION: 'gzip'
CELERY_TASK_IGNORE_RESULT: True
CELERY_ACCEPT_CONTENT: ['pickle', 'json', 'msgpack', 'yaml']
CELERY_BROKER_URL: 'redis://localhost:6379/0'

DEBUG: False
INSTALLED_APPS:
    ('django.contrib.auth',
 'django.contrib.contenttypes',
 'django.contrib.sessions',
 'django.contrib.sites',
 'django.contrib.messages',
 'django.contrib.admin',
 'django.contrib.sitemaps',
 'django.contrib.staticfiles',
 'django.contrib.humanize',
 'django.contrib.redirects',
 'django.contrib.gis',

 'django_extensions',
)

Passos para reproduzir

No Ubuntu 16.04.

1. configurar o aipo via systemd de acordo com os documentos do aipo
2. comece a comer aipo com sudo service <your-celery-service-name> start
3. tente recarregar o aipo via sudo service <your-celery-service-name> reload

Comportamento esperado

O aipo deve recarregar normalmente.

Comportamento real

O aipo não recarrega normalmente. Ele desliga, mas nunca reinicia.

Depois de emitir a solicitação de recarregamento (e subsequente falha ao iniciar), a saída de sudo journalctl -xe não parece mostrar nada de nefasto:

Nov 15 04:33:04 ip-172-31-44-219 sudo[12972]:   ubuntu : TTY=pts/0 ; PWD=/home/ubuntu ; USER=root ; COMMAND=/usr/sbin/service myproj-celery reload
Nov 15 04:33:04 ip-172-31-44-219 sudo[12972]: pam_unix(sudo:session): session opened for user root by ubuntu(uid=0)
Nov 15 04:33:04 ip-172-31-44-219 systemd[1]: Reloading myproj celery worker.
-- Subject: Unit myproj-celery.service has begun reloading its configuration
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit myproj-celery.service has begun reloading its configuration
Nov 15 04:33:07 ip-172-31-44-219 sh[12979]: celery multi v4.1.0 (latentcall)
Nov 15 04:33:07 ip-172-31-44-219 sh[12979]: > Stopping nodes...
Nov 15 04:33:07 ip-172-31-44-219 sh[12979]:         > worker1@ip-172-31-44-219: TERM -> 12962
Nov 15 04:33:07 ip-172-31-44-219 sh[12979]: > Waiting for 1 node -> 12962.....
Nov 15 04:33:07 ip-172-31-44-219 sh[12979]:         > worker1@ip-172-31-44-219: OK
Nov 15 04:33:07 ip-172-31-44-219 sh[12979]: > Restarting node worker1@ip-172-31-44-219: OK
Nov 15 04:33:07 ip-172-31-44-219 sh[12979]: > Waiting for 1 node -> None...
Nov 15 04:33:07 ip-172-31-44-219 sh[12992]: celery multi v4.1.0 (latentcall)
Nov 15 04:33:07 ip-172-31-44-219 sh[12992]: > worker1@ip-172-31-44-219: DOWN
Nov 15 04:33:07 ip-172-31-44-219 systemd[1]: Reloaded myproj celery worker.
-- Subject: Unit myproj-celery.service has finished reloading its configuration
-- Defined-By: systemd
-- Support: http://lists.freedesktop.org/mailman/listinfo/systemd-devel
--
-- Unit myproj-celery.service has finished reloading its configuration
--
-- The result is done.
Nov 15 04:33:07 ip-172-31-44-219 sudo[12972]: pam_unix(sudo:session): session closed for user root

Talvez CELERYD_PID_FILE="/var/run/celery/%N.pid" ser desconhecido seja o problema? Ou seja, não podemos dizer ao Systemd (via PIDFile= ) qual é a localização real do arquivo PID, porque ele é gerado pelo próprio aipo, não?