Adds an option to exit on encounters with unexpected container states

Closes #57

Author: funkyfuture

deck_chores/config.py

@@ -69,6 +69,9 @@ def generate_config() -> None:
     )
     cfg.debug = trueish(getenv('DEBUG', 'no'))
     cfg.default_max = int(getenv('DEFAULT_MAX', 1))
+    cfg.exit_on_unexpected_container_states = trueish(
+        getenv('EXIT_ON_UNEXPECTED_CONTAINER_STATES', 'no')
+    )
     cfg.job_executor_pool_size = int(getenv('JOB_POOL_SIZE', 10))
     cfg.job_name_regex = getenv("JOB_NAME_REGEX", "[a-z0-9-]+")
     cfg.label_ns = getenv('LABEL_NAMESPACE', 'deck-chores') + '.'

deck_chores/jobs.py

@@ -1,4 +1,6 @@
 import logging
+import os
+from signal import SIGUSR2
 from typing import Dict, Iterator, Mapping, Tuple
 
 from apscheduler import events
@@ -83,21 +85,27 @@ def on_missed(event: events.JobExecutionEvent) -> None:
 ####
 
 
+def handle_exec_error(message: str):
+    log.error(message)
+    if cfg.exit_on_unexpected_container_states:
+        os.kill(os.getpid(), SIGUSR2)
+
+
 def exec_job(**definition) -> Tuple[int, bytes]:
     job_id = definition['job_id']
     container_id = definition['container_id']
     log.info(f"{container_name(container_id)}: Executing '{definition['job_name']}'.")
 
-    # some sanity checks, to be removed eventually
+    # some sanity checks
     assert scheduler.get_job(job_id) is not None
     if cfg.client.containers.list(filters={'id': container_id, 'status': 'paused'}):
-        raise AssertionError('Container is paused.')
+        handle_exec_error('Container is paused.')
 
     if not cfg.client.containers.list(
         filters={'id': container_id, 'status': 'running'}
     ):
         assert scheduler.get_job(job_id) is None
-        raise AssertionError('Container is not running.')
+        handle_exec_error('Container is not running.')
     # end of sanity checks
 
     return cfg.client.containers.get(container_id).exec_run(

deck_chores/main.py

@@ -3,7 +3,7 @@
 import os
 import sys
 from datetime import datetime, timedelta, timezone
-from signal import signal, SIGINT, SIGTERM, SIGUSR1
+from signal import signal, SIGINT, SIGTERM, SIGUSR1, SIGUSR2
 from typing import Optional
 
 from apscheduler.schedulers import SchedulerNotRunningError
@@ -62,9 +62,15 @@ def sigusr1_handler(signum, frame):
         log.info(job.kwargs)
 
 
+def sigusr2_handler(signum, frame):
+    log.info("SIGUSR2 received, exiting with error.")
+    raise SystemExit(1)
+
+
 signal(SIGINT, sigint_handler)
 signal(SIGTERM, sigterm_handler)
 signal(SIGUSR1, sigusr1_handler)
+signal(SIGUSR2, sigusr2_handler)
 
 
 ####

docs/usage.rst

@@ -374,6 +374,14 @@ deck-chore's behaviour is defined by these environment variables:
 
     The URL of the Docker daemon to connect to.
 
+.. envvar:: EXIT_ON_UNEXPECTED_CONTAINER_STATES
+
+    default: ``no``
+
+    Exit *deck-chores* when it finds itself confused with unexpected states of containers.
+    If you expect a restart, configure your process manager like *systemd* or the *Docker daemon*
+    accordingly.
+
 .. envvar:: STDERR_LEVEL
 
     default: ``NOTSET``

tests/test_config.py

@@ -32,6 +32,7 @@ def every_file_exists(*args, **kwargs):
         'debug': False,
         'default_max': 1,
         'default_flags': ('image', 'service'),
+        'exit_on_unexpected_container_states': False,
         'job_executor_pool_size': 10,
         'job_name_regex': '[a-z0-9-]+',
         'label_ns': 'deck-chores.',