def wait(self):
"""Loop waiting on children to die and respawning as necessary."""
systemd.notify_once()
LOG.debug('Full set of CONF:')
CONF.log_opt_values(LOG, logging.DEBUG)
try:
while True:
self.handle_signal()
self._respawn_children()
# No signal means that stop was called. Don't clean up here.
if not self.sigcaught:
return
signame = _signo_to_signame(self.sigcaught)
LOG.info(_LI('Caught %s, stopping children'), signame)
if not _is_sighup_and_daemon(self.sigcaught):
break
cfg.CONF.reload_config_files()
for service in set(
[wrap.service for wrap in self.children.values()]):
service.reset()
for pid in self.children:
os.kill(pid, signal.SIGHUP)
self.running = True
self.sigcaught = None
except eventlet.greenlet.GreenletExit:
LOG.info(_LI("Wait called after thread killed. Cleaning up."))
self.stop()
def _start_child(self, wrap):
if len(wrap.forktimes) > wrap.workers:
# Limit ourselves to one process a second (over the period of
# number of workers * 1 second). This will allow workers to
# start up quickly but ensure we don't fork off children that
# die instantly too quickly.
if time.time() - wrap.forktimes[0] < wrap.workers:
LOG.info(_LI('Forking too fast, sleeping'))
time.sleep(1)
wrap.forktimes.pop(0)
wrap.forktimes.append(time.time())
pid = os.fork()
if pid == 0:
launcher = self._child_process(wrap.service)
while True:
self._child_process_handle_signal()
status, signo = self._child_wait_for_exit_or_signal(launcher)
if not _is_sighup_and_daemon(signo):
break
launcher.restart()
os._exit(status)
LOG.info(_LI('Started child %d'), pid)
wrap.children.add(pid)
self.children[pid] = wrap
return pid
def _wait_child(self):
try:
# Don't block if no child processes have exited
pid, status = os.waitpid(0, os.WNOHANG)
if not pid:
return None
except OSError as exc:
if exc.errno not in (errno.EINTR, errno.ECHILD):
raise
return None
if os.WIFSIGNALED(status):
sig = os.WTERMSIG(status)
LOG.info(_LI('Child %(pid)d killed by signal %(sig)d'),
dict(pid=pid, sig=sig))
else:
code = os.WEXITSTATUS(status)
LOG.info(_LI('Child %(pid)s exited with status %(code)d'),
dict(pid=pid, code=code))
if pid not in self.children:
LOG.warning(_LW('pid %d not in child list'), pid)
return None
wrap = self.children.pop(pid)
wrap.children.remove(pid)
return wrap
def _add_periodic_task(cls, task):
"""Add a periodic task to the list of periodic tasks.
The task should already be decorated by @periodic_task.
:return: whether task was actually enabled
"""
name = task._periodic_name
if task._periodic_spacing < 0:
LOG.info(_LI('Skipping periodic task %(task)s because '
'its interval is negative'),
{'task': name})
return False
if not task._periodic_enabled:
LOG.info(_LI('Skipping periodic task %(task)s because '
'it is disabled'),
{'task': name})
return False
# A periodic spacing of zero indicates that this task should
# be run on the default interval to avoid running too
# frequently.
if task._periodic_spacing == 0:
task._periodic_spacing = DEFAULT_INTERVAL
cls._periodic_tasks.append((name, task))
cls._periodic_spacing[name] = task._periodic_spacing
return True
def _pipe_watcher(self):
# This will block until the write end is closed when the parent
# dies unexpectedly
self.readpipe.read()
LOG.info(_LI('Parent process has died unexpectedly, exiting'))
sys.exit(1)
def stop(self):
"""Terminate child processes and wait on each."""
self.running = False
for pid in self.children:
try:
os.kill(pid, signal.SIGTERM)
except OSError as exc:
if exc.errno != errno.ESRCH:
raise
# Wait for children to die
if self.children:
LOG.info(_LI('Waiting on %d children to exit'), len(self.children))
while self.children:
self._wait_child()
def _wait_for_exit_or_signal(self, ready_callback=None):
status = None
signo = 0
LOG.debug('Full set of CONF:')
CONF.log_opt_values(LOG, logging.DEBUG)
try:
if ready_callback:
ready_callback()
super(ServiceLauncher, self).wait()
except SignalExit as exc:
signame = _signo_to_signame(exc.signo)
LOG.info(_LI('Caught %s, exiting'), signame)
status = exc.code
signo = exc.signo
except SystemExit as exc:
status = exc.code
finally:
self.stop()
return status, signo
def _child_wait_for_exit_or_signal(self, launcher):
status = 0
signo = 0
# NOTE(johannes): All exceptions are caught to ensure this
# doesn't fallback into the loop spawning children. It would
# be bad for a child to spawn more children.
try:
launcher.wait()
except SignalExit as exc:
signame = _signo_to_signame(exc.signo)
LOG.info(_LI('Child caught %s, exiting'), signame)
status = exc.code
signo = exc.signo
except SystemExit as exc:
status = exc.code
except BaseException:
LOG.exception(_LE('Unhandled exception'))
status = 2
finally:
launcher.stop()
return status, signo
def launch_service(self, service, workers=1):
wrap = ServiceWrapper(service, workers)
LOG.info(_LI('Starting %d workers'), wrap.workers)
while self.running and len(wrap.children) < wrap.workers:
self._start_child(wrap)
