def wait_for_daemon_dead(self, name, timeout=10):
deadline = time.time() + timeout
while time.time() < deadline:
stats = utils.get_agent_stats(name, self.celery)
if not stats:
return
self.logger.info('Waiting for daemon {0} to stop...'
.format(name))
time.sleep(1)
raise RuntimeError('Failed waiting for daemon {0} to stop. Waited '
'for {1} seconds'.format(name, timeout))
def start(self,
interval=defaults.START_INTERVAL,
timeout=defaults.START_TIMEOUT,
delete_amqp_queue=defaults.DELETE_AMQP_QUEUE_BEFORE_START):
"""
Starts the daemon process.
:param interval: the interval in seconds to sleep when waiting for
the daemon to be ready.
:param timeout: the timeout in seconds to wait for the daemon to be
ready.
:param delete_amqp_queue: delete any queues with the name of the
current daemon queue in the broker.
:raise DaemonStartupTimeout: in case the agent failed to start in the
given amount of time.
:raise DaemonException: in case an error happened during the agent
startup.
"""
if delete_amqp_queue:
self._logger.debug('Deleting AMQP queues')
self._delete_amqp_queues()
start_command = self.start_command()
self._logger.info('Starting daemon with command: {0}'
.format(start_command))
self._runner.run(start_command)
end_time = time.time() + timeout
while time.time() < end_time:
self._logger.debug('Querying daemon {0} stats'.format(self.name))
stats = utils.get_agent_stats(self.name, self._celery)
if stats:
# make sure the status command recognizes the daemon is up
status = self.status()
if status:
self._logger.debug('Daemon {0} has started'
.format(self.name))
return
self._logger.debug('Daemon {0} has not started yet. '
'Sleeping for {1} seconds...'
.format(self.name, interval))
time.sleep(interval)
self._logger.debug('Verifying there were no un-handled '
'exception during startup')
self._verify_no_celery_error()
raise exceptions.DaemonStartupTimeout(timeout, self.name)
def delete(self, force=defaults.DAEMON_FORCE_DELETE):
self._logger.debug('Retrieving daemon stats')
stats = utils.get_agent_stats(self.name, self._celery)
if stats:
if not force:
raise exceptions.DaemonStillRunningException(self.name)
self.stop()
self._logger.info('Removing {0} service'.format(
self.name))
self._runner.run('{0} remove {1} confirm'.format(
self.nssm_path,
self.name))
self._logger.debug('Deleting {0}'.format(self.config_path))
if os.path.exists(self.config_path):
os.remove(self.config_path)
def stop(self,
interval=defaults.STOP_INTERVAL,
timeout=defaults.STOP_TIMEOUT):
"""
Stops the daemon process.
:param interval: the interval in seconds to sleep when waiting for
the daemon to stop.
:param timeout: the timeout in seconds to wait for the daemon to stop.
:raise DaemonShutdownTimeout: in case the agent failed to be stopped
in the given amount of time.
:raise DaemonException: in case an error happened during the agent
shutdown.
"""
stop_command = self.stop_command()
self._logger.info('Stopping daemon with command: {0}'
.format(stop_command))
self._runner.run(stop_command)
end_time = time.time() + timeout
while time.time() < end_time:
self._logger.debug('Querying daemon {0} stats'.format(self.name))
# check the process has shutdown
stats = utils.get_agent_stats(self.name, self._celery)
if not stats:
# make sure the status command also recognizes the
# daemon is down
status = self.status()
if not status:
self._logger.debug('Daemon {0} has shutdown'
.format(self.name, interval))
self._logger.debug('Deleting AMQP queues')
self._delete_amqp_queues()
return
self._logger.debug('Daemon {0} is still running. '
'Sleeping for {1} seconds...'
.format(self.name, interval))
time.sleep(interval)
self._logger.debug('Verifying there were no un-handled '
'exception during startup')
self._verify_no_celery_error()
raise exceptions.DaemonShutdownTimeout(timeout, self.name)
def delete(self, force=defaults.DAEMON_FORCE_DELETE):
self._logger.debug('Retrieving daemon stats')
stats = utils.get_agent_stats(self.name, self._celery)
if stats:
if not force:
raise exceptions.DaemonStillRunningException(self.name)
self.stop()
if os.path.exists(self.script_path):
self._logger.debug('Deleting {0}'.format(self.script_path))
self._runner.run('sudo rm {0}'.format(self.script_path))
if os.path.exists(self.config_path):
self._logger.debug('Deleting {0}'.format(self.config_path))
self._runner.run('sudo rm {0}'.format(self.config_path))
if os.path.exists(self.includes_path):
self._logger.debug('Deleting {0}'.format(self.includes_path))
self._runner.run('sudo rm {0}'.format(self.includes_path))
def start(cloudify_agent, installer, **_):
if ctx.type == context.NODE_INSTANCE:
remote_execution = _get_remote_execution()
if remote_execution:
ctx.logger.info("Starting Agent {0}".format(cloudify_agent["name"]))
installer.start_agent()
else:
# if remote_execution is False, and this operation was invoked
# (install_agent is True), it means that some other process is
# installing the agent (e.g userdata). All that is left for us
# to do is wait for the agent to start.
stats = utils.get_agent_stats(cloudify_agent["name"], app)
if stats:
ctx.logger.info("Agent has started")
else:
return ctx.operation.retry(message="Waiting for Agent to start...", retry_after=5)
else:
ctx.logger.info("Starting Agent {0}".format(cloudify_agent["name"]))
installer.start_agent()
def assert_daemon_dead(self, name):
stats = utils.get_agent_stats(name, self.celery)
self.assertTrue(stats is None)
