def add_active_task(self, task_id):
if task_id in self.finished_tasks:
log.error("Found an active task (%s) in finished tasks?" % task_id)
return
task_monitor = TaskMonitor(self._pathspec, task_id)
if not task_monitor.get_state().header:
log.info('Unable to load task "%s"' % task_id)
return
sandbox = task_monitor.get_state().header.sandbox
resource_monitor = self._resource_monitor(task_monitor, sandbox)
resource_monitor.start()
self._active_tasks[task_id] = ActiveObservedTask(
task_id=task_id, pathspec=self._pathspec, task_monitor=task_monitor, resource_monitor=resource_monitor
)
def add_active_task(self, task_id):
if task_id in self.finished_tasks:
log.error('Found an active task (%s) in finished tasks?' % task_id)
return
task_monitor = TaskMonitor(self._pathspec, task_id)
if not task_monitor.get_state().header:
log.info('Unable to load task "%s"' % task_id)
return
sandbox = task_monitor.get_state().header.sandbox
resource_monitor = self._resource_monitor(task_monitor, sandbox)
resource_monitor.start()
self._active_tasks[task_id] = ActiveObservedTask(
task_id=task_id, pathspec=self._pathspec,
task_monitor=task_monitor, resource_monitor=resource_monitor
)
def test_basic_as_job(self):
proxy_driver = ProxyDriver()
with temporary_dir() as tempdir:
te = AuroraExecutor(runner_provider=make_provider(tempdir),
sandbox_provider=DefaultTestSandboxProvider())
te.launchTask(proxy_driver,
make_task(MESOS_JOB(task=HELLO_WORLD), instanceId=0))
te.runner_started.wait()
while te._status_manager is None:
time.sleep(0.1)
te.terminated.wait()
tm = TaskMonitor(TaskPath(root=tempdir),
task_id=HELLO_WORLD_TASK_ID)
runner_state = tm.get_state()
assert 'hello_world_hello_world-001' in runner_state.processes, (
'Could not find processes, got: %s' %
' '.join(runner_state.processes))
updates = proxy_driver.method_calls['sendStatusUpdate']
assert len(updates) == 3
status_updates = [arg_tuple[0][0] for arg_tuple in updates]
assert status_updates[0].state == mesos_pb2.TASK_STARTING
assert status_updates[1].state == mesos_pb2.TASK_RUNNING
assert status_updates[2].state == mesos_pb2.TASK_FINISHED
def test_pg_is_killed(self):
runner = self.start_runner()
tm = TaskMonitor(runner.tempdir, runner.task_id)
self.wait_until_running(tm)
process_state, run_number = tm.get_active_processes()[0]
assert process_state.process == 'process'
assert run_number == 0
child_pidfile = os.path.join(runner.sandbox, runner.task_id,
'child.txt')
while not os.path.exists(child_pidfile):
time.sleep(0.1)
parent_pidfile = os.path.join(runner.sandbox, runner.task_id,
'parent.txt')
while not os.path.exists(parent_pidfile):
time.sleep(0.1)
with open(child_pidfile) as fp:
child_pid = int(fp.read().rstrip())
with open(parent_pidfile) as fp:
parent_pid = int(fp.read().rstrip())
ps = ProcessProviderFactory.get()
ps.collect_all()
assert parent_pid in ps.pids()
assert child_pid in ps.pids()
assert child_pid in ps.children_of(parent_pid)
with open(os.path.join(runner.sandbox, runner.task_id, 'exit.txt'),
'w') as fp:
fp.write('go away!')
while tm.task_state() is not TaskState.SUCCESS:
time.sleep(0.1)
state = tm.get_state()
assert state.processes['process'][0].state == ProcessState.SUCCESS
# Another test case may have called setup_child_subreaping(). We therefore have to reap any
# terminated re-parented child processes to ensure that we don't list already terminated
# processes (i.e. zombies) in ps.pids() below.
TaskRunnerHelper.reap_children()
ps.collect_all()
assert parent_pid not in ps.pids()
assert child_pid not in ps.pids()
def test_coordinator_dead_kill(self):
runner = self.start_runner()
tm = TaskMonitor(runner.pathspec, runner.task_id)
self.wait_until_running(tm)
process_state, run_number = tm.get_active_processes()[0]
assert process_state.process == 'ignorant_process'
assert run_number == 0
os.kill(runner.po.pid, signal.SIGKILL)
os.kill(process_state.coordinator_pid, signal.SIGKILL)
os.kill(process_state.pid, signal.SIGKILL)
killer = TaskRunner.get(runner.task_id, runner.root)
assert killer is not None
killer.kill(force=True)
state = tm.get_state()
assert len(state.processes['ignorant_process']) == 1
assert state.processes['ignorant_process'][0].state == ProcessState.LOST
def test_coordinator_kill(self):
runner = self.start_runner()
tm = TaskMonitor(runner.tempdir, runner.task_id)
self.wait_until_running(tm)
process_state, run_number = tm.get_active_processes()[0]
assert process_state.process == 'ignorant_process'
assert run_number == 0
os.kill(process_state.coordinator_pid, signal.SIGKILL)
while True:
active_procs = tm.get_active_processes()
if active_procs and active_procs[0][1] > 0:
break
time.sleep(0.2)
self.wait_until_running(tm)
process_state, run_number = tm.get_active_processes()[0]
assert process_state.process == 'ignorant_process'
assert run_number == 1
os.kill(process_state.pid, signal.SIGKILL)
while True:
active_procs = tm.get_active_processes()
if active_procs and active_procs[0][1] > 1:
break
time.sleep(0.2)
self.wait_until_running(tm)
os.kill(runner.po.pid, signal.SIGKILL)
try:
state = tm.get_state()
assert state.processes['ignorant_process'][
0].state == ProcessState.LOST
assert state.processes['ignorant_process'][
1].state == ProcessState.KILLED
assert state.processes['ignorant_process'][
2].state == ProcessState.RUNNING
finally:
os.kill(state.processes['ignorant_process'][2].coordinator_pid,
signal.SIGKILL)
os.kill(state.processes['ignorant_process'][2].pid, signal.SIGKILL)
def test_coordinator_dead_kill(self):
runner = self.start_runner()
tm = TaskMonitor(runner.tempdir, runner.task_id)
self.wait_until_running(tm)
process_state, run_number = tm.get_active_processes()[0]
assert process_state.process == 'ignorant_process'
assert run_number == 0
os.kill(runner.po.pid, signal.SIGKILL)
os.kill(process_state.coordinator_pid, signal.SIGKILL)
os.kill(process_state.pid, signal.SIGKILL)
killer = TaskRunner.get(runner.task_id, runner.root)
assert killer is not None
killer.kill(force=True)
state = tm.get_state()
assert len(state.processes['ignorant_process']) == 1
assert state.processes['ignorant_process'][
0].state == ProcessState.LOST
def test_basic(self):
proxy_driver = ProxyDriver()
with temporary_dir() as tempdir:
te = AuroraExecutor(runner_provider=make_provider(tempdir), sandbox_provider=DefaultTestSandboxProvider())
te.launchTask(proxy_driver, make_task(HELLO_WORLD_MTI))
te.terminated.wait()
tm = TaskMonitor(tempdir, task_id=HELLO_WORLD_TASK_ID)
runner_state = tm.get_state()
assert "hello_world_hello_world-001" in runner_state.processes, "Could not find processes, got: %s" % " ".join(
runner_state.processes
)
updates = proxy_driver.method_calls["sendStatusUpdate"]
assert len(updates) == 3
status_updates = [arg_tuple[0][0] for arg_tuple in updates]
assert status_updates[0].state == mesos_pb2.TASK_STARTING
assert status_updates[1].state == mesos_pb2.TASK_RUNNING
assert status_updates[2].state == mesos_pb2.TASK_FINISHED
def test_basic(self):
proxy_driver = ProxyDriver()
with temporary_dir() as tempdir:
te = AuroraExecutor(runner_provider=make_provider(tempdir),
sandbox_provider=DefaultTestSandboxProvider())
te.launchTask(proxy_driver, make_task(HELLO_WORLD_MTI))
te.terminated.wait()
tm = TaskMonitor(tempdir, task_id=HELLO_WORLD_TASK_ID)
runner_state = tm.get_state()
assert 'hello_world_hello_world-001' in runner_state.processes, (
'Could not find processes, got: %s' %
' '.join(runner_state.processes))
updates = proxy_driver.method_calls['sendStatusUpdate']
assert len(updates) == 3
status_updates = [arg_tuple[0][0] for arg_tuple in updates]
assert status_updates[0].state == mesos_pb2.TASK_STARTING
assert status_updates[1].state == mesos_pb2.TASK_RUNNING
assert status_updates[2].state == mesos_pb2.TASK_FINISHED
def test_pg_is_killed(self):
runner = self.start_runner()
tm = TaskMonitor(runner.tempdir, runner.task_id)
self.wait_until_running(tm)
process_state, run_number = tm.get_active_processes()[0]
assert process_state.process == 'process'
assert run_number == 0
child_pidfile = os.path.join(runner.sandbox, runner.task_id,
'child.txt')
while not os.path.exists(child_pidfile):
time.sleep(0.1)
parent_pidfile = os.path.join(runner.sandbox, runner.task_id,
'parent.txt')
while not os.path.exists(parent_pidfile):
time.sleep(0.1)
with open(child_pidfile) as fp:
child_pid = int(fp.read().rstrip())
with open(parent_pidfile) as fp:
parent_pid = int(fp.read().rstrip())
ps = ProcessProviderFactory.get()
ps.collect_all()
assert parent_pid in ps.pids()
assert child_pid in ps.pids()
assert child_pid in ps.children_of(parent_pid)
with open(os.path.join(runner.sandbox, runner.task_id, 'exit.txt'),
'w') as fp:
fp.write('go away!')
while tm.task_state() is not TaskState.SUCCESS:
time.sleep(0.1)
state = tm.get_state()
assert state.processes['process'][0].state == ProcessState.SUCCESS
ps.collect_all()
assert parent_pid not in ps.pids()
assert child_pid not in ps.pids()
def test_coordinator_kill(self):
runner = self.start_runner()
tm = TaskMonitor(runner.pathspec, runner.task_id)
self.wait_until_running(tm)
process_state, run_number = tm.get_active_processes()[0]
assert process_state.process == 'ignorant_process'
assert run_number == 0
os.kill(process_state.coordinator_pid, signal.SIGKILL)
while True:
active_procs = tm.get_active_processes()
if active_procs and active_procs[0][1] > 0:
break
time.sleep(0.2)
self.wait_until_running(tm)
process_state, run_number = tm.get_active_processes()[0]
assert process_state.process == 'ignorant_process'
assert run_number == 1
os.kill(process_state.pid, signal.SIGKILL)
while True:
active_procs = tm.get_active_processes()
if active_procs and active_procs[0][1] > 1:
break
time.sleep(0.2)
self.wait_until_running(tm)
os.kill(runner.po.pid, signal.SIGKILL)
try:
state = tm.get_state()
assert state.processes['ignorant_process'][0].state == ProcessState.LOST
assert state.processes['ignorant_process'][1].state == ProcessState.KILLED
assert state.processes['ignorant_process'][2].state == ProcessState.RUNNING
finally:
os.kill(state.processes['ignorant_process'][2].coordinator_pid, signal.SIGKILL)
os.kill(state.processes['ignorant_process'][2].pid, signal.SIGKILL)
def test_basic_as_job(self):
proxy_driver = ProxyDriver()
with temporary_dir() as tempdir:
te = AuroraExecutor(
runner_provider=make_provider(tempdir),
sandbox_provider=DefaultTestSandboxProvider())
te.launchTask(proxy_driver, make_task(MESOS_JOB(task=HELLO_WORLD), instanceId=0))
te.runner_started.wait()
while te._status_manager is None:
time.sleep(0.1)
te.terminated.wait()
tm = TaskMonitor(tempdir, task_id=HELLO_WORLD_TASK_ID)
runner_state = tm.get_state()
assert 'hello_world_hello_world-001' in runner_state.processes, (
'Could not find processes, got: %s' % ' '.join(runner_state.processes))
updates = proxy_driver.method_calls['sendStatusUpdate']
assert len(updates) == 3
status_updates = [arg_tuple[0][0] for arg_tuple in updates]
assert status_updates[0].state == mesos_pb2.TASK_STARTING
assert status_updates[1].state == mesos_pb2.TASK_RUNNING
assert status_updates[2].state == mesos_pb2.TASK_FINISHED
def test_pg_is_killed(self):
runner = self.start_runner()
tm = TaskMonitor(runner.pathspec, runner.task_id)
self.wait_until_running(tm)
process_state, run_number = tm.get_active_processes()[0]
assert process_state.process == 'process'
assert run_number == 0
child_pidfile = os.path.join(runner.sandbox, runner.task_id, 'child.txt')
while not os.path.exists(child_pidfile):
time.sleep(0.1)
parent_pidfile = os.path.join(runner.sandbox, runner.task_id, 'parent.txt')
while not os.path.exists(parent_pidfile):
time.sleep(0.1)
with open(child_pidfile) as fp:
child_pid = int(fp.read().rstrip())
with open(parent_pidfile) as fp:
parent_pid = int(fp.read().rstrip())
ps = ProcessProviderFactory.get()
ps.collect_all()
assert parent_pid in ps.pids()
assert child_pid in ps.pids()
assert child_pid in ps.children_of(parent_pid)
with open(os.path.join(runner.sandbox, runner.task_id, 'exit.txt'), 'w') as fp:
fp.write('go away!')
while tm.task_state() is not TaskState.SUCCESS:
time.sleep(0.1)
state = tm.get_state()
assert state.processes['process'][0].state == ProcessState.SUCCESS
ps.collect_all()
assert parent_pid not in ps.pids()
assert child_pid not in ps.pids()
