#!/usr/bin/python # pylint: disable=missing-docstring import logging, os, signal, unittest import common import mock from autotest_lib.client.common_lib import enum, global_config, host_protections from autotest_lib.database import database_connection from autotest_lib.frontend import setup_django_environment from autotest_lib.frontend.afe import frontend_test_utils, models from autotest_lib.frontend.afe import model_attributes from autotest_lib.scheduler import drone_manager, email_manager from autotest_lib.scheduler import monitor_db, scheduler_models from autotest_lib.scheduler import scheduler_config from autotest_lib.scheduler import scheduler_lib HqeStatus = models.HostQueueEntry.Status HostStatus = models.Host.Status class NullMethodObject(object): _NULL_METHODS = () def __init__(self): def null_method(*args, **kwargs): pass for method_name in self._NULL_METHODS: setattr(self, method_name, null_method) # the SpecialTask names here must match the suffixes used on the SpecialTask # results directories _PidfileType = enum.Enum('verify', 'cleanup', 'repair', 'job', 'gather', 'parse', 'archive', 'reset', 'provision') _PIDFILE_TO_PIDFILE_TYPE = { drone_manager.AUTOSERV_PID_FILE: _PidfileType.JOB, drone_manager.CRASHINFO_PID_FILE: _PidfileType.GATHER, drone_manager.PARSER_PID_FILE: _PidfileType.PARSE, drone_manager.ARCHIVER_PID_FILE: _PidfileType.ARCHIVE, } _PIDFILE_TYPE_TO_PIDFILE = dict((value, key) for key, value in _PIDFILE_TO_PIDFILE_TYPE.iteritems()) class MockConnectionManager(object): """docstring for MockConnectionManager""" db = None def __init__(self): super(MockConnectionManager, self).__init__() def get_connection(self): assert MockConnectionManager.db return MockConnectionManager.db class MockDroneManager(NullMethodObject): """ Public attributes: max_runnable_processes_value: value returned by max_runnable_processes(). tests can change this to activate throttling. """ _NULL_METHODS = ('reinitialize_drones', 'copy_to_results_repository', 'copy_results_on_drone', 'trigger_refresh', 'sync_refresh') class _DummyPidfileId(object): """ Object to represent pidfile IDs that is opaque to the scheduler code but still debugging-friendly for us. """ def __init__(self, working_directory, pidfile_name, num_processes=None): self._working_directory = working_directory self._pidfile_name = pidfile_name self._num_processes = num_processes self._paired_with_pidfile = None def key(self): """Key for MockDroneManager._pidfile_index""" return (self._working_directory, self._pidfile_name) def __str__(self): return os.path.join(self._working_directory, self._pidfile_name) def __repr__(self): return '<_DummyPidfileId: %s>' % str(self) def __init__(self): super(MockDroneManager, self).__init__() self.process_capacity = 100 # maps result_dir to set of tuples (file_path, file_contents) self._attached_files = {} # maps pidfile IDs to PidfileContents self._pidfiles = {} # pidfile IDs that haven't been created yet self._future_pidfiles = [] # maps _PidfileType to the most recently created pidfile ID of that type self._last_pidfile_id = {} # maps (working_directory, pidfile_name) to pidfile IDs self._pidfile_index = {} # maps process to pidfile IDs self._process_index = {} # tracks pidfiles of processes that have been killed self._pids_to_signals_received = {} # pidfile IDs that have just been unregistered (so will disappear on the # next cycle) self._unregistered_pidfiles = set() # Pids to write exit status for at end of tick self._set_pidfile_exit_status_queue = [] # utility APIs for use by the test def finish_process(self, pidfile_type, exit_status=0): pidfile_id = self._last_pidfile_id[pidfile_type] self._set_pidfile_exit_status(pidfile_id, exit_status) def finish_specific_process(self, working_directory, pidfile_name): pidfile_id = self.pidfile_from_path(working_directory, pidfile_name) self._set_pidfile_exit_status(pidfile_id, 0) def finish_active_process_on_host(self, host_id): match = 'hosts/host%d/' % host_id for pidfile_id in self.nonfinished_pidfile_ids(): if pidfile_id._working_directory.startswith(match): self._set_pidfile_exit_status(pidfile_id, 0) break else: raise KeyError('No active process matched %s' % match) def _set_pidfile_exit_status(self, pidfile_id, exit_status): assert pidfile_id is not None contents = self._pidfiles[pidfile_id] contents.exit_status = exit_status contents.num_tests_failed = 0 def was_last_process_killed(self, pidfile_type, sigs): pidfile_id = self._last_pidfile_id[pidfile_type] return sigs == self._pids_to_signals_received[pidfile_id] def nonfinished_pidfile_ids(self): return [pidfile_id for pidfile_id, pidfile_contents in self._pidfiles.iteritems() if pidfile_contents.exit_status is None] def running_pidfile_ids(self): return [pidfile_id for pidfile_id in self.nonfinished_pidfile_ids() if self._pidfiles[pidfile_id].process is not None] def pidfile_from_path(self, working_directory, pidfile_name): return self._pidfile_index[(working_directory, pidfile_name)] def attached_files(self, working_directory): """ Return dict mapping path to contents for attached files with specified paths. """ return dict((path, contents) for path, contents in self._attached_files.get(working_directory, []) if path is not None) # DroneManager emulation APIs for use by monitor_db def get_orphaned_autoserv_processes(self): return set() def total_running_processes(self): return sum(pidfile_id._num_processes for pidfile_id in self.nonfinished_pidfile_ids()) def max_runnable_processes(self, username, drone_hostnames_allowed): return self.process_capacity - self.total_running_processes() def refresh(self): for pidfile_id in self._unregistered_pidfiles: # intentionally handle non-registered pidfiles silently self._pidfiles.pop(pidfile_id, None) self._unregistered_pidfiles = set() def execute_actions(self): # executing an "execute_command" causes a pidfile to be created for pidfile_id in self._future_pidfiles: # Process objects are opaque to monitor_db process = object() self._pidfiles[pidfile_id].process = process self._process_index[process] = pidfile_id self._future_pidfiles = [] for pidfile_id in self._set_pidfile_exit_status_queue: self._set_pidfile_exit_status(pidfile_id, 271) self._set_pidfile_exit_status_queue = [] def attach_file_to_execution(self, result_dir, file_contents, file_path=None): self._attached_files.setdefault(result_dir, set()).add((file_path, file_contents)) return 'attach_path' def _initialize_pidfile(self, pidfile_id): if pidfile_id not in self._pidfiles: assert pidfile_id.key() not in self._pidfile_index self._pidfiles[pidfile_id] = drone_manager.PidfileContents() self._pidfile_index[pidfile_id.key()] = pidfile_id def _set_last_pidfile(self, pidfile_id, working_directory, pidfile_name): if working_directory.startswith('hosts/'): # such paths look like hosts/host1/1-verify, we'll grab the end type_string = working_directory.rsplit('-', 1)[1] pidfile_type = _PidfileType.get_value(type_string) else: pidfile_type = _PIDFILE_TO_PIDFILE_TYPE[pidfile_name] self._last_pidfile_id[pidfile_type] = pidfile_id def execute_command(self, command, working_directory, pidfile_name, num_processes, log_file=None, paired_with_pidfile=None, username=None, drone_hostnames_allowed=None): logging.debug('Executing %s in %s', command, working_directory) pidfile_id = self._DummyPidfileId(working_directory, pidfile_name) if pidfile_id.key() in self._pidfile_index: pidfile_id = self._pidfile_index[pidfile_id.key()] pidfile_id._num_processes = num_processes pidfile_id._paired_with_pidfile = paired_with_pidfile self._future_pidfiles.append(pidfile_id) self._initialize_pidfile(pidfile_id) self._pidfile_index[(working_directory, pidfile_name)] = pidfile_id self._set_last_pidfile(pidfile_id, working_directory, pidfile_name) return pidfile_id def get_pidfile_contents(self, pidfile_id, use_second_read=False): if pidfile_id not in self._pidfiles: logging.debug('Request for nonexistent pidfile %s', pidfile_id) return self._pidfiles.get(pidfile_id, drone_manager.PidfileContents()) def is_process_running(self, process): return True def register_pidfile(self, pidfile_id): self._initialize_pidfile(pidfile_id) def unregister_pidfile(self, pidfile_id): self._unregistered_pidfiles.add(pidfile_id) def declare_process_count(self, pidfile_id, num_processes): pidfile_id.num_processes = num_processes def absolute_path(self, path): return 'absolute/' + path def write_lines_to_file(self, file_path, lines, paired_with_process=None): # TODO: record this pass def get_pidfile_id_from(self, execution_tag, pidfile_name): default_pidfile = self._DummyPidfileId(execution_tag, pidfile_name, num_processes=0) return self._pidfile_index.get((execution_tag, pidfile_name), default_pidfile) def kill_process(self, process, sig=signal.SIGKILL): pidfile_id = self._process_index[process] if pidfile_id not in self._pids_to_signals_received: self._pids_to_signals_received[pidfile_id] = set() self._pids_to_signals_received[pidfile_id].add(sig) if signal.SIGKILL == sig: self._set_pidfile_exit_status_queue.append(pidfile_id) class MockEmailManager(NullMethodObject): _NULL_METHODS = ('send_queued_emails', 'send_email') def enqueue_notify_email(self, subject, message): logging.warning('enqueue_notify_email: %s', subject) logging.warning(message) class SchedulerFunctionalTest(unittest.TestCase, frontend_test_utils.FrontendTestMixin): # some number of ticks after which the scheduler is presumed to have # stabilized, given no external changes _A_LOT_OF_TICKS = 10 def setUp(self): self._frontend_common_setup() self._set_stubs() self._set_global_config_values() self._create_dispatcher() logging.basicConfig(level=logging.DEBUG) def _create_dispatcher(self): self.dispatcher = monitor_db.Dispatcher() def tearDown(self): self._database.disconnect() self._frontend_common_teardown() def _set_stubs(self): self.mock_config = global_config.FakeGlobalConfig() self.god.stub_with(global_config, 'global_config', self.mock_config) self.mock_drone_manager = MockDroneManager() drone_manager._set_instance(self.mock_drone_manager) self.mock_email_manager = MockEmailManager() self.god.stub_with(email_manager, 'manager', self.mock_email_manager) self._database = ( database_connection.TranslatingDatabase.get_test_database( translators=scheduler_lib._DB_TRANSLATORS)) self._database.connect(db_type='django') self.god.stub_with(monitor_db, '_db', self._database) self.god.stub_with(scheduler_models, '_db', self._database) MockConnectionManager.db = self._database scheduler_lib.ConnectionManager = MockConnectionManager monitor_db.initialize_globals() scheduler_models.initialize_globals() patcher = mock.patch( 'autotest_lib.scheduler.luciferlib.is_lucifer_enabled', lambda: False) patcher.start() self.addCleanup(patcher.stop) def _set_global_config_values(self): self.mock_config.set_config_value('SCHEDULER', 'pidfile_timeout_mins', 1) self.mock_config.set_config_value('SCHEDULER', 'gc_stats_interval_mins', 999999) self.mock_config.set_config_value('SCHEDULER', 'enable_archiving', True) self.mock_config.set_config_value('SCHEDULER', 'clean_interval_minutes', 60) self.mock_config.set_config_value('SCHEDULER', 'max_parse_processes', 50) self.mock_config.set_config_value('SCHEDULER', 'max_transfer_processes', 50) self.mock_config.set_config_value('SCHEDULER', 'clean_interval_minutes', 50) self.mock_config.set_config_value('SCHEDULER', 'max_provision_retries', 1) self.mock_config.set_config_value('SCHEDULER', 'max_repair_limit', 1) self.mock_config.set_config_value( 'SCHEDULER', 'secs_to_wait_for_atomic_group_hosts', 600) self.mock_config.set_config_value( 'SCHEDULER', 'inline_host_acquisition', True) scheduler_config.config.read_config() def _initialize_test(self): self.dispatcher.initialize() def _run_dispatcher(self): for _ in xrange(self._A_LOT_OF_TICKS): self.dispatcher.tick() def test_idle(self): self._initialize_test() self._run_dispatcher() def _assert_process_executed(self, working_directory, pidfile_name): process_was_executed = self.mock_drone_manager.was_process_executed( 'hosts/host1/1-verify', drone_manager.AUTOSERV_PID_FILE) self.assert_(process_was_executed, '%s/%s not executed' % (working_directory, pidfile_name)) def _update_instance(self, model_instance): return type(model_instance).objects.get(pk=model_instance.pk) def _check_statuses(self, queue_entry, queue_entry_status, host_status=None): self._check_entry_status(queue_entry, queue_entry_status) if host_status: self._check_host_status(queue_entry.host, host_status) def _check_entry_status(self, queue_entry, status): # update from DB queue_entry = self._update_instance(queue_entry) self.assertEquals(queue_entry.status, status) def _check_host_status(self, host, status): # update from DB host = self._update_instance(host) self.assertEquals(host.status, status) def _run_pre_job_verify(self, queue_entry): self._run_dispatcher() # launches verify self._check_statuses(queue_entry, HqeStatus.VERIFYING, HostStatus.VERIFYING) self.mock_drone_manager.finish_process(_PidfileType.VERIFY) def test_simple_job(self): self._initialize_test() job, queue_entry = self._make_job_and_queue_entry() self._run_pre_job_verify(queue_entry) self._run_dispatcher() # launches job self._check_statuses(queue_entry, HqeStatus.RUNNING, HostStatus.RUNNING) self._finish_job(queue_entry) self._check_statuses(queue_entry, HqeStatus.COMPLETED, HostStatus.READY) self._assert_nothing_is_running() def _setup_for_pre_job_reset(self): self._initialize_test() job, queue_entry = self._make_job_and_queue_entry() job.reboot_before = model_attributes.RebootBefore.ALWAYS job.save() return queue_entry def _run_pre_job_reset_job(self, queue_entry): self._run_dispatcher() # reset self._check_statuses(queue_entry, HqeStatus.RESETTING, HostStatus.RESETTING) self.mock_drone_manager.finish_process(_PidfileType.RESET) self._run_dispatcher() # job self._finish_job(queue_entry) def test_pre_job_reset(self): queue_entry = self._setup_for_pre_job_reset() self._run_pre_job_reset_job(queue_entry) def _run_pre_job_reset_one_failure(self): queue_entry = self._setup_for_pre_job_reset() self._run_dispatcher() # reset self.mock_drone_manager.finish_process(_PidfileType.RESET, exit_status=256) self._run_dispatcher() # repair self._check_statuses(queue_entry, HqeStatus.QUEUED, HostStatus.REPAIRING) self.mock_drone_manager.finish_process(_PidfileType.REPAIR) return queue_entry def test_pre_job_reset_failure(self): queue_entry = self._run_pre_job_reset_one_failure() # from here the job should run as normal self._run_pre_job_reset_job(queue_entry) def test_pre_job_reset_double_failure(self): # TODO (showard): this test isn't perfect. in reality, when the second # reset fails, it copies its results over to the job directory using # copy_results_on_drone() and then parses them. since we don't handle # that, there appear to be no results at the job directory. the # scheduler handles this gracefully, parsing gets effectively skipped, # and this test passes as is. but we ought to properly test that # behavior. queue_entry = self._run_pre_job_reset_one_failure() self._run_dispatcher() # second reset self.mock_drone_manager.finish_process(_PidfileType.RESET, exit_status=256) self._run_dispatcher() self._check_statuses(queue_entry, HqeStatus.FAILED, HostStatus.REPAIR_FAILED) # nothing else should run self._assert_nothing_is_running() def _assert_nothing_is_running(self): self.assertEquals(self.mock_drone_manager.running_pidfile_ids(), []) def _setup_for_post_job_cleanup(self): self._initialize_test() job, queue_entry = self._make_job_and_queue_entry() job.reboot_after = model_attributes.RebootAfter.ALWAYS job.save() return queue_entry def _run_post_job_cleanup_failure_up_to_repair(self, queue_entry, include_verify=True): if include_verify: self._run_pre_job_verify(queue_entry) self._run_dispatcher() # job self.mock_drone_manager.finish_process(_PidfileType.JOB) self._run_dispatcher() # parsing + cleanup self.mock_drone_manager.finish_process(_PidfileType.PARSE) self.mock_drone_manager.finish_process(_PidfileType.CLEANUP, exit_status=256) self._run_dispatcher() # repair, HQE unaffected return queue_entry def test_post_job_cleanup_failure(self): queue_entry = self._setup_for_post_job_cleanup() self._run_post_job_cleanup_failure_up_to_repair(queue_entry) self._check_statuses(queue_entry, HqeStatus.COMPLETED, HostStatus.REPAIRING) self.mock_drone_manager.finish_process(_PidfileType.REPAIR) self._run_dispatcher() self._check_statuses(queue_entry, HqeStatus.COMPLETED, HostStatus.READY) def test_post_job_cleanup_failure_repair_failure(self): queue_entry = self._setup_for_post_job_cleanup() self._run_post_job_cleanup_failure_up_to_repair(queue_entry) self.mock_drone_manager.finish_process(_PidfileType.REPAIR, exit_status=256) self._run_dispatcher() self._check_statuses(queue_entry, HqeStatus.COMPLETED, HostStatus.REPAIR_FAILED) def _ensure_post_job_process_is_paired(self, queue_entry, pidfile_type): pidfile_name = _PIDFILE_TYPE_TO_PIDFILE[pidfile_type] queue_entry = self._update_instance(queue_entry) pidfile_id = self.mock_drone_manager.pidfile_from_path( queue_entry.execution_path(), pidfile_name) self.assert_(pidfile_id._paired_with_pidfile) def _finish_job(self, queue_entry): self._check_statuses(queue_entry, HqeStatus.RUNNING) self.mock_drone_manager.finish_process(_PidfileType.JOB) self._run_dispatcher() # launches parsing self._check_statuses(queue_entry, HqeStatus.PARSING) self._ensure_post_job_process_is_paired(queue_entry, _PidfileType.PARSE) self._finish_parsing() def _finish_parsing(self): self.mock_drone_manager.finish_process(_PidfileType.PARSE) self._run_dispatcher() def _create_reverify_request(self): host = self.hosts[0] models.SpecialTask.schedule_special_task( host=host, task=models.SpecialTask.Task.VERIFY) return host def test_requested_reverify(self): host = self._create_reverify_request() self._run_dispatcher() self._check_host_status(host, HostStatus.VERIFYING) self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() self._check_host_status(host, HostStatus.READY) def test_requested_reverify_failure(self): host = self._create_reverify_request() self._run_dispatcher() self.mock_drone_manager.finish_process(_PidfileType.VERIFY, exit_status=256) self._run_dispatcher() # repair self._check_host_status(host, HostStatus.REPAIRING) self.mock_drone_manager.finish_process(_PidfileType.REPAIR) self._run_dispatcher() self._check_host_status(host, HostStatus.READY) def _setup_for_do_not_verify(self): self._initialize_test() job, queue_entry = self._make_job_and_queue_entry() queue_entry.host.protection = host_protections.Protection.DO_NOT_VERIFY queue_entry.host.save() return queue_entry def test_do_not_verify_job(self): queue_entry = self._setup_for_do_not_verify() self._run_dispatcher() # runs job directly self._finish_job(queue_entry) def test_do_not_verify_job_with_cleanup(self): queue_entry = self._setup_for_do_not_verify() queue_entry.job.reboot_before = model_attributes.RebootBefore.ALWAYS queue_entry.job.save() self._run_dispatcher() # cleanup self.mock_drone_manager.finish_process(_PidfileType.CLEANUP) self._run_dispatcher() # job self._finish_job(queue_entry) def test_do_not_verify_pre_job_cleanup_failure(self): queue_entry = self._setup_for_do_not_verify() queue_entry.job.reboot_before = model_attributes.RebootBefore.ALWAYS queue_entry.job.save() self._run_dispatcher() # cleanup self.mock_drone_manager.finish_process(_PidfileType.CLEANUP, exit_status=256) self._run_dispatcher() # failure ignored; job runs self._finish_job(queue_entry) def test_do_not_verify_post_job_cleanup_failure(self): queue_entry = self._setup_for_do_not_verify() queue_entry.job.reboot_after = model_attributes.RebootAfter.ALWAYS queue_entry.job.save() self._run_post_job_cleanup_failure_up_to_repair(queue_entry, include_verify=False) # failure ignored, host still set to Ready self._check_statuses(queue_entry, HqeStatus.COMPLETED, HostStatus.READY) self._run_dispatcher() # nothing else runs self._assert_nothing_is_running() def test_do_not_verify_requested_reverify_failure(self): host = self._create_reverify_request() host.protection = host_protections.Protection.DO_NOT_VERIFY host.save() self._run_dispatcher() self.mock_drone_manager.finish_process(_PidfileType.VERIFY, exit_status=256) self._run_dispatcher() self._check_host_status(host, HostStatus.READY) # ignore failure self._assert_nothing_is_running() def test_job_abort_in_verify(self): self._initialize_test() job = self._create_job(hosts=[1]) queue_entries = list(job.hostqueueentry_set.all()) self._run_dispatcher() # launches verify self._check_statuses(queue_entries[0], HqeStatus.VERIFYING) job.hostqueueentry_set.update(aborted=True) self._run_dispatcher() # kills verify, launches cleanup self.assert_(self.mock_drone_manager.was_last_process_killed( _PidfileType.VERIFY, set([signal.SIGKILL]))) self.mock_drone_manager.finish_process(_PidfileType.CLEANUP) self._run_dispatcher() def test_job_abort(self): self._initialize_test() job = self._create_job(hosts=[1]) job.run_reset = False job.save() queue_entries = list(job.hostqueueentry_set.all()) self._run_dispatcher() # launches job self._check_statuses(queue_entries[0], HqeStatus.RUNNING) job.hostqueueentry_set.update(aborted=True) self._run_dispatcher() # kills job, launches gathering self._check_statuses(queue_entries[0], HqeStatus.GATHERING) self.mock_drone_manager.finish_process(_PidfileType.GATHER) self._run_dispatcher() # launches parsing + cleanup queue_entry = job.hostqueueentry_set.all()[0] self._finish_parsing() # The abort will cause gathering to launch a cleanup. self.mock_drone_manager.finish_process(_PidfileType.CLEANUP) self._run_dispatcher() self.mock_drone_manager.finish_process(_PidfileType.RESET) self._run_dispatcher() def test_job_abort_queued_synchronous(self): self._initialize_test() job = self._create_job(hosts=[1,2]) job.synch_count = 2 job.save() job.hostqueueentry_set.update(aborted=True) self._run_dispatcher() for host_queue_entry in job.hostqueueentry_set.all(): self.assertEqual(host_queue_entry.status, HqeStatus.ABORTED) def test_no_pidfile_leaking(self): self._initialize_test() self.test_simple_job() self.mock_drone_manager.refresh() self.assertEquals(self.mock_drone_manager._pidfiles, {}) self.test_job_abort_in_verify() self.mock_drone_manager.refresh() self.assertEquals(self.mock_drone_manager._pidfiles, {}) self.test_job_abort() self.mock_drone_manager.refresh() self.assertEquals(self.mock_drone_manager._pidfiles, {}) def _make_job_and_queue_entry(self): job = self._create_job(hosts=[1]) queue_entry = job.hostqueueentry_set.all()[0] return job, queue_entry def test_recover_running_no_process(self): # recovery should re-execute a Running HQE if no process is found _, queue_entry = self._make_job_and_queue_entry() queue_entry.status = HqeStatus.RUNNING queue_entry.execution_subdir = '1-myuser/host1' queue_entry.save() queue_entry.host.status = HostStatus.RUNNING queue_entry.host.save() self._initialize_test() self._run_dispatcher() self._finish_job(queue_entry) def test_recover_verifying_hqe_no_special_task(self): # recovery should move a Resetting HQE with no corresponding # Verify or Reset SpecialTask back to Queued. _, queue_entry = self._make_job_and_queue_entry() queue_entry.status = HqeStatus.RESETTING queue_entry.save() # make some dummy SpecialTasks that shouldn't count models.SpecialTask.objects.create( host=queue_entry.host, task=models.SpecialTask.Task.RESET, requested_by=models.User.current_user()) models.SpecialTask.objects.create( host=queue_entry.host, task=models.SpecialTask.Task.CLEANUP, queue_entry=queue_entry, is_complete=True, requested_by=models.User.current_user()) self._initialize_test() self._check_statuses(queue_entry, HqeStatus.QUEUED) def _test_recover_verifying_hqe_helper(self, task, pidfile_type): _, queue_entry = self._make_job_and_queue_entry() queue_entry.status = HqeStatus.VERIFYING queue_entry.save() special_task = models.SpecialTask.objects.create( host=queue_entry.host, task=task, queue_entry=queue_entry) self._initialize_test() self._run_dispatcher() self.mock_drone_manager.finish_process(pidfile_type) self._run_dispatcher() # don't bother checking the rest of the job execution, as long as the # SpecialTask ran def test_recover_verifying_hqe_with_cleanup(self): # recover an HQE that was in pre-job cleanup self._test_recover_verifying_hqe_helper(models.SpecialTask.Task.CLEANUP, _PidfileType.CLEANUP) def test_recover_verifying_hqe_with_verify(self): # recover an HQE that was in pre-job verify self._test_recover_verifying_hqe_helper(models.SpecialTask.Task.VERIFY, _PidfileType.VERIFY) def test_recover_parsing(self): self._initialize_test() job, queue_entry = self._make_job_and_queue_entry() job.run_verify = False job.run_reset = False job.reboot_after = model_attributes.RebootAfter.NEVER job.save() self._run_dispatcher() # launches job self.mock_drone_manager.finish_process(_PidfileType.JOB) self._run_dispatcher() # launches parsing # now "restart" the scheduler self._create_dispatcher() self._initialize_test() self._run_dispatcher() self.mock_drone_manager.finish_process(_PidfileType.PARSE) self._run_dispatcher() def test_recover_parsing__no_process_already_aborted(self): _, queue_entry = self._make_job_and_queue_entry() queue_entry.execution_subdir = 'host1' queue_entry.status = HqeStatus.PARSING queue_entry.aborted = True queue_entry.save() self._initialize_test() self._run_dispatcher() def test_job_scheduled_just_after_abort(self): # test a pretty obscure corner case where a job is aborted while queued, # another job is ready to run, and throttling is active. the post-abort # cleanup must not be pre-empted by the second job. # This test kind of doesn't make sense anymore after verify+cleanup # were merged into reset. It should maybe just be removed. job1, queue_entry1 = self._make_job_and_queue_entry() queue_entry1.save() job2, queue_entry2 = self._make_job_and_queue_entry() job2.reboot_before = model_attributes.RebootBefore.IF_DIRTY job2.save() self.mock_drone_manager.process_capacity = 0 self._run_dispatcher() # schedule job1, but won't start verify job1.hostqueueentry_set.update(aborted=True) self.mock_drone_manager.process_capacity = 100 self._run_dispatcher() # reset must run here, not verify for job2 self._check_statuses(queue_entry1, HqeStatus.ABORTED, HostStatus.RESETTING) self.mock_drone_manager.finish_process(_PidfileType.RESET) self._run_dispatcher() # now verify starts for job2 self._check_statuses(queue_entry2, HqeStatus.RUNNING, HostStatus.RUNNING) def test_reverify_interrupting_pre_job(self): # ensure things behave sanely if a reverify is scheduled in the middle # of pre-job actions _, queue_entry = self._make_job_and_queue_entry() self._run_dispatcher() # pre-job verify self._create_reverify_request() self.mock_drone_manager.finish_process(_PidfileType.VERIFY, exit_status=256) self._run_dispatcher() # repair self.mock_drone_manager.finish_process(_PidfileType.REPAIR) self._run_dispatcher() # reverify runs now self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() # pre-job verify self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() # and job runs... self._check_statuses(queue_entry, HqeStatus.RUNNING, HostStatus.RUNNING) self._finish_job(queue_entry) # reverify has been deleted self._check_statuses(queue_entry, HqeStatus.COMPLETED, HostStatus.READY) self._assert_nothing_is_running() def test_reverify_while_job_running(self): # once a job is running, a reverify must not be allowed to preempt # Gathering _, queue_entry = self._make_job_and_queue_entry() self._run_pre_job_verify(queue_entry) self._run_dispatcher() # job runs self._create_reverify_request() # make job end with a signal, so gathering will run self.mock_drone_manager.finish_process(_PidfileType.JOB, exit_status=271) self._run_dispatcher() # gathering must start self.mock_drone_manager.finish_process(_PidfileType.GATHER) self._run_dispatcher() # parsing and cleanup self._finish_parsing() self._run_dispatcher() # now reverify runs self._check_statuses(queue_entry, HqeStatus.FAILED, HostStatus.VERIFYING) self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() self._check_host_status(queue_entry.host, HostStatus.READY) def test_reverify_while_host_pending(self): # ensure that if a reverify is scheduled while a host is in Pending, it # won't run until the host is actually free job = self._create_job(hosts=[1,2]) queue_entry = job.hostqueueentry_set.get(host__hostname='host1') job.synch_count = 2 job.save() host2 = self.hosts[1] host2.locked = True host2.save() self._run_dispatcher() # verify host1 self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() # host1 Pending self._check_statuses(queue_entry, HqeStatus.PENDING, HostStatus.PENDING) self._create_reverify_request() self._run_dispatcher() # nothing should happen here self._check_statuses(queue_entry, HqeStatus.PENDING, HostStatus.PENDING) # now let the job run host2.locked = False host2.save() self._run_dispatcher() # verify host2 self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() # run job self._finish_job(queue_entry) # the reverify should now be running self._check_statuses(queue_entry, HqeStatus.COMPLETED, HostStatus.VERIFYING) self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() self._check_host_status(queue_entry.host, HostStatus.READY) def test_throttling(self): job = self._create_job(hosts=[1,2,3]) job.synch_count = 3 job.save() queue_entries = list(job.hostqueueentry_set.all()) def _check_hqe_statuses(*statuses): for queue_entry, status in zip(queue_entries, statuses): self._check_statuses(queue_entry, status) self.mock_drone_manager.process_capacity = 2 self._run_dispatcher() # verify runs on 1 and 2 queue_entries = list(job.hostqueueentry_set.all()) _check_hqe_statuses(HqeStatus.QUEUED, HqeStatus.VERIFYING, HqeStatus.VERIFYING) self.assertEquals(len(self.mock_drone_manager.running_pidfile_ids()), 2) self.mock_drone_manager.finish_specific_process( 'hosts/host3/1-verify', drone_manager.AUTOSERV_PID_FILE) self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() # verify runs on 3 _check_hqe_statuses(HqeStatus.VERIFYING, HqeStatus.PENDING, HqeStatus.PENDING) self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() # job won't run due to throttling _check_hqe_statuses(HqeStatus.STARTING, HqeStatus.STARTING, HqeStatus.STARTING) self._assert_nothing_is_running() self.mock_drone_manager.process_capacity = 3 self._run_dispatcher() # now job runs _check_hqe_statuses(HqeStatus.RUNNING, HqeStatus.RUNNING, HqeStatus.RUNNING) self.mock_drone_manager.process_capacity = 2 self.mock_drone_manager.finish_process(_PidfileType.JOB, exit_status=271) self._run_dispatcher() # gathering won't run due to throttling _check_hqe_statuses(HqeStatus.GATHERING, HqeStatus.GATHERING, HqeStatus.GATHERING) self._assert_nothing_is_running() self.mock_drone_manager.process_capacity = 3 self._run_dispatcher() # now gathering runs self.mock_drone_manager.process_capacity = 0 self.mock_drone_manager.finish_process(_PidfileType.GATHER) self._run_dispatcher() # parsing runs despite throttling _check_hqe_statuses(HqeStatus.PARSING, HqeStatus.PARSING, HqeStatus.PARSING) def test_abort_starting_while_throttling(self): self._initialize_test() job = self._create_job(hosts=[1,2], synchronous=True) queue_entry = job.hostqueueentry_set.all()[0] job.run_verify = False job.run_reset = False job.reboot_after = model_attributes.RebootAfter.NEVER job.save() self.mock_drone_manager.process_capacity = 0 self._run_dispatcher() # go to starting, but don't start job self._check_statuses(queue_entry, HqeStatus.STARTING, HostStatus.PENDING) job.hostqueueentry_set.update(aborted=True) self._run_dispatcher() self._check_statuses(queue_entry, HqeStatus.GATHERING, HostStatus.RUNNING) self.mock_drone_manager.process_capacity = 5 self._run_dispatcher() self._check_statuses(queue_entry, HqeStatus.ABORTED, HostStatus.CLEANING) def test_simple_metahost_assignment(self): job = self._create_job(metahosts=[1]) self._run_dispatcher() entry = job.hostqueueentry_set.all()[0] self.assertEquals(entry.host.hostname, 'host1') self._check_statuses(entry, HqeStatus.VERIFYING, HostStatus.VERIFYING) self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() self._check_statuses(entry, HqeStatus.RUNNING, HostStatus.RUNNING) # rest of job proceeds normally def test_metahost_fail_verify(self): self.hosts[1].labels.add(self.labels[0]) # put label1 also on host2 job = self._create_job(metahosts=[1]) self._run_dispatcher() # assigned to host1 self.mock_drone_manager.finish_process(_PidfileType.VERIFY, exit_status=256) self._run_dispatcher() # host1 failed, gets reassigned to host2 entry = job.hostqueueentry_set.all()[0] self.assertEquals(entry.host.hostname, 'host2') self._check_statuses(entry, HqeStatus.VERIFYING, HostStatus.VERIFYING) self._check_host_status(self.hosts[0], HostStatus.REPAIRING) self.mock_drone_manager.finish_process(_PidfileType.VERIFY) self._run_dispatcher() self._check_statuses(entry, HqeStatus.RUNNING, HostStatus.RUNNING) def test_hostless_job(self): job = self._create_job(hostless=True) entry = job.hostqueueentry_set.all()[0] self._run_dispatcher() self._check_entry_status(entry, HqeStatus.RUNNING) self.mock_drone_manager.finish_process(_PidfileType.JOB) self._run_dispatcher() self._check_entry_status(entry, HqeStatus.PARSING) self.mock_drone_manager.finish_process(_PidfileType.PARSE) self._run_dispatcher() self._check_entry_status(entry, HqeStatus.COMPLETED) def test_pre_job_keyvals(self): job = self._create_job(hosts=[1]) job.run_verify = False job.run_reset = False job.reboot_before = model_attributes.RebootBefore.NEVER job.save() models.JobKeyval.objects.create(job=job, key='mykey', value='myvalue') self._run_dispatcher() self._finish_job(job.hostqueueentry_set.all()[0]) attached_files = self.mock_drone_manager.attached_files( '1-autotest_system/host1') job_keyval_path = '1-autotest_system/host1/keyval' self.assert_(job_keyval_path in attached_files, attached_files) keyval_contents = attached_files[job_keyval_path] keyval_dict = dict(line.strip().split('=', 1) for line in keyval_contents.splitlines()) self.assert_('job_queued' in keyval_dict, keyval_dict) self.assertEquals(keyval_dict['mykey'], 'myvalue') # This tests the scheduler functions with archiving step disabled class SchedulerFunctionalTestNoArchiving(SchedulerFunctionalTest): def _set_global_config_values(self): super(SchedulerFunctionalTestNoArchiving, self )._set_global_config_values() self.mock_config.set_config_value('SCHEDULER', 'enable_archiving', False) def _finish_parsing(self): self.mock_drone_manager.finish_process(_PidfileType.PARSE) self._run_dispatcher() def _run_post_job_cleanup_failure_up_to_repair(self, queue_entry, include_verify=True): if include_verify: self._run_pre_job_verify(queue_entry) self._run_dispatcher() # job self.mock_drone_manager.finish_process(_PidfileType.JOB) self._run_dispatcher() # parsing + cleanup self.mock_drone_manager.finish_process(_PidfileType.PARSE) self.mock_drone_manager.finish_process(_PidfileType.CLEANUP, exit_status=256) self._run_dispatcher() # repair, HQE unaffected return queue_entry def test_hostless_job(self): job = self._create_job(hostless=True) entry = job.hostqueueentry_set.all()[0] self._run_dispatcher() self._check_entry_status(entry, HqeStatus.RUNNING) self.mock_drone_manager.finish_process(_PidfileType.JOB) self._run_dispatcher() self._check_entry_status(entry, HqeStatus.PARSING) self.mock_drone_manager.finish_process(_PidfileType.PARSE) self._run_dispatcher() self._check_entry_status(entry, HqeStatus.COMPLETED) def test_synchronous_with_reset(self): # For crbug/621257. job = self._create_job(hosts=[1, 2]) job.synch_count = 2 job.reboot_before = model_attributes.RebootBefore.ALWAYS job.save() hqe1 = job.hostqueueentry_set.get(host__hostname='host1') hqe2 = job.hostqueueentry_set.get(host__hostname='host2') self._run_dispatcher() self._check_statuses(hqe1, HqeStatus.RESETTING, HostStatus.RESETTING) self._check_statuses(hqe2, HqeStatus.RESETTING, HostStatus.RESETTING) self.mock_drone_manager.finish_active_process_on_host(1) self._run_dispatcher() self._check_statuses(hqe1, HqeStatus.PENDING, HostStatus.PENDING) self._check_statuses(hqe2, HqeStatus.RESETTING, HostStatus.RESETTING) self.mock_drone_manager.finish_active_process_on_host(2) self._run_dispatcher() self._check_statuses(hqe1, HqeStatus.RUNNING, HostStatus.RUNNING) self._check_statuses(hqe2, HqeStatus.RUNNING, HostStatus.RUNNING) if __name__ == '__main__': unittest.main()