class Spawner(object): def __init__(self): self.num_workers = int(Config.get('workers', 'init_count')) self.secret = os.environ.get('_abaco_secret') self.queue = os.environ.get('queue', 'default') self.cmd_ch = CommandChannel(name=self.queue) self.tot_workers = 0 try: self.host_id = Config.get('spawner', 'host_id') except Exception as e: logger.critical("Spawner not configured with a host_id! Aborting! Exception: {}".format(e)) raise e def run(self): while True: # check resource threshold before subscribing while True: if self.overloaded(): logger.critical("METRICS - SPAWNER FOR HOST {} OVERLOADED!!!".format(self.host_id)) # self.update_status to OVERLOADED time.sleep(5) else: break cmd, msg_obj = self.cmd_ch.get_one() # directly ack the messages from the command channel; problems generated from starting workers are # handled downstream; e.g., by setting the actor in an ERROR state; command messages should not be re-queued msg_obj.ack() try: self.process(cmd) except Exception as e: logger.error("spawner got an exception trying to process cmd: {}. " "Exception type: {}. Exception: {}".format(cmd, type(e), e)) def get_tot_workers(self): logger.debug("top of get_tot_workers") self.tot_workers = 0 logger.debug('spawner host_id: {}'.format(self.host_id)) for k,v in workers_store.items(): for wid, worker in v.items(): if worker.get('host_id') == self.host_id: self.tot_workers += 1 logger.debug("returning total workers: {}".format(self.tot_workers)) return self.tot_workers def overloaded(self): logger.debug("top of overloaded") self.get_tot_workers() logger.info("total workers for this host: {}".format(self.tot_workers)) if self.tot_workers >= MAX_WORKERS: return True def stop_workers(self, actor_id, worker_ids): """Stop existing workers; used when updating an actor's image.""" logger.debug("Top of stop_workers() for actor: {}.".format(actor_id)) try: workers_dict = workers_store[actor_id] except KeyError: logger.debug("workers_store had no workers for actor: {}".format(actor_id)) workers_dict = {} # if there are existing workers, we need to close the actor message channel and # gracefully shutdown the existing worker processes. if len(workers_dict.items()) > 0: logger.info("Found {} workers to stop.".format(len(workers_dict.items()))) # first, close the actor msg channel to prevent any new messages from being pulled # by the old workers. actor_ch = ActorMsgChannel(actor_id) actor_ch.close() logger.info("Actor channel closed for actor: {}".format(actor_id)) # now, send messages to workers for a graceful shutdown: for _, worker in workers_dict.items(): # don't stop the new workers: if worker['id'] not in worker_ids: ch = WorkerChannel(worker_id=worker['id']) # since this is an update, there are new workers being started, so # don't delete the actor msg channel: ch.put('stop-no-delete') logger.info("Sent 'stop-no-delete' message to worker_id: {}".format(worker['id'])) ch.close() else: logger.debug("skipping worker {} as it it not in worker_ids.".format(worker)) else: logger.info("No workers to stop.") def process(self, cmd): """Main spawner method for processing a command from the CommandChannel.""" logger.info("top of process; cmd: {}".format(cmd)) actor_id = cmd['actor_id'] try: actor = Actor.from_db(actors_store[actor_id]) except Exception as e: msg = f"Exception in spawner trying to retrieve actor object from store. Aborting. Exception: {e}" logger.error(msg) return worker_id = cmd['worker_id'] image = cmd['image'] tenant = cmd['tenant'] stop_existing = cmd.get('stop_existing', True) num_workers = 1 logger.debug("spawner command params: actor_id: {} worker_id: {} image: {} tenant: {}" "stop_existing: {} num_workers: {}".format(actor_id, worker_id, image, tenant, stop_existing, num_workers)) # if the worker was sent a delete request before spawner received this message to create the worker, # the status will be SHUTDOWN_REQUESTED, not REQUESTED. in that case, we simply abort and remove the # worker from the collection. try: logger.debug("spawner checking worker's status for SHUTDOWN_REQUESTED") worker = Worker.get_worker(actor_id, worker_id) logger.debug(f"spawner got worker; worker: {worker}") except Exception as e: logger.error(f"spawner got exception trying to retrieve worker. " f"actor_id: {actor_id}; worker_id: {worker_id}; e: {e}") return status = worker.get('status') if not status == REQUESTED: logger.debug(f"worker was NOT in REQUESTED status. status: {status}") if status == SHUTDOWN_REQUESTED or status == SHUTTING_DOWN or status == ERROR: logger.debug(f"worker status was {status}; spawner deleting worker and returning..") try: Worker.delete_worker(actor_id, worker_id) logger.debug("spawner deleted worker because it was SHUTDOWN_REQUESTED.") return except Exception as e: logger.error(f"spawner got exception trying to delete a worker in SHUTDOWN_REQUESTED status." f"actor_id: {actor_id}; worker_id: {worker_id}; e: {e}") return else: logger.error(f"spawner found worker in unexpected status: {status}. Not processing command and returning.") return # worker status was REQUESTED; moving on to SPAWNER_SETUP ---- Worker.update_worker_status(actor_id, worker_id, SPAWNER_SETUP) logger.debug("spawner has updated worker status to SPAWNER_SETUP; worker_id: {}".format(worker_id)) client_id = None client_secret = None client_access_token = None client_refresh_token = None api_server = None client_secret = None # ---- Oauth client generation for the worker ------- # check if tenant and instance configured for client generation - try: generate_clients = Config.get('workers', f'{tenant}_generate_clients').lower() except: logger.debug(f"Did not find a {tenant}_generate_clients config. Looking for a global config.") generate_clients = Config.get('workers', 'generate_clients').lower() logger.debug(f"final generate_clients: {generate_clients}") if generate_clients == "true": logger.debug("client generation was configured to be available; now checking the actor's token attr.") # updated 1.3.0-- check whether the actor requires a token: if actor.token: logger.debug("spawner starting client generation") client_id, \ client_access_token, \ client_refresh_token, \ api_server, \ client_secret = self.client_generation(actor_id, worker_id, tenant) else: logger.debug("actor's token attribute was False. Not generating client.") ch = SpawnerWorkerChannel(worker_id=worker_id) logger.debug("spawner attempting to start worker; worker_id: {}".format(worker_id)) try: worker = self.start_worker( image, tenant, actor_id, worker_id, client_id, client_access_token, client_refresh_token, ch, api_server, client_secret ) except Exception as e: msg = "Spawner got an exception from call to start_worker. Exception:{}".format(e) logger.error(msg) self.error_out_actor(actor_id, worker_id, msg) if client_id: self.delete_client(tenant, actor_id, worker_id, client_id, client_secret) return logger.debug("Returned from start_worker; Created new worker: {}".format(worker)) ch.close() logger.debug("Client channel closed") if stop_existing: logger.info("Stopping existing workers: {}".format(worker_id)) # TODO - update status to stop_requested self.stop_workers(actor_id, [worker_id]) def client_generation(self, actor_id, worker_id, tenant): client_ch = ClientsChannel() try: client_msg = client_ch.request_client( tenant=tenant, actor_id=actor_id, worker_id=worker_id, secret=self.secret ) except Exception as e: logger.error("Got a ChannelTimeoutException trying to generate a client for " "actor_id: {}; worker_id: {}; exception: {}".format(actor_id, worker_id, e)) # put worker in an error state and return self.error_out_actor(actor_id, worker_id, "Abaco was unable to generate an OAuth client for a new " "worker for this actor. System administrators have been notified.") client_ch.close() Worker.update_worker_status(actor_id, worker_id, ERROR) logger.critical("Client generation FAILED.") raise e client_ch.close() if client_msg.get('status') == 'error': logger.error("Error generating client: {}".format(client_msg.get('message'))) self.error_out_actor(actor_id, worker_id, "Abaco was unable to generate an OAuth client for a new " "worker for this actor. System administrators have been notified.") Worker.update_worker_status(actor_id, worker_id, ERROR) raise SpawnerException("Error generating client") #TODO - clean up error message # else, client was generated successfully: else: logger.info("Got a client: {}, {}, {}".format(client_msg['client_id'], client_msg['access_token'], client_msg['refresh_token'])) return client_msg['client_id'], \ client_msg['access_token'], \ client_msg['refresh_token'], \ client_msg['api_server'], \ client_msg['client_secret'] def delete_client(self, tenant, actor_id, worker_id, client_id, secret): clients_ch = ClientsChannel() msg = clients_ch.request_delete_client(tenant=tenant, actor_id=actor_id, worker_id=worker_id, client_id=client_id, secret=secret) if msg['status'] == 'ok': logger.info("Client delete request completed successfully for " "worker_id: {}, client_id: {}.".format(worker_id, client_id)) else: logger.error("Error deleting client for " "worker_id: {}, client_id: {}. Message: {}".format(worker_id, msg['message'], client_id, msg)) clients_ch.close() def start_worker(self, image, tenant, actor_id, worker_id, client_id, client_access_token, client_refresh_token, ch, api_server, client_secret): # start an actor executor container and wait for a confirmation that image was pulled. attempts = 0 # worker = get_worker(worker_id) # worker['status'] = PULLING_IMAGE Worker.update_worker_status(actor_id, worker_id, PULLING_IMAGE) try: logger.debug("Worker pulling image {}...".format(image)) pull_image(image) except DockerError as e: # return a message to the spawner that there was an error pulling image and abort # this is not necessarily an error state: the user simply could have provided an # image name that does not exist in the registry. This is the first time we would # find that out. logger.info("worker got a DockerError trying to pull image. Error: {}.".format(e)) raise e logger.info("Image {} pulled successfully.".format(image)) # Done pulling image # Run Worker Container while True: try: Worker.update_worker_status(actor_id, worker_id, CREATING_CONTAINER) logger.debug('spawner creating worker container') worker_dict = run_worker( image, actor_id, worker_id, client_id, client_access_token, client_refresh_token, tenant, api_server, client_secret ) logger.debug(f'finished run worker; worker dict: {worker_dict}') except DockerError as e: logger.error("Spawner got a docker exception from run_worker; Exception: {}".format(e)) if 'read timeout' in e.message: logger.info("Exception was a read timeout; trying run_worker again..") time.sleep(5) attempts = attempts + 1 if attempts > 20: msg = "Spawner continued to get DockerError for 20 attempts. Exception: {}".format(e) logger.critical(msg) # todo - should we be calling kill_worker here? (it is called in the exception block of the else below) raise SpawnerException(msg) continue else: logger.info("Exception was NOT a read timeout; quiting on this worker.") # delete this worker from the workers store: try: self.kill_worker(actor_id, worker_id) except WorkerException as e: logger.info("Got WorkerException from delete_worker(). " "worker_id: {}" "Exception: {}".format(worker_id, e)) raise SpawnerException(message="Unable to start worker; error: {}".format(e)) break logger.debug('finished loop') worker_dict['ch_name'] = WorkerChannel.get_name(worker_id) # if the actor is not already in READY status, set actor status to READY before worker status has been # set to READY. # it is possible the actor status is already READY because this request is the autoscaler starting a new worker # for an existing actor. actor = Actor.from_db(actors_store[actor_id]) if not actor.status == READY: try: Actor.set_status(actor_id, READY, status_message=" ") except KeyError: # it is possible the actor was already deleted during worker start up; if # so, the worker should have a stop message waiting for it. starting subscribe # as usual should allow this process to work as expected. pass # finalize worker with READY status worker = Worker(tenant=tenant, **worker_dict) logger.info("calling add_worker for worker: {}.".format(worker)) Worker.add_worker(actor_id, worker) ch.put('READY') # step 4 logger.info('sent message through channel') def error_out_actor(self, actor_id, worker_id, message): """In case of an error, put the actor in error state and kill all workers""" logger.debug("top of error_out_actor for worker: {}_{}".format(actor_id, worker_id)) Actor.set_status(actor_id, ERROR, status_message=message) # first we try to stop workers using the "graceful" approach - try: self.stop_workers(actor_id, worker_ids=[]) logger.info("Spawner just stopped worker {}_{} in error_out_actor".format(actor_id, worker_id)) return except Exception as e: logger.error("spawner got exception trying to run stop_workers. Exception: {}".format(e)) try: self.kill_worker(actor_id, worker_id) logger.info("Spawner just killed worker {}_{} in error_out_actor".format(actor_id, worker_id)) except DockerError as e: logger.info("Received DockerError trying to kill worker: {}. Exception: {}".format(worker_id, e)) logger.info("Spawner will continue on since this is exception processing.") def kill_worker(self, actor_id, worker_id): try: Worker.delete_worker(actor_id, worker_id) except WorkerException as e: logger.info("Got WorkerException from delete_worker(). " "worker_id: {}" "Exception: {}".format(worker_id, e)) except Exception as e: logger.error("Got an unexpected exception from delete_worker(). " "worker_id: {}" "Exception: {}".format(worker_id, e))
class Spawner(object): def __init__(self): self.num_workers = int(Config.get('workers', 'init_count')) self.secret = os.environ.get('_abaco_secret') self.cmd_ch = CommandChannel() self.tot_workers = 0 try: self.host_id = Config.get('spawner', 'host_id') except Exception as e: logger.critical("Spawner not configured with a host_id! Aborting! Exception: {}".format(e)) raise e def run(self): while True: # check resource threshold before subscribing while True: if self.overloaded(): logger.critical("METRICS - SPAWNER FOR HOST {} OVERLOADED!!!".format(self.host_id)) # self.update_status to OVERLOADED time.sleep(5) else: break cmd, msg_obj = self.cmd_ch.get_one() # directly ack the messages from the command channel; problems generated from starting workers are # handled downstream; e.g., by setting the actor in an ERROR state; command messages should not be re-queued msg_obj.ack() self.process(cmd) def get_tot_workers(self): logger.debug("top of get_tot_workers") self.tot_workers = 0 logger.debug('spawner host_id: {}'.format(self.host_id)) for k,v in workers_store.items(): for wid, worker in v.items(): if worker.get('host_id') == self.host_id: self.tot_workers += 1 logger.debug("returning total workers: {}".format(self.tot_workers)) return self.tot_workers def overloaded(self): logger.debug("top of overloaded") self.get_tot_workers() logger.info("total workers for this host: {}".format(self.tot_workers)) if self.tot_workers >= MAX_WORKERS: return True def stop_workers(self, actor_id, worker_ids): """Stop existing workers; used when updating an actor's image.""" logger.debug("Top of stop_workers() for actor: {}.".format(actor_id)) try: workers_dict = workers_store[actor_id] except KeyError: logger.debug("workers_store had no workers for actor: {}".format(actor_id)) workers_dict = {} # if there are existing workers, we need to close the actor message channel and # gracefully shutdown the existing worker processes. if len(workers_dict.items()) > 0: logger.info("Found {} workers to stop.".format(len(workers_dict.items()))) # first, close the actor msg channel to prevent any new messages from being pulled # by the old workers. actor_ch = ActorMsgChannel(actor_id) actor_ch.close() logger.info("Actor channel closed for actor: {}".format(actor_id)) # now, send messages to workers for a graceful shutdown: for _, worker in workers_dict.items(): # don't stop the new workers: if worker['id'] not in worker_ids: ch = WorkerChannel(worker_id=worker['id']) # since this is an update, there are new workers being started, so # don't delete the actor msg channel: ch.put('stop-no-delete') logger.info("Sent 'stop-no-delete' message to worker_id: {}".format(worker['id'])) ch.close() else: logger.info("No workers to stop.") def process(self, cmd): """Main spawner method for processing a command from the CommandChannel.""" logger.info("Spawner processing new command:{}".format(cmd)) actor_id = cmd['actor_id'] worker_ids = cmd['worker_ids'] image = cmd['image'] tenant = cmd['tenant'] stop_existing = cmd.get('stop_existing', True) num_workers = cmd.get('num', self.num_workers) logger.info("command params: actor_id: {} worker_ids: {} image: {} stop_existing: {} mum_workers: {}".format( actor_id, worker_ids, image, tenant, stop_existing, num_workers)) try: new_channels, anon_channels, new_workers = self.start_workers(actor_id, worker_ids, image, tenant, num_workers) except SpawnerException as e: # for now, start_workers will do clean up for a SpawnerException, so we just need # to return back to the run loop. logger.info("Spawner returning to main run loop.") return logger.info("Created new workers: {}".format(new_workers)) # stop any existing workers: if stop_existing: logger.info("Stopping existing workers: {}".format(worker_ids)) self.stop_workers(actor_id, worker_ids) # add workers to store first so that the records will be there when the workers go # to update their status if not stop_existing: # if we're not stopping the existing workers, we need to add each worker to the # actor's collection. for _, worker in new_workers.items(): logger.info("calling add_worker for worker: {}.".format(worker)) Worker.add_worker(actor_id, worker) else: # since we're stopping the existing workers, the actor's collection should just # be equal to the new_workers. workers_store[actor_id] = new_workers logger.info("workers_store set to new_workers: {}.".format(new_workers)) # Tell new worker to subscribe to the actor channel. # If abaco is configured to generate clients for the workers, generate them now # and send new workers their clients. generate_clients = Config.get('workers', 'generate_clients').lower() logger.info("Sending messages to new workers over anonymous channels to subscribe to inbox.") for idx, channel in enumerate(anon_channels): if generate_clients == 'true': worker_id = new_workers[list(new_workers)[idx]]['id'] logger.info("Getting client for worker number {}, id: {}".format(idx, worker_id)) client_ch = ClientsChannel() try: client_msg = client_ch.request_client(tenant=tenant, actor_id=actor_id, # new_workers is a dictionary of dictionaries; list(d) creates a # list of keys for a dictionary d. hence, the idx^th entry # of list(ner_workers) should be the key. worker_id=worker_id, secret=self.secret) except ChannelTimeoutException as e: logger.error("Got a ChannelTimeoutException trying to generate a client for " "actor_id: {}; worker_id: {}; exception: {}".format(actor_id, worker_id, e)) # put actor in an error state and return self.error_out_actor(actor_id, worker_id, "Abaco was unable to generate an OAuth client for a new " "worker for this actor. System administrators have been notified.") client_ch.close() return client_ch.close() # we need to ignore errors when generating clients because it's possible it is not set up for a specific # tenant. we log it instead. if client_msg.get('status') == 'error': logger.error("Error generating client: {}".format(client_msg.get('message'))) channel.put({'status': 'ok', 'actor_id': actor_id, 'tenant': tenant, 'client': 'no'}) logger.debug("Sent OK message over anonymous worker channel.") # else, client was generated successfully: else: logger.info("Got a client: {}, {}, {}".format(client_msg['client_id'], client_msg['access_token'], client_msg['refresh_token'])) channel.put({'status': 'ok', 'actor_id': actor_id, 'tenant': tenant, 'client': 'yes', 'client_id': client_msg['client_id'], 'client_secret': client_msg['client_secret'], 'access_token': client_msg['access_token'], 'refresh_token': client_msg['refresh_token'], 'api_server': client_msg['api_server'], }) logger.debug("Sent OK message AND client over anonymous worker channel.") else: logger.info("Not generating clients. Config value was: {}".format(generate_clients)) channel.put({'status': 'ok', 'actor_id': actor_id, 'tenant': tenant, 'client': 'no'}) logger.debug("Sent OK message over anonymous worker channel.") # @TODO - # delete the anonymous channel from this thread but sleep first to avoid the race condition. time.sleep(1.5) channel.delete() # due to the race condition deleting channels (potentially before all workers have received all messages) # we put a sleep here. time.sleep(1) for ch in new_channels: try: # the new_channels are the spawnerworker channels so they can be deleted. ch.delete() except Exception as e: logger.error("Got exception trying to delete spawnerworker channel: {}".format(e)) logger.info("Done processing command.") def start_workers(self, actor_id, worker_ids, image, tenant, num_workers): logger.info("starting {} workers. actor_id: {} image: {}".format(str(self.num_workers), actor_id, image)) channels = [] anon_channels = [] workers = {} try: for i in range(num_workers): worker_id = worker_ids[i] logger.info("starting worker {} with id: {}".format(i, worker_id)) ch, anon_ch, worker = self.start_worker(image, tenant, actor_id, worker_id) logger.debug("channel for worker {} is: {}".format(str(i), ch.name)) channels.append(ch) anon_channels.append(anon_ch) workers[worker_id] = worker except SpawnerException as e: logger.info("Caught SpawnerException:{}".format(str(e))) # in case of an error, put the actor in error state and kill all workers self.error_out_actor(actor_id, worker_id, e.message) raise SpawnerException(message=e.message) return channels, anon_channels, workers def start_worker(self, image, tenant, actor_id, worker_id): ch = SpawnerWorkerChannel(worker_id=worker_id) # start an actor executor container and wait for a confirmation that image was pulled. attempts = 0 while True: try: worker_dict = run_worker(image, actor_id, worker_id) except DockerError as e: logger.error("Spawner got a docker exception from run_worker; Exception: {}".format(e)) if 'read timeout' in e.message: logger.info("Exception was a read timeout; trying run_worker again..") time.sleep(5) attempts = attempts + 1 if attempts > 20: msg = "Spawner continued to get DockerError for 20 attempts. Exception: {}".format(e) logger.critical(msg) raise SpawnerException(msg) continue else: logger.info("Exception was NOT a read timeout; quiting on this worker.") # delete this worker from the workers store: try: self.kill_worker(actor_id, worker_id) except WorkerException as e: logger.info("Got WorkerException from delete_worker(). " "worker_id: {}" "Exception: {}".format(worker_id, e)) raise SpawnerException(message="Unable to start worker; error: {}".format(e)) break worker_dict['ch_name'] = WorkerChannel.get_name(worker_id) worker = Worker(tenant=tenant, **worker_dict) logger.info("worker started successfully, waiting on ack that image was pulled...") result = ch.get() logger.debug("Got response back from worker. Response: {}".format(result)) if result.get('status') == 'error': # there was a problem pulling the image; put the actor in an error state: msg = "Got an error back from the worker. Message: {}",format(result) logger.info(msg) if 'msg' in result: raise SpawnerException(message=result['msg']) else: logger.error("Spawner received invalid message from worker. 'msg' field missing. Message: {}".format(result)) raise SpawnerException(message="Internal error starting worker process.") elif result['value']['status'] == 'ok': logger.debug("received ack from worker.") return ch, result['reply_to'], worker else: msg = "Got an error status from worker: {}. Raising an exception.".format(str(result)) logger.error("Spawner received an invalid message from worker. Message: ".format(result)) raise SpawnerException(msg) def error_out_actor(self, actor_id, worker_id, message): """In case of an error, put the actor in error state and kill all workers""" Actor.set_status(actor_id, ERROR, status_message=message) try: self.kill_worker(actor_id, worker_id) except DockerError as e: logger.info("Received DockerError trying to kill worker: {}. Exception: {}".format(worker_id, e)) logger.info("Spawner will continue on since this is exception processing.") def kill_worker(self, actor_id, worker_id): try: Worker.delete_worker(actor_id, worker_id) except WorkerException as e: logger.info("Got WorkerException from delete_worker(). " "worker_id: {}" "Exception: {}".format(worker_id, e)) except Exception as e: logger.error("Got an unexpected exception from delete_worker(). " "worker_id: {}" "Exception: {}".format(worker_id, e))