• 设为首页
  • 点击收藏
  • 手机版
    手机扫一扫访问
    迪恩网络手机版
  • 关注官方公众号
    微信扫一扫关注
    公众号

Python model.Worker类代码示例

原作者: [db:作者] 来自: [db:来源] 收藏 邀请

本文整理汇总了Python中pulp.server.db.model.Worker的典型用法代码示例。如果您正苦于以下问题:Python Worker类的具体用法?Python Worker怎么用?Python Worker使用的例子?那么恭喜您, 这里精选的类代码示例或许可以为您提供帮助。



在下文中一共展示了Worker类的13个代码示例,这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞,您的评价将有助于我们的系统推荐出更棒的Python代码示例。

示例1: handle_worker_heartbeat

def handle_worker_heartbeat(worker_name):
    """
    This is a generic function for updating worker heartbeat records.

    Existing Worker objects are searched for one to update. If an existing one is found, it is
    updated. Otherwise a new Worker entry is created. Logging at the info level is also done.

    :param worker_name: The hostname of the worker
    :type  worker_name: basestring
    """
    start = datetime.utcnow()
    existing_worker = Worker.objects(name=worker_name).first()

    if not existing_worker:
        msg = _("New worker '%s' discovered") % worker_name
        _logger.info(msg)

    timestamp = datetime.utcnow()
    msg = _("Worker heartbeat from '{name}' at time {timestamp}").format(timestamp=timestamp,
                                                                         name=worker_name)
    _logger.debug(msg)

    Worker.objects(name=worker_name).update_one(set__last_heartbeat=timestamp,
                                                upsert=True)

    if(datetime.utcnow() - start > timedelta(seconds=PULP_PROCESS_HEARTBEAT_INTERVAL)):
        sec = (datetime.utcnow() - start).total_seconds()
        msg = _("Worker {name} heartbeat time {time}s exceeds heartbeat interval. Consider "
                "adjusting the worker_timeout setting.").format(time=sec, name=worker_name)
        _logger.warn(msg)
开发者ID:alexxa,项目名称:pulp,代码行数:30,代码来源:worker_watcher.py


示例2: _delete_worker

def _delete_worker(name, normal_shutdown=False):
    """
    Delete the Worker with _id name from the database, cancel any associated tasks and reservations

    If the worker shutdown normally, no message is logged, otherwise an error level message is
    logged. Default is to assume the worker did not shut down normally.

    Any resource reservations associated with this worker are cleaned up by this function.

    Any tasks associated with this worker are explicitly canceled.

    :param name:            The name of the worker you wish to delete.
    :type  name:            basestring
    :param normal_shutdown: True if the worker shutdown normally, False otherwise.  Defaults to
                            False.
    :type normal_shutdown:  bool
    """
    if normal_shutdown is False:
        msg = _('The worker named %(name)s is missing. Canceling the tasks in its queue.')
        msg = msg % {'name': name}
        _logger.error(msg)

    # Delete the worker document
    Worker.objects(name=name).delete()

    # Delete all reserved_resource documents for the worker
    ReservedResource.objects(worker_name=name).delete()

    # Cancel all of the tasks that were assigned to this worker's queue
    for task_status in TaskStatus.objects(worker_name=name,
                                          state__in=constants.CALL_INCOMPLETE_STATES):
        cancel(task_status['task_id'])

    # Delete working directory
    common_utils.delete_worker_working_directory(name)
开发者ID:hjensas,项目名称:pulp,代码行数:35,代码来源:tasks.py


示例3: test_resource_not_in_resource_map

    def test_resource_not_in_resource_map(self):
        """
        Test _release_resource() with a resource that is not in the database. This should be
        gracefully handled, and result in no changes to the database.
        """
        # Set up two workers
        worker_1 = Worker(WORKER_1, datetime.utcnow())
        worker_1.save()
        worker_2 = Worker(WORKER_2, datetime.utcnow())
        worker_2.save()
        # Set up two resource reservations, using our workers from above
        reserved_resource_1 = ReservedResource(str(uuid.uuid4()), worker_1.name, 'resource_1')
        reserved_resource_1.save()
        reserved_resource_2 = ReservedResource(str(uuid.uuid4()), worker_2.name, 'resource_2')
        reserved_resource_2.save()

        # This should not raise any Exception, but should also not alter either the Worker
        # collection or the ReservedResource collection
        tasks._release_resource('made_up_resource_id')

        # Make sure that the workers collection has not been altered
        self.assertEqual(Worker.objects().count(), 2)
        worker_1 = Worker.objects().get(name=worker_1.name)
        self.assertTrue(worker_1)
        worker_2 = Worker.objects().get(name=worker_2.name)
        self.assertTrue(worker_2)
        # Make sure that the reserved resources collection has not been altered
        self.assertEqual(ReservedResource.objects.count(), 2)
        rr_1 = ReservedResource.objects.get(task_id=reserved_resource_1.task_id)
        self.assertEqual(rr_1['worker_name'], reserved_resource_1.worker_name)
        self.assertEqual(rr_1['resource_id'], 'resource_1')
        rr_2 = ReservedResource.objects.get(task_id=reserved_resource_2.task_id)
        self.assertEqual(rr_2['worker_name'], reserved_resource_2.worker_name)
        self.assertEqual(rr_2['resource_id'], 'resource_2')
开发者ID:credativ,项目名称:pulp,代码行数:34,代码来源:test_tasks.py


示例4: get_resource_manager_lock

def get_resource_manager_lock(name):
    """
    Tries to acquire the resource manager lock.

    If the lock cannot be acquired immediately, it will wait until the
    currently active instance becomes unavailable, at which point the worker
    cleanup routine will clear the lock for us to acquire. A worker record will
    be created so that the waiting resource manager will appear in the Status
    API. This worker record will be cleaned up through the regular worker
    shutdown routine.

    :param name:   The hostname of the worker
    :type  name:   basestring
    """
    assert name.startswith(constants.RESOURCE_MANAGER_WORKER_NAME)

    lock = ResourceManagerLock(name=name)

    # Whether this is the first lock availability check for this instance
    _first_check = True

    while True:

        now = dateutils.ensure_tz(datetime.utcnow())
        old_timestamp = now - timedelta(seconds=PULP_PROCESS_TIMEOUT_INTERVAL)

        ResourceManagerLock.objects(timestamp__lte=old_timestamp).delete()

        # Create / update the worker record so that Pulp knows we exist
        Worker.objects(name=name).update_one(set__last_heartbeat=datetime.utcnow(),
                                             upsert=True)
        try:
            lock.timestamp = now
            lock.save()

            msg = _("Resource manager '%s' has acquired the resource manager lock") % name
            _logger.debug(msg)

            if not _first_check:
                msg = _("Failover occurred: '%s' is now the primary resource manager") % name
                _logger.warning(msg)

            break
        except mongoengine.NotUniqueError:
            # Only log the message the first time
            if _first_check:
                _logger.info(_("Hot spare pulp_resource_manager instance '%(name)s' detected.")
                             % {'name': name})
                _first_check = False

            time.sleep(PULP_PROCESS_HEARTBEAT_INTERVAL)
开发者ID:alexxa,项目名称:pulp,代码行数:51,代码来源:app.py


示例5: _get_unreserved_worker

def _get_unreserved_worker():
    """
    Return the Worker instance that has no reserved_resource entries
    associated with it. If there are no unreserved workers a
    pulp.server.exceptions.NoWorkers exception is raised.

    :raises NoWorkers: If all workers have reserved_resource entries associated with them.

    :returns:          The Worker instance that has no reserved_resource
                       entries associated with it.
    :rtype:            pulp.server.db.model.resources.Worker
    """

    # Build a mapping of queue names to Worker objects
    workers_dict = dict((worker['name'], worker) for worker in Worker.objects())
    worker_names = workers_dict.keys()
    reserved_names = [r['worker_name'] for r in ReservedResource.objects.all()]

    # Find an unreserved worker using set differences of the names, and filter
    # out workers that should not be assigned work.
    # NB: this is a little messy but set comprehensions are in python 2.7+
    unreserved_workers = set(filter(_is_worker, worker_names)) - set(reserved_names)

    try:
        return workers_dict[unreserved_workers.pop()]
    except KeyError:
        # All workers are reserved
        raise NoWorkers()
开发者ID:hjensas,项目名称:pulp,代码行数:28,代码来源:tasks.py


示例6: _delete_worker

def _delete_worker(name, normal_shutdown=False):
    """
    Delete the Worker with _id name from the database, cancel any associated tasks and reservations

    If the worker shutdown normally, no message is logged, otherwise an error level message is
    logged. Default is to assume the worker did not shut down normally.

    Any resource reservations associated with this worker are cleaned up by this function.

    Any tasks associated with this worker are explicitly canceled.

    :param name:            The name of the worker you wish to delete.
    :type  name:            basestring
    :param normal_shutdown: True if the worker shutdown normally, False otherwise.  Defaults to
                            False.
    :type normal_shutdown:  bool
    """
    if normal_shutdown is False:
        msg = _('The worker named %(name)s is missing. Canceling the tasks in its queue.')
        msg = msg % {'name': name}
        _logger.error(msg)
    else:
        msg = _("Cleaning up shutdown worker '%s'.") % name
        _logger.info(msg)

    # Delete the worker document
    Worker.objects(name=name).delete()

    # Delete all reserved_resource documents for the worker
    ReservedResource.objects(worker_name=name).delete()

    # If the worker is a resource manager, we also need to delete the associated lock
    if name.startswith(RESOURCE_MANAGER_WORKER_NAME):
        ResourceManagerLock.objects(name=name).delete()

    # If the worker is a scheduler, we also need to delete the associated lock
    if name.startswith(SCHEDULER_WORKER_NAME):
        CeleryBeatLock.objects(name=name).delete()

    # Cancel all of the tasks that were assigned to this worker's queue
    for task_status in TaskStatus.objects(worker_name=name,
                                          state__in=constants.CALL_INCOMPLETE_STATES):
        cancel(task_status['task_id'], revoke_task=False)
开发者ID:alexxa,项目名称:pulp,代码行数:43,代码来源:tasks.py


示例7: get_resource_manager_lock

def get_resource_manager_lock(name):
    """
    Tries to acquire the resource manager lock.

    If the lock cannot be acquired immediately, it will wait until the
    currently active instance becomes unavailable, at which point the worker
    cleanup routine will clear the lock for us to acquire. A worker record will
    be created so that the waiting resource manager will appear in the Status
    API. This worker record will be cleaned up through the regular worker
    shutdown routine.

    :param name:   The hostname of the worker
    :type  name:   basestring
    """
    assert name.startswith(constants.RESOURCE_MANAGER_WORKER_NAME)

    lock = ResourceManagerLock(name=name)

    # Whether this is the first lock availability check for this instance
    _first_check = True

    while True:
        # Create / update the worker record so that Pulp knows we exist
        Worker.objects(name=name).update_one(set__last_heartbeat=datetime.utcnow(),
                                             upsert=True)
        try:
            lock.save()

            msg = _("Resource manager '%s' has acquired the resource manager lock") % name
            _logger.info(msg)
            break
        except mongoengine.NotUniqueError:
            # Only log the message the first time
            if _first_check:
                msg = _("Resource manager '%(name)s' attempted to acquire the the resource manager "
                        "lock but was unable to do so. It will retry every %(interval)d seconds "
                        "until the lock can be acquired.") % \
                    {'name': name, 'interval': constants.CELERY_CHECK_INTERVAL}
                _logger.info(msg)
                _first_check = False

            time.sleep(constants.CELERY_CHECK_INTERVAL)
开发者ID:pulp,项目名称:pulp,代码行数:42,代码来源:app.py


示例8: handle_worker_heartbeat

def handle_worker_heartbeat(event):
    """
    Celery event handler for 'worker-heartbeat' events.

    The event is first parsed and logged.  Then the existing Worker objects are
    searched for one to update. If an existing one is found, it is updated.
    Otherwise a new Worker entry is created. Logging at the info and debug
    level is also done.

    :param event: A celery event to handle.
    :type event: dict
    """
    event_info = _parse_and_log_event(event)
    worker = Worker.objects(name=event_info['worker_name']).first()

    if not worker:
        msg = _("New worker '%(worker_name)s' discovered") % event_info
        _logger.info(msg)

    Worker.objects(name=event_info['worker_name']).\
        update_one(set__last_heartbeat=event_info['local_received'], upsert=True)
开发者ID:alanoe,项目名称:pulp,代码行数:21,代码来源:worker_watcher.py


示例9: test_resource_in_resource_map

    def test_resource_in_resource_map(self):
        """
        Test _release_resource() with a valid resource. This should remove the resource from the
        database.
        """
        # Set up two workers
        now = datetime.utcnow()
        worker_1 = Worker(name=WORKER_1, last_heartbeat=now)
        worker_1.save()
        worker_2 = Worker(name=WORKER_2, last_heartbeat=now)
        worker_2.save()
        # Set up two reserved resources
        reserved_resource_1 = ReservedResource(task_id=str(uuid.uuid4()),
                                               worker_name=worker_1.name, resource_id='resource_1')
        reserved_resource_1.save()
        reserved_resource_2 = ReservedResource(task_id=str(uuid.uuid4()),
                                               worker_name=worker_2.name, resource_id='resource_2')
        reserved_resource_2.save()

        # This should remove resource_2 from the _resource_map.
        tasks._release_resource(reserved_resource_2.task_id)

        # resource_2 should have been removed from the database
        self.assertEqual(ReservedResource.objects.count(), 1)
        rr_1 = ReservedResource.objects.get(task_id=reserved_resource_1.task_id)
        self.assertEqual(rr_1['worker_name'], reserved_resource_1.worker_name)
        self.assertEqual(rr_1['resource_id'], 'resource_1')
开发者ID:zjhuntin,项目名称:pulp,代码行数:27,代码来源:test_tasks.py


示例10: get_worker_for_reservation

def get_worker_for_reservation(resource_id):
    """
    Return the Worker instance that is associated with a reservation of type resource_id. If
    there are no workers with that reservation_id type a pulp.server.exceptions.NoWorkers
    exception is raised.

    :param resource_id:    The name of the resource you wish to reserve for your task.

    :raises NoWorkers:     If all workers have reserved_resource entries associated with them.

    :type resource_id:     basestring
    :returns:              The Worker instance that has a reserved_resource entry of type
                           `resource_id` associated with it.
    :rtype:                pulp.server.db.model.resources.Worker
    """
    reservation = ReservedResource.objects(resource_id=resource_id).first()
    if reservation:
        return Worker.objects(name=reservation['worker_name']).first()
    else:
        raise NoWorkers()
开发者ID:hjensas,项目名称:pulp,代码行数:20,代码来源:tasks.py


示例11: check_workers

    def check_workers(self):
        """
        Look for missing workers, and dispatch a cleanup task if one goes missing.

        To find a missing worker, filter the Workers model for entries older than
        utcnow() - WORKER_TIMEOUT_SECONDS. The heartbeat times are stored in native UTC, so this is
        a comparable datetime.

        For each missing worker found, dispatch a _delete_worker task requesting that the resource
        manager delete the Worker and cleanup any associated work.

        This method logs and the debug and error levels.
        """
        msg = _(
            'Looking for workers missing for more than %s seconds') % self.WORKER_TIMEOUT_SECONDS
        _logger.debug(msg)
        oldest_heartbeat_time = datetime.utcnow() - timedelta(seconds=self.WORKER_TIMEOUT_SECONDS)
        worker_list = Worker.objects(last_heartbeat__lt=oldest_heartbeat_time)
        for worker in worker_list:
            msg = _("Workers '%s' has gone missing, removing from list of workers") % worker.name
            _logger.error(msg)
            _delete_worker(worker.name)
开发者ID:hjensas,项目名称:pulp,代码行数:22,代码来源:scheduler.py


示例12: get_workers

def get_workers():
    """
    :returns:          list of workers with their heartbeats
    :rtype:            list
    """
    return Worker.objects()
开发者ID:pcreech,项目名称:pulp,代码行数:6,代码来源:status.py


示例13: tearDown

 def tearDown(self):
     Worker.objects().delete()
     ReservedResource.objects.delete()
     TaskStatus.objects().delete()
开发者ID:credativ,项目名称:pulp,代码行数:4,代码来源:base.py



注:本文中的pulp.server.db.model.Worker类示例由纯净天空整理自Github/MSDocs等源码及文档管理平台,相关代码片段筛选自各路编程大神贡献的开源项目,源码版权归原作者所有,传播和使用请参考对应项目的License;未经允许,请勿转载。


鲜花

握手

雷人

路过

鸡蛋
该文章已有0人参与评论

请发表评论

全部评论

专题导读
上一篇:
Python auth.Permission类代码示例发布时间:2022-05-25
下一篇:
Python model.TaskStatus类代码示例发布时间:2022-05-25
热门推荐
阅读排行榜

扫描微信二维码

查看手机版网站

随时了解更新最新资讯

139-2527-9053

在线客服(服务时间 9:00~18:00)

在线QQ客服
地址:深圳市南山区西丽大学城创智工业园
电邮:jeky_zhao#qq.com
移动电话:139-2527-9053

Powered by 互联科技 X3.4© 2001-2213 极客世界.|Sitemap