Newton版Openstack創建云主機失敗怎么辦

這篇文章主要介紹Newton版Openstack創建云主機失敗怎么辦，文中介紹的非常詳細，具有一定的參考價值，感興趣的小伙伴們一定要看完！

1.錯誤背景

安裝了N版Openstack Ha，在創建云主機時源選擇鏡像并創建新卷時，云主機創建失敗。檢查發現，卷其實創建成功了并且可用，可以通過創建的卷啟動云主機。使用cirros測試鏡像，不會出現問題，使用CentOS7.3會出現這種報錯，不僅同時創建2個，就連一次只創建一個也會報錯。

2.檢查日志

2017-06-06 09:30:28.513 665097 WARNING nova.virt.block_device [req-c5016b48-399b-44f8-adea-a86421f1e129 c49d86f8a39148bb809d2da44811ca3b f19126d9035d4209b363386e8e267129 - - -] Failed to delete volume: dd04f3d2-63cb-45f2-9a31-9ecc523bd4
0f due to Invalid input received: Invalid volume: Volume status must be available or error or error_restoring or error_extending and must not be migrating, attached, belong to a group or have snapshots. (HTTP 400) (Request-ID: req-0bd8e
4b4-6989-4c34-a623-20d2bff98816)
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [req-c5016b48-399b-44f8-adea-a86421f1e129 c49d86f8a39148bb809d2da44811ca3b f19126d9035d4209b363386e8e267129 - - -] [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94] Instance faile
d block device setup
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94] Traceback (most recent call last):
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 1586, in _prep_block_device
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     wait_func=self._await_block_device_map_created)
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/virt/block_device.py", line 514, in attach_block_devices
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     map(_log_and_attach, block_device_mapping)
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/virt/block_device.py", line 512, in _log_and_attach
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     bdm.attach(*attach_args, **attach_kwargs)
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/virt/block_device.py", line 404, in attach
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     self._call_wait_func(context, wait_func, volume_api, vol['id'])
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/virt/block_device.py", line 363, in _call_wait_func
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     {'volume_id': volume_id, 'exc': exc})
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/oslo_utils/excutils.py", line 220, in __exit__
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     self.force_reraise()
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/oslo_utils/excutils.py", line 196, in force_reraise
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     six.reraise(self.type_, self.value, self.tb)
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/virt/block_device.py", line 353, in _call_wait_func
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     wait_func(context, volume_id)
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 1258, in _await_block_device_map_created
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     volume_status=volume_status)
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94] VolumeNotCreated: Volume dd04f3d2-63cb-45f2-9a31-9ecc523bd40f did not finish being created even after we waited 189 seconds or 61
attempts. And its status is downloading.
2017-06-06 09:30:28.514 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]
2017-06-06 09:30:28.638 665097 WARNING nova.scheduler.client.report [req-c5016b48-399b-44f8-adea-a86421f1e129 c49d86f8a39148bb809d2da44811ca3b f19126d9035d4209b363386e8e267129 - - -] No authentication information found for placement API
. Placement is optional in Newton, but required in Ocata. Please enable the placement service before upgrading.
2017-06-06 09:30:28.638 665097 WARNING nova.scheduler.client.report [req-c5016b48-399b-44f8-adea-a86421f1e129 c49d86f8a39148bb809d2da44811ca3b f19126d9035d4209b363386e8e267129 - - -] Unable to refresh my resource provider record
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [req-c5016b48-399b-44f8-adea-a86421f1e129 c49d86f8a39148bb809d2da44811ca3b f19126d9035d4209b363386e8e267129 - - -] [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94] Build of insta
nce 30e8a024-d1c0-4a64-aae7-a830e8c6be94 aborted: Block Device Mapping is Invalid.
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94] Traceback (most recent call last):
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 1783, in _do_build_and_run_instance
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     filter_properties)
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 1943, in _build_and_run_instance
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     'create.error', fault=e)
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/oslo_utils/excutils.py", line 220, in __exit__
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     self.force_reraise()
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/oslo_utils/excutils.py", line 196, in force_reraise
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     six.reraise(self.type_, self.value, self.tb)
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 1908, in _build_and_run_instance
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     block_device_mapping) as resources:
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib64/python2.7/contextlib.py", line 17, in __enter__
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     return self.gen.next()
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]   File "/usr/lib/python2.7/site-packages/nova/compute/manager.py", line 2071, in _build_resources
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]     reason=e.format_message())
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94] BuildAbortException: Build of instance 30e8a024-d1c0-4a64-aae7-a830e8c6be94 aborted: Block Device Mapping is Invalid.
2017-06-06 09:30:28.639 665097 ERROR nova.compute.manager [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94]
2017-06-06 09:30:29.364 665097 INFO nova.compute.manager [req-c5016b48-399b-44f8-adea-a86421f1e129 c49d86f8a39148bb809d2da44811ca3b f19126d9035d4209b363386e8e267129 - - -] [instance: 30e8a024-d1c0-4a64-aae7-a830e8c6be94] Took 0.72 secon
ds to deallocate network for instance.

3.錯誤分析：

底層存儲Ceph創建新卷的時候速度太慢，導致nova連接卷超時或者超出連接次數 兩個解決辦法： 一個是提高ceph讀寫速度（增加ssd作為journal），條件不允許,此方法作罷。 另一個是尋找是否有相關參數設定連接時間或者連接次數。

4.源碼查看

錯誤信息中看到了_await_block_device_map_created 找到了對應的函數

def _await_block_device_map_created(self, context, vol_id):
        # TODO(yamahata): creating volume simultaneously
        #                 reduces creation time?
        # TODO(yamahata): eliminate dumb polling
        start = time.time()
        retries = CONF.block_device_allocate_retries
        if retries < 0:
            LOG.warning(_LW("Treating negative config value (%(retries)s) for "
                            "'block_device_retries' as 0."),
                        {'retries': retries})
        # (1) treat  negative config value as 0
        # (2) the configured value is 0, one attempt should be made
        # (3) the configured value is > 0, then the total number attempts
        #      is (retries + 1)
        attempts = 1
        if retries >= 1:
            attempts = retries + 1
        for attempt in range(1, attempts + 1):
            volume = self.volume_api.get(context, vol_id)
            volume_status = volume['status']
            if volume_status not in ['creating', 'downloading']:
                if volume_status == 'available':
                    return attempt
                LOG.warning(_LW("Volume id: %(vol_id)s finished being "
                                "created but its status is %(vol_status)s."),
                            {'vol_id': vol_id,
                             'vol_status': volume_status})
                break
            greenthread.sleep(CONF.block_device_allocate_retries_interval)
        raise exception.VolumeNotCreated(volume_id=vol_id,
                                         seconds=int(time.time() - start),
                                         attempts=attempt,
                                         volume_status=volume_status)

有兩個參數block_device_allocate_retries和block_device_allocate_retries_interval，看名字可以知道一個是嘗試次數，一個是嘗試間隔

5.解決經過

參數默認值

block_device_allocate_retries_interval=3
block_device_allocate_retries=60

先修改間隔
block_device_allocate_retries_interval=10

重啟nova-compute

systemctl restart openstack-nova-compute.service

再以同樣的方式創建一臺云主機，發現可以創建成功。

以上是“Newton版Openstack創建云主機失敗怎么辦”這篇文章的所有內容，感謝各位的閱讀！希望分享的內容對大家有幫助，更多相關知識，歡迎關注億速云行業資訊頻道！