yahoo-eng-team team mailing list archive

[kevinzhao <1816543@xxxxxxxxxxxxxxxxxx>]

Public bug reported:

Description
===========
nova service-delete report  ComputeHostNotFound when deleting nova-compute service after I delete other nova service(nova-consoleauth) on the same compute node.
The compute_node should be removed according to the binary of the service to be deleted. When the binary of the service to be deleted is nova-compute, it is appropriate to delete the compute_node.

Steps to reproduce
==================
1) nail1 is an all in one environment,there are nova-compute and nova-consoleauth service on host nail1
2) remove all instances on hypervisor nail1
[root@nail1 ~]# nova service-list
+--------------------------------------+------------------+-----------------+----------+---------+-------+----------------------------+-----------------+-------------+
| Id                                   | Binary           | Host            | Zone     | Status  | State | Updated_at                 | Disabled Reason | Forced down |
+--------------------------------------+------------------+-----------------+----------+---------+-------+----------------------------+-----------------+-------------+
| b4ca49a8-c3a9-4fc8-b9a8-f2d662e26060 | nova-conductor   | nail1           | internal | enabled | up    | 2019-02-19T06:39:49.000000 | -               | False       |
| e6ae7de7-d8dc-4364-84ed-1845fe967cb6 | nova-scheduler   | nail1           | internal | enabled | up    | 2019-02-19T06:39:43.000000 | -               | False       |
| ea3689d5-ace1-4561-acab-369b4e067053 | nova-compute     | nail1           | nova     | enabled | down  | 2019-02-19T06:35:41.000000 | -               | False       |
| 25da267f-9b7c-4cef-8044-9b26fc2aa18a | nova-compute     | nail2           | nova     | enabled | up    | 2019-02-19T06:39:50.000000 | -               | False       |
| 90686f1f-6a16-4c97-af9d-bdedb9ebec7d | nova-consoleauth | nail1           | internal | enabled | down  | 2019-02-19T06:37:48.000000 | -               | False       |

+--------------------------------------+------------------+-----------------+----------+---------+-------+----------------------------+-----------------+-------------+

3) delete nova-consoleauth service on nail1
[root@nail1 ~]# nova service-delete 90686f1f-6a16-4c97-af9d-bdedb9ebec7d


4) delete nova-compute service on hypervisor nail1

Actual result
=============
[root@nail1 ~]# nova service-delete ea3689d5-ace1-4561-acab-369b4e067053
ERROR (ClientException): Unexpected API Error. Please report this at http://bugs.launchpad.net/nova/ and attach the Nova API log if possible.
<class 'nova.exception.ComputeHostNotFound'> (HTTP 500) (Request-ID: req-f283de97-7f00-4eae-af77-9155a7b9395d)


Environment
===========
[root@nail1 ~]# rpm -qa|grep openstack-nova-compute
openstack-nova-compute-18.0.2-1.el7.noarch

hypervisor: Libvirt + KVM


The relevant code is as follows:
nova/db/sqlalchemy/api.py

@pick_context_manager_writer
def service_destroy(context, service_id):
    service = service_get(context, service_id)

    model_query(context, models.Service).\
                filter_by(id=service_id).\
                soft_delete(synchronize_session=False)

    # TODO(sbauza): Remove the service_id filter in a later release
    # once we are sure that all compute nodes report the host field
    model_query(context, models.ComputeNode).\
                filter(or_(models.ComputeNode.service_id == service_id,
                           models.ComputeNode.host == service['host'])).\
                soft_delete(synchronize_session=False)

** Affects: nova
     Importance: Undecided
         Status: New

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1816543

Title:
  nova service-delete report  ComputeHostNotFound when delete compute
  service after I delete  other nova service on the same compute node

Status in OpenStack Compute (nova):
  New

Bug description:
  Description
  ===========
  nova service-delete report  ComputeHostNotFound when deleting nova-compute service after I delete other nova service(nova-consoleauth) on the same compute node.
  The compute_node should be removed according to the binary of the service to be deleted. When the binary of the service to be deleted is nova-compute, it is appropriate to delete the compute_node.

  Steps to reproduce
  ==================
  1) nail1 is an all in one environment,there are nova-compute and nova-consoleauth service on host nail1
  2) remove all instances on hypervisor nail1
  [root@nail1 ~]# nova service-list
  +--------------------------------------+------------------+-----------------+----------+---------+-------+----------------------------+-----------------+-------------+
  | Id                                   | Binary           | Host            | Zone     | Status  | State | Updated_at                 | Disabled Reason | Forced down |
  +--------------------------------------+------------------+-----------------+----------+---------+-------+----------------------------+-----------------+-------------+
  | b4ca49a8-c3a9-4fc8-b9a8-f2d662e26060 | nova-conductor   | nail1           | internal | enabled | up    | 2019-02-19T06:39:49.000000 | -               | False       |
  | e6ae7de7-d8dc-4364-84ed-1845fe967cb6 | nova-scheduler   | nail1           | internal | enabled | up    | 2019-02-19T06:39:43.000000 | -               | False       |
  | ea3689d5-ace1-4561-acab-369b4e067053 | nova-compute     | nail1           | nova     | enabled | down  | 2019-02-19T06:35:41.000000 | -               | False       |
  | 25da267f-9b7c-4cef-8044-9b26fc2aa18a | nova-compute     | nail2           | nova     | enabled | up    | 2019-02-19T06:39:50.000000 | -               | False       |
  | 90686f1f-6a16-4c97-af9d-bdedb9ebec7d | nova-consoleauth | nail1           | internal | enabled | down  | 2019-02-19T06:37:48.000000 | -               | False       |

  +--------------------------------------+------------------+-----------------+----------+---------+-------+----------------------------+-----------------+-------------+

  3) delete nova-consoleauth service on nail1
  [root@nail1 ~]# nova service-delete 90686f1f-6a16-4c97-af9d-bdedb9ebec7d

  
  4) delete nova-compute service on hypervisor nail1

  Actual result
  =============
  [root@nail1 ~]# nova service-delete ea3689d5-ace1-4561-acab-369b4e067053
  ERROR (ClientException): Unexpected API Error. Please report this at http://bugs.launchpad.net/nova/ and attach the Nova API log if possible.
  <class 'nova.exception.ComputeHostNotFound'> (HTTP 500) (Request-ID: req-f283de97-7f00-4eae-af77-9155a7b9395d)

  
  Environment
  ===========
  [root@nail1 ~]# rpm -qa|grep openstack-nova-compute
  openstack-nova-compute-18.0.2-1.el7.noarch

  hypervisor: Libvirt + KVM


  The relevant code is as follows:
  nova/db/sqlalchemy/api.py

  @pick_context_manager_writer
  def service_destroy(context, service_id):
      service = service_get(context, service_id)

      model_query(context, models.Service).\
                  filter_by(id=service_id).\
                  soft_delete(synchronize_session=False)

      # TODO(sbauza): Remove the service_id filter in a later release
      # once we are sure that all compute nodes report the host field
      model_query(context, models.ComputeNode).\
                  filter(or_(models.ComputeNode.service_id == service_id,
                             models.ComputeNode.host == service['host'])).\
                  soft_delete(synchronize_session=False)

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1816543/+subscriptions