Failed to operate legacy cnf

[Ayumu Ueha]

When I performed cnf operation (legacy) according to the procedure described in the document [1], the following error occurred.

1. vnf create error caused by policies parameter
2. vnf scale error due to key error of policy[]
3. no retry interval during vnf delete wait

[1. vnf create error caused by policies parameter]
==================================================
When vnf create operation, the following error occurred.

$ openstack vnf descriptor create --vnfd-file tosca-vnfd-containerized.yaml VNFD1
tosca-parser failed: -
The pre-parsed input failed validation with the following error(s):
        MissingRequiredFieldError: "properties" of template "SP1" is missing required field "['increment', 'default_instances']".
                File /usr/local/lib/python3.6/dist-packages/eventlet/greenthread.py, line 221, in main

This error is due to a change that now causes tosca-parser to check the required parameter of policies [2].

To avoid the above error, I added the required parameter to VNFD policies and re-executed create, but the following error occurred.

$ openstack vnf create --vnfd-name VNFD1 --vim-name vim-kubernetes VNF1
Found unsupported keys for ['default_instances', 'increment']

To avoid this "Found unsupported keys" error, it is necessary to remove the checking for unsupported keys of policies [3].

[2. vnf scale error due to key error of policy[]]
=================================================
After I fixed above create error, It failed to scale vnf because of following error.

$ openstack vnf scale --scaling-policy-name SP1 --scaling-type out VNF1
Request Failed: internal server error while processing your request.

[req-1ef50574-e3c8-4423-bc68-61daf38ecafc admin admin] create failed: No details.: KeyError: 'vnf_instance_id'
Traceback (most recent call last):
  File "/opt/stack/tacker/tacker/api/v1/resource.py", line 77, in resource
    result = method(request=request, **args)
  File "/opt/stack/tacker/tacker/api/v1/base.py", line 394, in create
    obj = obj_creator(request.context, **kwargs)
  File "/opt/stack/tacker/tacker/vnfm/plugin.py", line 912, in create_vnf_scale
    self._handle_vnf_scaling(context, policy_)
  File "/opt/stack/tacker/tacker/vnfm/plugin.py", line 850, in _handle_vnf_scaling
    last_event_id = _vnf_policy_action()
  File "/opt/stack/tacker/tacker/vnfm/plugin.py", line 814, in _vnf_policy_action
    _handle_vnf_scaling_post(constants.ERROR)
  File "/usr/local/lib/python3.6/dist-packages/oslo_utils/excutils.py", line 220, in __exit__
    self.force_reraise()
  File "/usr/local/lib/python3.6/dist-packages/oslo_utils/excutils.py", line 196, in force_reraise
    six.reraise(self.type_, self.value, self.tb)
  File "/usr/local/lib/python3.6/dist-packages/six.py", line 703, in reraise
    raise value
  File "/opt/stack/tacker/tacker/vnfm/plugin.py", line 802, in _vnf_policy_action
    region_name=region_name
  File "/opt/stack/tacker/tacker/common/driver_manager.py", line 71, in invoke
    return getattr(driver, method_name)(**kwargs)
  File "/opt/stack/tacker/tacker/common/log.py", line 35, in wrapper
    return method(*args, **kwargs)
  File "/opt/stack/tacker/tacker/vnfm/infra_drivers/kubernetes/kubernetes_driver.py", line 1127, in scale
    context, policy['vnf_instance_id'])
KeyError: 'vnf_instance_id'

`policy['vnf_instance_id']` is added for SOL CNF scale route in vnflcmDriver in Wallaby release development [4].

But in legacy route, `vnf_instance_id` is not added to `policy[]`, so this error is occurred.
The current implementation determines legacy or sol based on the existence of vnf_resource.

To avoid this error, it is necessary to change the legacy or sol decision, and I think it's best to use `policy ['vnf_instance_id']`.

[3. no retry interval during vnf delete wait]
=============================================
I found a problem that the wait process ends immediately, because there is no interval of wait retry process [5].

If the deletion finishes immediately, there is no problem, but if it takes time, an error will occur.

Reference
======
[1] https://docs.openstack.org/tacker/latest/user/containerized_vnf_usage_guide.html
[2] https://review.opendev.org/c/openstack/tosca-parser/+/763144
[3] https://opendev.org/openstack/tacker/src/branch/master/tacker/vnfm/infra_drivers/kubernetes/k8s/translate_inputs.py#L192-L193
[4] https://review.opendev.org/c/openstack/tacker/+/765786
[5] https://opendev.org/openstack/tacker/src/branch/master/tacker/vnfm/infra_drivers/kubernetes/kubernetes_driver.py#L928-L930

[Ayumu Ueha (ueha)]

Fix proposed to branch: master
Review: https://review.opendev.org/c/openstack/tacker/+/783626

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/tacker 5.0.0.0rc3 release candidate.

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/tacker 6.0.0.0rc1 release candidate.