In addition to the reproducible the Vm has to be tagged with "lifecycle: Fully retire VM and remove from Provider" else it does not even try to run the refresh.

Ash,

What is the RHV version, used by the customer please?

Correction to comment 4:
RHV version was 3.6.10-0.1.el6, and not 4.1.6.

Boris,

I wanted to verify it using same verification steps I used for the 5.8.2 bug, mentioned here:
https://bugzilla.redhat.com/show_bug.cgi?id=1484613 , comment 6:

1) In ovirt gem, at lib/ovirt/services.rb in the method resource_verb, inside the block:
resource.send(verb, *args)
add the following lines (in the current gem version 0.17.1 its line 319 of the file):
>>>>>
if resource.url.include?("storagedomains") && resource.url.include?("disks")
  response = OpenStruct.new(code: 404)
end
<<<<<
...
However,
On CFME-5.9.0.14, the services.rb file (/opt/rh/rh-ruby23/root/usr/share/gems/gems/ovirt-engine-sdk-4.2.1/lib/ovirtsdk4/services.rb)  is different.

So I can no longer follow this step 1).

Is there an alternative change I can do in this file (or elsewhere), to achieve the desired Storage Domain Error please?

Thanks.

With Boris's help,

Verified on CFME-5.9.0.16/RHV-3.6.12-4,
Using this flow:

1) In ovirt gem, at file:  
/opt/rh/cfme-gemset/gems/ovirt-0.18.1/lib/ovirt/service.rb
in the method resource_verb, inside the block:
resource.send(verb, *args)
add the following lines (in the current gem version 0.18.1 its line 316 in this file):

   def resource_verb(path, verb, *args)
      log_header = "#{self.class.name}#resource_#{verb}"
      resource   = create_resource(path)
      logger.info "#{log_header}: Sending URL: <#{resource.url}>"
      logger.debug "#{log_header}: With args: <#{args.inspect}>"
      resource.send(verb, *args) do |response, request, result, &block|
>>>>>
        if resource.url.include?("storagedomains") && resource.url.include?("disks")
          response = OpenStruct.new(code: 404)
        end
<<<<<
        case response.code

2)     # vmdb
   vmdb# rake evm:restart

3) Added RHV-3.6 that has a VM with 2 disks.

4) In the VM details page, under "Policy"-> "Edit tags": Set the tags for the vm to be: 
"LifeCycle", "Fully retire VM and remove from Provider"

5) In the VMs page select the VM and under Lifecycle choose "retire selected items"

In rhevm.log there are only 2 related ERRORs [1],
and then such error did not repeat.

The VM was removed.

[1]

[----] E, [2018-01-11T07:17:50.184435 #63952:d8b134] ERROR -- : Ovirt::Service#parse_error_response Return from URL: <https://<RHV-FQDN>/ovirt-engine/api/storagedomains/8f5fe6de-fbd7-4b5f-a9dd-a4b3032e1185/disks/776d40ea-d16a-46dc-84a4-2ae4e290b103> Data:#<OpenStruct code=404>
[----] I, [2018-01-11T07:17:50.184931 #63952:d8b134]  INFO -- : Ovirt::Service#resource_get: Sending URL: <https://<RHV-FQDN>/ovirt-engine/api/storagedomains/8f5fe6de-fbd7-4b5f-a9dd-a4b3032e1185/disks/359d962b-2f45-4883-ad71-9bf97ee9380a>

[----] E, [2018-01-11T07:17:50.329430 #63952:d8b134] ERROR -- : Ovirt::Service#parse_error_response Return from URL: <https://<RHV-FQDN>/ovirt-engine/api/storagedomains/8f5fe6de-fbd7-4b5f-a9dd-a4b3032e1185/disks/359d962b-2f45-4883-ad71-9bf97ee9380a> Data:#<OpenStruct code=404>
[----] I, [2018-01-11T07:17:59.585753 #63924:d8b134]  INFO -- : Ovirt::Service#resource_get: Sending URL: <https://<RHV-FQDN>/ovirt-engine/api/users/0000001a-001a-001a-001a-0000000001f7>