hey neil,
neutron.conf has a section called ' [keystone_authtoken]’ which has credentials 
to query keystone as neutron. you can read the config as you’d typically do 
from the mechanism driver for any other property using oslo.config.
you could then use python-keystoneclient with those creds to query the mapping. 
a sample is given in the keystoneclient repo [1].
via telegram
[1] 
https://github.com/openstack/python-keystoneclient/blob/650716d0dd30a73ccabe3f0ec20eb722ca0d70d4/keystoneclient/v3/client.py#L102-L116
 On Tue, Jul 17, 2018 at 9:58 PM, Neil Jerram <[email protected]> wrote:
On Tue, Jul 17, 2018 at 3:55 PM Jay Pipes < [email protected] 
[[email protected]] > wrote:
On 07/17/2018 03:36 AM, Neil Jerram wrote:
> Can someone help me with how to look up a project name (aka tenant name)
> for a known project/tenant ID, from code (specifically a mechanism
> driver) running in the Neutron server?
>
> I believe that means I need to make a GET REST call as here:
> https://developer.openstack.org/api-ref/identity/v3/index.html#projects 
> [https://developer.openstack.org/api-ref/identity/v3/index.html#projects] . 
> But
> I don't yet understand how a piece of Neutron server code can ensure
> that it has the right credentials to do that. If someone happens to
> have actual code for doing this, I'm sure that would be very helpful.
>
> (I'm aware that whenever the Neutron server processes an API request,
> the project name for the project that generated that request is added
> into the request context. That is great when my code is running in an
> API request context. But there are other times when the code isn't in a
> request context and still needs to map from a project ID to project
> name; hence the question here.)

Hi Neil,

You basically answered your own question above :) The neutron request
context gets built from oslo.context's Context.from_environ() [1] which
has this note in the implementation [2]:

# Load a new context object from the environment variables set by
# auth_token middleware. See:
#
https://docs.openstack.org/keystonemiddleware/latest/api/keystonemiddleware.auth_token.html#what-auth-token-adds-to-the-request-for-use-by-the-openstack-service
 
[https://docs.openstack.org/keystonemiddleware/latest/api/keystonemiddleware.auth_token.html#what-auth-token-adds-to-the-request-for-use-by-the-openstack-service]

So, basically, simply look at the HTTP headers for HTTP_X_PROJECT_NAME.
If you don't have access to a HTTP headers, then you'll need to pass
some context object/struct to the code you're referring to. Might as
well pass the neutron RequestContext (derived from oslo_context.Context)
to the code you're referring to and you get all this for free.

Best,
-jay

[1]
https://github.com/openstack/oslo.context/blob/4abd5377e4d847102a4e87a528d689e31cc1713c/oslo_context/context.py#L424
 
[https://github.com/openstack/oslo.context/blob/4abd5377e4d847102a4e87a528d689e31cc1713c/oslo_context/context.py#L424]

[2]
https://github.com/openstack/oslo.context/blob/4abd5377e4d847102a4e87a528d689e31cc1713c/oslo_context/context.py#L433-L435
 
[https://github.com/openstack/oslo.context/blob/4abd5377e4d847102a4e87a528d689e31cc1713c/oslo_context/context.py#L433-L435]
Many thanks for this reply, Jay.
If I'm understanding fully, I believe it all works beautifully so long as the 
Neutron server is processing a specific API request, e.g. a port CRUD 
operation. Then, as you say, the RequestContext includes the name of the 
project/tenant that originated that request.
I have an additional requirement, though, to do a occasional audit of standing 
resources in the Neutron DB, and to check that my mechanism driver's 
programming for them is correct. To do that, I have an independent eventlet 
thread that runs in admin context and occasionally queries Neutron resources, 
e.g. all the ports. For each port, the Neutron DB data includes the project_id, 
but not project_name, and I'd like at that point to be able to map from the 
project_id for each port to project_name.
Do you have any thoughts on how I could do that? (E.g. perhaps there is some 
way of generating and looping round a request with the project_id, such that 
the middleware populates the project_name... but that sounds a bit baroque; I 
would hope that there would be a way of doing a simpler Keystone DB lookup.)
Regards, Neil

__________________________________________________________________________ 
OpenStack Development Mailing List (not for usage questions) Unsubscribe: 
[email protected]?subject:unsubscribe 
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: [email protected]?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev

Reply via email to