Re: [Openstack] Authorization Question

2013-03-04 Thread Dolph Mathews
That's correct. Right now, all endpoints registered in keystone are
returned to all users, regardless of whether they actually have any sort of
authorization on those endpoints.

I suspect we'll be having a planning session at the design summit on this
topic -- I'd be helpful to better understand your ideal use case in
suppressing endpoints from the catalog? In Grizzly, users, groups,
projects, and domains have absolutely no relationship with services and
endpoints within keystone. That becomes deployment specific when you
consider how RBAC is applied service-side with policy.json, etc, but those
services are not necessarily aware of how they appear in the service
catalog, nor does keystone interpret policy files other than it's own.


-Dolph


On Mon, Mar 4, 2013 at 2:32 PM, Miller, Mark M (EB SW Cloud - RD -
Corvallis) mark.m.mil...@hp.com wrote:

  Hello,

 ** **

 I have been looking over the Keystone v3 API documentation as well as the
 database table columns. My question concerns endpoint access restrictions.
 I don’t see any noticeable way to associate endpoints with domains which
 means that any user can access any endpoint of any domain. Is this correct?
 The only database column that might come into play is the region column of
 the endpoint table.

 ** **

 Regards,

 ** **

 Mark Miller

 ___
 Mailing list: https://launchpad.net/~openstack
 Post to : openstack@lists.launchpad.net
 Unsubscribe : https://launchpad.net/~openstack
 More help   : https://help.launchpad.net/ListHelp


___
Mailing list: https://launchpad.net/~openstack
Post to : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help   : https://help.launchpad.net/ListHelp


Re: [Openstack] Authorization Question

2013-03-04 Thread Miller, Mark M (EB SW Cloud - RD - Corvallis)
Dolph,

At this point in time I am still gaining a grasp of the Keystone V3 changes and 
how domains and groups can be used. I noticed that the service catalog is 
returned in the GET token response and also in the PKI token when a user 
obtains a scoped token. The catalog data could be a large amount  of extra data 
to pass around in the PKI token and was wondering why it was included. I 
thought maybe there was a link between the user's domain and the endpoints 
included in the catalog but did not see any linking information in the database 
or API documentation. You have just clarified what I thought was true.

Thanks again,

Mark Miller

From: Dolph Mathews [mailto:dolph.math...@gmail.com]
Sent: Monday, March 04, 2013 2:03 PM
To: Miller, Mark M (EB SW Cloud - RD - Corvallis)
Cc: openstack@lists.launchpad.net (openstack@lists.launchpad.net); Brownell, 
Jonathan C (Corvallis)
Subject: Re: [Openstack] Authorization Question

That's correct. Right now, all endpoints registered in keystone are returned to 
all users, regardless of whether they actually have any sort of authorization 
on those endpoints.

I suspect we'll be having a planning session at the design summit on this topic 
-- I'd be helpful to better understand your ideal use case in suppressing 
endpoints from the catalog? In Grizzly, users, groups, projects, and domains 
have absolutely no relationship with services and endpoints within keystone. 
That becomes deployment specific when you consider how RBAC is applied 
service-side with policy.json, etc, but those services are not necessarily 
aware of how they appear in the service catalog, nor does keystone interpret 
policy files other than it's own.


-Dolph

On Mon, Mar 4, 2013 at 2:32 PM, Miller, Mark M (EB SW Cloud - RD - Corvallis) 
mark.m.mil...@hp.commailto:mark.m.mil...@hp.com wrote:
Hello,

I have been looking over the Keystone v3 API documentation as well as the 
database table columns. My question concerns endpoint access restrictions. I 
don't see any noticeable way to associate endpoints with domains which means 
that any user can access any endpoint of any domain. Is this correct? The only 
database column that might come into play is the region column of the endpoint 
table.

Regards,

Mark Miller

___
Mailing list: https://launchpad.net/~openstack
Post to : 
openstack@lists.launchpad.netmailto:openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help   : https://help.launchpad.net/ListHelp

___
Mailing list: https://launchpad.net/~openstack
Post to : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help   : https://help.launchpad.net/ListHelp


Re: [Openstack] Authorization Question

2013-03-04 Thread Nathanael Burton
Dolph,

In our deployments we often want to restrict projects to particular
endpoints or regions.  We've currently hacked that in to our Folsom systems
by adding a 'regions' list to the 'extra' column of the tenant table.  With
only a few minor tweaks to keystone to return the filtered service catalog
based on 'regions' and some minor tweaks to Horizon it all works out fairly
well.  We would prefer to do this in a much more supported configuration,
something like attributes or roles on projects to achieve the same result.

Thanks,

Nate


On Mon, Mar 4, 2013 at 5:02 PM, Dolph Mathews dolph.math...@gmail.comwrote:

 That's correct. Right now, all endpoints registered in keystone are
 returned to all users, regardless of whether they actually have any sort of
 authorization on those endpoints.

 I suspect we'll be having a planning session at the design summit on this
 topic -- I'd be helpful to better understand your ideal use case in
 suppressing endpoints from the catalog? In Grizzly, users, groups,
 projects, and domains have absolutely no relationship with services and
 endpoints within keystone. That becomes deployment specific when you
 consider how RBAC is applied service-side with policy.json, etc, but those
 services are not necessarily aware of how they appear in the service
 catalog, nor does keystone interpret policy files other than it's own.


 -Dolph


 On Mon, Mar 4, 2013 at 2:32 PM, Miller, Mark M (EB SW Cloud - RD -
 Corvallis) mark.m.mil...@hp.com wrote:

  Hello,

 ** **

 I have been looking over the Keystone v3 API documentation as well as the
 database table columns. My question concerns endpoint access restrictions.
 I don’t see any noticeable way to associate endpoints with domains which
 means that any user can access any endpoint of any domain. Is this correct?
 The only database column that might come into play is the region column of
 the endpoint table.

 ** **

 Regards,

 ** **

 Mark Miller

 ___
 Mailing list: https://launchpad.net/~openstack
 Post to : openstack@lists.launchpad.net
 Unsubscribe : https://launchpad.net/~openstack
 More help   : https://help.launchpad.net/ListHelp



 ___
 Mailing list: https://launchpad.net/~openstack
 Post to : openstack@lists.launchpad.net
 Unsubscribe : https://launchpad.net/~openstack
 More help   : https://help.launchpad.net/ListHelp


___
Mailing list: https://launchpad.net/~openstack
Post to : openstack@lists.launchpad.net
Unsubscribe : https://launchpad.net/~openstack
More help   : https://help.launchpad.net/ListHelp