[Openstack-operators] [keystone] Catalog Backend in Deployments (Templated, SQL, etc) [x-posted from dev-list]

Jesse Pretorius jesse.pretorius at gmail.com
Wed Apr 23 07:58:36 UTC 2014

On 22 April 2014 21:39, Morgan Fainberg <m at metacloud.com> wrote:

> During the weekly Keystone meeting, the topic of improving the Catalog was
> brought up. This topic is in the context of preparing for the design summit
> session on the Service Catalog. There are currently limitations in the
> templated catalog that do not exist in the SQL backed catalog. In an effort
> to provide the best support for the catalog going forward, the Keystone
> team would like to get feedback on the use of the various catalog backends.
> What we are looking for:
>    1. In your OpenStack deployments, which catalog backend are you using?
>    2. Which Keystone API version are you using?
> Interesting - as a deployer I found myself surprised that there were
options. I have a vague awareness that options were being worked on, but
not that they were actively available.

I tried to do a little research to understand the options mentioned, but I
can't seem to find anything to describe what they are, what advantages
there are for using one over another and how one would deploy any of them.
The only information available seems to be the code itself.

The developer documentation only lists service back-ends. [1]
The configuration reference [2] shows many back-end drivers in use, but
there's no description of options.
The install guides presume default back-end configuration.
The admin guide [3] mentions other back-ends available but doesn't describe

While I see mention made of the SQL back-end, which appears to be the
default, and the template back-end (for some services) I see nothing about
Dogpile except in the keystone developer configuration reference [4].

Generally speaking as a deployer I try not to use undocumented features as
doing so ends up causing pain in the form of either the feature
disappearing, or in the form of having to try to educate someone else
rather than rely on them using the standard resources (documentation,
google) to understand them. It becomes a technical debt.

Considering that the SQL back-end is the default Catalog Back-end, it's
safe to assume that we're using that. We're also still using the Keystone
v2 API as we're still looking to see more v3 API support from the other
projects and access to use them from more friendly interfaces (CLI,
Horizon) so that our administrators don't need to do curl queries on a
day-to-day basis and our customers can simply use the stock tools provided.
I've not evaluated Icehouse yet in this regard, but Havana still had a
little way to go.

[4] http://docs.openstack.org/developer/keystone/configuration.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20140423/bad91a14/attachment.html>

More information about the OpenStack-operators mailing list