[openstack-dev] [TripleO] gnocchi backport exception for stable/mitaka
James Slagle
james.slagle at gmail.com
Tue Mar 29 21:42:21 UTC 2016
On Tue, Mar 29, 2016 at 12:27 PM, Steven Hardy <shardy at redhat.com> wrote:
> On Tue, Mar 29, 2016 at 11:05:00AM -0400, Pradeep Kilambi wrote:
>> Hi Everyone:
>>
>> As Mitaka branch was cut yesterday, I would like to request a backport
>> exception to get gnocchi patches[1][2][3] into stable/mitaka. It
>> should low risk feature as we decided not to set ceilometer to use
>> gnocchi by default. So ceilometer would work as is and gnocchi is
>> deployed along side as a new service but not used out of the box. So
>> this should make upgrades pretty much a non issues as things should
>> work exactly like before. If someone want to use gnocchi backend, they
>> can add an env template file to override the backend. In Netwon, we'll
>> flip the switch to make gnocchi the default backend.
>>
>> If we can please vote to agree to get this in as an exception it would
>> be super useful.
>
> +1, provided we're able to confirm this plays nicely wrt upgrades I think
> we should allow this.
>
> We're taking a much stricter stance re backports for stable/mitaka, but I
> think this is justified for the following reasons:
>
> - The patches have been posted in plenty of time, but have suffered from a
> lack of reviews and a lot of issues getting CI passing, were it not for
> those issues this should really have landed by now.
>
> - The Ceilometer community have been moving towards replacing the database
> dispatcher with gnocchi since kilo, and it should provide us with a
> (better performing) alternative the current setup AIUI.
>
> Thus I think this is a case where an exception is probably justified, but
> to be clear I'm generally opposed to granting exceptions for mitaka beyond
> the few things we may discover in the next few days prior to the
> coordinated release (in Newton I hope we can formalize this to be more
> aligned with the normal feature-freeze and RC process).
+1.
--
-- James Slagle
--
More information about the OpenStack-dev
mailing list