Open Stack

+1 to simple mode removal

On Tue, Jan 27, 2015 at 4:44 PM, Stanislaw Bogatkin <sbogatkin at mirantis.com>
wrote:

> +1
>
> On Tue, Jan 27, 2015 at 4:05 PM, Aleksandr Didenko <adidenko at mirantis.com>
> wrote:
>
>> Hi,
>>
>> After starting implementing granular deployment we've faced a bunch of
>> issues that would make further development of this feature much more
>> complicated if we have to support both Simple and HA deployment modes. For
>> example: simple mode does not require cluster (corosync, pacemaker, vips,
>> etc), so we had to skip this task for Simple mode somehow - we can use
>> conditional tasks, or conditional manifests in our tasks, or create
>> separate task graphs for different deployment modes, etc - either way it's
>> pretty much doubling the amount of work for some parts of Fuel and our
>> development cycle.
>>
>> At the moment, CI blocks us from further development of fuel-library
>> modularization BP [2] because we still use Simple mode in CI. So in order
>> to proceed with this BP we have two options:
>>
>> 1) remove Simple mode from CI/QA and thus drop it completely from Fuel
>> 2) double our efforts to support both Simple and HA modes in granular
>> deployment
>>
>> We have a BP about single-controller HA [1]. HA with single controller
>> works just fine at the moment. So if you want to test Fuel on a minimum set
>> of nodes, you can do this on 3 nodes (Fuel master, controller, compute),
>> just like with Simple mode before. I suppose, it's time to finally drop
>> support for Simple mode in Fuel :)
>>
>> [1] https://blueprints.launchpad.net/fuel/+spec/single-controller-ha
>> [2]
>> https://blueprints.launchpad.net/fuel/+spec/fuel-library-modularization
>>
>> --
>> Regards,
>> Aleksandr Didenko
>>
>>
>> On Tue, Aug 26, 2014 at 9:25 AM, Mike Scherbakov <
>> mscherbakov at mirantis.com> wrote:
>>
>>> Definitely fuel spec is needed :)
>>>
>>>
>>> On Mon, Aug 25, 2014 at 8:45 PM, Evgeniy L <eli at mirantis.com> wrote:
>>>
>>>> Hi Andrew,
>>>>
>>>> I have some comments regarding to you action items
>>>>
>>>> >> 2) Removing simple mode from the ui and tests
>>>> >> 3) Removing simple mode support from nailgun (maybe we leave it) and
>>>> cli
>>>>
>>>> We shouldn't do it, because nailgun should handle both versions of
>>>> cluster.
>>>> What we have to do here is to use openstack.yaml to keep all possible
>>>> modes.
>>>> For new release there will be only ha, to manage previous releases we
>>>> have
>>>> to create data migrations in nailgun to create the filed with modes
>>>> i.e. multinode
>>>> and ha.
>>>>
>>>> Also fixes for ui are required too, I think it mostly related to
>>>> wizard, 'mode' tab
>>>> where use can chose ha or non ha cluster in case of new release there
>>>> should
>>>> be only ha, and in case of old releases there should be ha and
>>>> multinode.
>>>>
>>>> Thanks,
>>>>
>>>>
>>>>
>>>>  On Mon, Aug 25, 2014 at 8:19 PM, Andrew Woodward <xarses at gmail.com>
>>>> wrote:
>>>>
>>>>>  Started a new thread so that we don't hijack the older thread.
>>>>>  as
>>>>>
>>>>>
>>>>>> Andrew, will you work on it in 6.0? What are remaining items there?
>>>>>> Also, it might affect our tests - simple mode runs faster so we use it for
>>>>>> smoke ISO test. Anastasia, please confirm that we can switch smoke to
>>>>>> one-ha-controller model, or even drop smoke at all and use BVT only
>>>>>> (running CentOS 3 HA controllers and same with Ubuntu).
>>>>>>
>>>>>
>>>>> The primary reason that we haven't disabled single yet is was due to
>>>>> [0] where we where having problems adding additional controllers. With the
>>>>> changes to galera and rabbit clustering it appears that we ended up fixing
>>>>> it already.
>>>>>
>>>>> The remaining issues are:
>>>>> 1) Ensuring we have good test coverage for the cases we expect to
>>>>> support [1]
>>>>> 2) Removing simple mode from the ui and tests
>>>>> 3) Removing simple mode support from nailgun (maybe we leave it) and
>>>>> cli
>>>>> 4) Updating documentation
>>>>>
>>>>> [0] https://bugs.launchpad.net/fuel/+bug/1350266
>>>>> [1] https://bugs.launchpad.net/fuel/+bug/1350266/comments/7
>>>>>
>>>>> --
>>>>> Andrew
>>>>> Mirantis
>>>>> Ceph community
>>>>>
>>>>> _______________________________________________
>>>>> OpenStack-dev mailing list
>>>>> OpenStack-dev at lists.openstack.org
>>>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>>>>
>>>>>
>>>>
>>>> _______________________________________________
>>>> OpenStack-dev mailing list
>>>> OpenStack-dev at lists.openstack.org
>>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>>>
>>>>
>>>
>>>
>>> --
>>> Mike Scherbakov
>>> #mihgen
>>>
>>>
>>> _______________________________________________
>>> OpenStack-dev mailing list
>>> OpenStack-dev at lists.openstack.org
>>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>>
>>>
>>
>> __________________________________________________________________________
>> OpenStack Development Mailing List (not for usage questions)
>> Unsubscribe:
>> OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
>> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>>
>>
>
> __________________________________________________________________________
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
>
>


-- 
Yours Faithfully,
Vladimir Kuklin,
Fuel Library Tech Lead,
Mirantis, Inc.
+7 (495) 640-49-04
+7 (926) 702-39-68
Skype kuklinvv
45bk3, Vorontsovskaya Str.
Moscow, Russia,
www.mirantis.com <http://www.mirantis.ru/>
www.mirantis.ru
vkuklin at mirantis.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20150127/7f6c1846/attachment.html>