Completely agree, and I think there are many people that know its an
issue, which is positive news.

There are always people who 'pave' over there clouds instead, which to me
is not positive news;

In fact I think such activities are detrimental to the whole community
(your opinion may vary).

To me this is why openstack **must** provide a reference cloud and do as
many of these rigorous activities "checking" automatically.

-Josh

On 10/24/13 3:03 PM, "Robert Collins" <[email protected]> wrote:

>On 25 October 2013 10:20, Shane Johnson <[email protected]>
>wrote:
>
>>
>>
>> But doesn't that responsibility lie with the package maintainers for
>> Openstack in each distro?
>
>No. The responsibility for making it possible to upgrade in place
>*reliably* and *with confidence* is an entirely upstream problem.
>Rigorous backwards compatibility, rigorous testing for performance
>regressions, reliable upgrade paths, being able to achieve low/no
>downtime on the cloud during upgrade. All upstream problems.
>
>Putting the code into packages and making the packages install well is
>a distribution problem. For distributions with orchestration code tied
>into them, orchestrating the upgrade is also a distribution problem
>(e.g. RDO and UO both offer fully orchestrated deployments, so any
>upgrade orchestration will be bundled there).
>
>-Rob
>
>
>-- 
>Robert Collins <[email protected]>
>Distinguished Technologist
>HP Converged Cloud
>
>_______________________________________________
>Mailing list: 
>http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
>Post to     : [email protected]
>Unsubscribe : 
>http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack


_______________________________________________
Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
Post to     : [email protected]
Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack

Reply via email to