[openstack-dev] [infra] Gerrit downtime on May 23 for project renames

Tom Fifield tom at openstack.org
Thu May 22 04:39:22 UTC 2014


On 22/05/14 11:06, Kyle Mestery wrote:
> On Wed, May 21, 2014 at 5:06 PM, Tom Fifield <tom at openstack.org> wrote:
>> On 22/05/14 05:48, James E. Blair wrote:
>>>
>>> Tom Fifield <tom at openstack.org> writes:
>>>
>>>> May I ask, will the old names have some kind of redirect to the new
>>>> names?
>>>
>>>
>>> Of course you may ask!  And it's a great question!  But sadly the answer
>>> is "no".  Unfortunately, Gerrit's support for renaming projects is not
>>> very good (which is why we need to take downtime to do it).
>>>
>>> I'm personally quite fond of stable URLs.  However, these started as an
>>> "experiment" so we were bound to get some things wrong (and will
>>> probably continue to do so) and it's better to try to fix them early.
>>
>>
>> This is a really poor outcome.
>>
>> Can we delay the migration until we have some time to think about the
>> communication strategy?
>>
>> At the least, I'd suggest a delay for renaming neutron-specs until until
>> after the peak of the Juno blueprint work is done. Say in ~3 weeks time.
>>
> I tend to agree with James that we should do this early and take the
> bullet on renaming now. The process for adding new Neutron specs is
> outlined here [1], and this will be updated once the repository is
> renamed. In addition, I'm working on adding/updating some Neutron wiki
> pages around the Neutron development process, and the specs repo will
> be highlighted there once that's done. It would be good to have the
> renaming done before then.

... and how would you propose we communicate this to the users we've 
been asking to do blueprint review specifically during this early 
period? We can't exactly send them an email saying "sorry, the link we 
mentioned earlier is now wrong" :)

What would you gain from doing it this week instead of later in the month?

We're really trying to engage users to help out with the spec review 
process, but it seems they weren't taken into account at all when 
planning this change. Seems like a bad precedent to set for our first 
"experiment".


Regards,


Tom



More information about the OpenStack-dev mailing list