[openstack-dev] [tripleo] TripleO Ocata release blockers

Ben Nemec openstack at nemebean.com
Tue Feb 14 19:14:52 UTC 2017

On 02/13/2017 12:52 PM, Ben Nemec wrote:
> On 02/12/2017 02:24 PM, Emilien Macchi wrote:
>> Quick updates:
>> - All FFEs patches have been merged except:
>> https://review.openstack.org/#/c/330050/ - so we're green on this
>> side.
>> - Upgrade team is still working on automation to upgrade from newton
>> to ocata.
>> - ovb-updates (CI job with IPv6) doesn't pass -
>> https://bugs.launchpad.net/tripleo/+bug/1663187 - See
>> https://review.openstack.org/#/c/432761/ for a potential fix.
>> - our CI is testing OpenStack from trunk, we have regular promotions.
>> At this stage, I believe we can release TripleO Ocata RC1 by Thursday
>> 17th, considering Congress will be landed, 1663187 fixed, no new
>> blocker in CI.
>> I think upgrade folks will still have some patches after Thursday, but
>> I don't think it's a big deal, we'll just backport them to
>> stable/ocata.
>> Please let us know asap if you see more blockers for Ocata RC1.
> I would like to propose https://bugs.launchpad.net/tripleo/+bug/1664331
> as a release blocker, mostly because if someone ran an undercloud
> upgrade without us fixing that I believe it could break their undercloud
> Heat in a way that would be difficult to recover from (should probably
> discuss that with the Heat team though to make sure I'm not wrong about
> the impact).
> If we can't fix it before release, then I think we should at least merge
> the _stackrc_exists fix and essentially block undercloud upgrades until
> we fix the member role logic.  That way we can backport the fix knowing
> that no one has run a bad upgrade.

I take it all back. :-)

After some more investigation into this, I discovered that this is a 
workaround for a problem that dates back to the downstream-only kilo 
release, so it's absence isn't actually a critical problem.  There's a 
simple manual workaround that users would have had to use in the 
intervening releases anyway.

It would still be nice to fix this, but it doesn't need to block the 

More information about the OpenStack-dev mailing list