On 03/07/2016 05:22 PM, Michał Jastrzębski wrote:
Upgrades are required, true, but not necessarily automatic ones.
People still can rebuild and redeploy containers using normal deploy.
It will be downtime causing and less optimal, but possible. Also with
backport of named volumes it won't be data-destroying. It will cause
total downtime of APIs, but well, it's first version.
So I'm -1 to porting it to 1.1.0.
But I would suggest another option, namely 1.2.0 with automatic
upgrades we have now. It will allow upgrade 1.1.0->1.2.0 and it will
not add more work to 1.1.0 which we need asap (we need it well tested
by Austin summt AT MOST). Adding upgrades might make it tight,
especially that infra upgrades aren't finished yet in master.
Cheers,
Michal
I like this idea.
I'm OK with the idea of backporting upgrade playbooks to stable/liberty
in general. But I think that the other Michal made a valid point that
1.1.0 has to be well tested soon and in the same moment we don't have
all upgrades yet even in master.
So,
-1 for backporting to 1.1.0
+1 for backporting later to 1.2.0
__________________________________________________________________________
OpenStack Development Mailing List (not for usage questions)
Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
