Open Stack

Hi all,

I want to recap the Boston Forum session "Skip-level upgrading - jumping 
ahead to catch up" which I and Andy co-moderated.

We had a good number of attendees from both operators and developers in 
the room and had positive discussion.
We agreed to bring this topic on dev ML for further discussion, so all 
feedback, comment are appreciated.

Here are some of the points from the session.
See etherpad [1] for details.

- Most of the ops in the room who did OpenStack upgrade did N-m to N 
(where m>=2, N:latest release)
- Variations of skip-level upgrade experiences was shared and the 
requirements for community support of skip-level upgrade was pointed out.
- There were some level of consensus among the Ops in the room that API 
downtime during the maintenance window is acceptable. Having instances 
running during the upgrade is critical for operators.
- There were proposal from Dev that they can try to have better 
description on N-2 to N upgrade impact in release notes (per project).
- All agreed to bring this proposal to dev-ML for further discussion.

[1] https://etherpad.openstack.org/p/BOS-forum-skip-level-upgrading

Regards,
Shintaro

-- 
Shintaro MIZUNO
NTT Software Innovation Center
TEL: 0422-59-4977
E-mail: mizuno.shintaro at lab.ntt.co.jp
         shintaro1018 at gmail.com