[Openstack-operators] [openstack-dev] [sig][upgrades][ansible][charms][tripleo][kolla][airship] reboot or poweroff?
Adam Spiers
aspiers at suse.com
Fri Aug 3 15:41:57 UTC 2018
[Adding openstack-sigs list too; apologies for the extreme
cross-posting, but I think in this case the discussion deserves wide
visibility. Happy to be corrected if there's a better way to handle
this.]
Hi James,
James Page <james.page at canonical.com> wrote:
>Hi All
>
>tl;dr we (the original founders) have not managed to invest the time to get
>the Upgrades SIG booted - time to hit reboot or time to poweroff?
TL;DR response: reboot, absolutely no question! My full response is
below.
>Since Vancouver, two of the original SIG chairs have stepped down leaving
>me in the hot seat with minimal participation from either deployment
>projects or operators in the IRC meetings. In addition I've only been able
>to make every 3rd IRC meeting, so they have generally not being happening.
>
>I think the current timing is not good for a lot of folk so finding a
>better slot is probably a must-have if the SIG is going to continue - and
>maybe moving to a monthly or bi-weekly schedule rather than the weekly slot
>we have now.
>
>In addition I need some willing folk to help with leadership in the SIG.
>If you have an interest and would like to help please let me know!
>
>I'd also like to better engage with all deployment projects - upgrades is
>something that deployment tools should be looking to encapsulate as
>features, so it would be good to get deployment projects engaged in the SIG
>with nominated representatives.
>
>Based on the attendance in upgrades sessions in Vancouver and
>developer/operator appetite to discuss all things upgrade at said sessions
>I'm assuming that there is still interest in having a SIG for Upgrades but
>I may be wrong!
>
>Thoughts?
As a SIG leader in a similar position (albeit with one other very
helpful person on board), let me throw my £0.02 in ...
With both upgrades and self-healing I think there is a big disparity
between supply (developers with time to work on the functionality) and
demand (operators who need the functionality). And perhaps also the
high demand leads to a lot of developers being interested in the topic
whilst not having much spare time to help out. That is probably why
we both see high attendance at the summit / PTG events but relatively
little activity in between.
I also freely admit that the inevitable conflicts with downstream
requirements mean that I have struggled to find time to be as
proactive with driving momentum as I had wanted, although I'm hoping
to pick this up again over the next weeks leading up to the PTG. It
sounds like maybe you have encountered similar challenges.
That said, I strongly believe that both of these SIGs offer a *lot* of
value, and even if we aren't yet seeing the level of online activity
that we would like, I think it's really important that they both
continue. If for no other reasons, the offline sessions at the
summits and PTGs are hugely useful for helping converge the community
on common approaches, and the associated repositories / wikis serve as
a great focal point too.
Regarding online collaboration, yes, building momentum for IRC
meetings is tough, especially with the timezone challenges. Maybe a
monthly cadence is a reasonable starting point, or twice a month in
alternating timezones - but maybe with both meetings within ~24 hours
of each other, to reduce accidental creation of geographic silos.
Another possibility would be to offer "open clinic" office hours, like
the TC and other projects have done. If the TC or anyone else has
established best practices in this space, it'd be great to hear them.
Either way, I sincerely hope that you decide to continue with the SIG,
and that other people step up to help out. These things don't develop
overnight but it is a tremendously worthwhile initiative; after all,
everyone needs to upgrade OpenStack. Keep the faith! ;-)
Cheers,
Adam
More information about the OpenStack-operators
mailing list