[QA] [Wikitech-l] #Beta-Cluster is now #Beta-Cluster-infrastructure and now there's #Beta-Cluster-reproducible
Greg Grossmeier
greg at wikimedia.org
Tue Oct 6 17:07:00 UTC 2015
<quote name="Jamison Lofthouse" date="2015-10-06" time="09:09:23 -0600">
> >
> > > How about "#unbreak-before-it-hits-production"?
>
>
> Or even better #unbreak-pre-production.
Not bad.
My only concern:
There might be tasks that are filed in #whatever-we-call-this that are
of this type: Valid issue, but not one worthy of blocking deployment.
So, we could still have the #beta-cluster-reproducible but use that
fancy "UBN!" priority and any task in that project with UBN! as
priority.... needs to be unbroken now(!).
Critique of my critique:
Are tasks that are a "valid issue, but not worthy of blocking
deployment" really needing to be in #beta-cluster-reproducible?
History:
We used to have bug 38865 "Next wmf deployment (tracking)" which was
used to track bugs that needed to be fixed before the next deployment.
https://phabricator.wikimedia.org/T40865
We closed that blocking task and (I) created #WMF-Server-Backports. The
point of this project was to collect tasks that are ready to be
backported to production. This project is now mostly moot with our new
world that has active SWAT deploys.
https://wikitech.wikimedia.org/wiki/SWAT_deploys
Unfortunately, we can't just trust the UBN! priority across all
projects:
https://phabricator.wikimedia.org/maniphest/query/E3qGuXy8ug.S/#R
There's tasks there that haven't been updated in over a month. And no
one wants to maintain a list of "projects where UBN! actually means UBN!
wrt to WMF production". Also, people, even within WMF engineering teams,
don't like it when you tell them they're using a status or task
incorrectly. Bazaar vs Cathedral and all.
I have no conclusion or summary, yet,
Greg
--
| Greg Grossmeier GPG: B2FA 27B1 F7EB D327 6B8E |
| identi.ca: @greg A18D 1138 8E47 FAC8 1C7D |
More information about the QA
mailing list