[openstack-dev] [openstack-announce] End of life for managed stable/icehouse branches

Ihar Hrachyshka ihrachys at redhat.com
Wed Jul 15 10:37:59 UTC 2015


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

On 07/14/2015 09:14 PM, Thomas Goirand wrote:
> On 07/14/2015 10:29 AM, Ihar Hrachyshka wrote:
>> On 07/14/2015 12:33 AM, Thomas Goirand wrote:
>>> I missed this announce...
>> 
>>> On 07/02/2015 05:32 AM, Jeremy Stanley wrote:
>>>> Per the Icehouse EOL discussion[1] last month, now that the 
>>>> final 2014.1.5 release[2] is behind us I have followed our
>>>> usual end of life steps for stable/icehouse branches on repos
>>>> under the control of the OpenStack Release Cycle Management
>>>> project-team. Specifically, for any repos with the
>>>> release:managed[3] tag, icehouse-specific test jobs were
>>>> removed from our CI system and all open change reviews were
>>>> abandoned for stable/icehouse. Then the final states of the
>>>> branches were tagged as "icehouse-eol" and the branches
>>>> subsequently deleted.
>> 
>>> I believe I asked you about 10 times to keep these branches
>>> alive, so that distributions could work together on a longer
>>> support, even without a CI behind it.
>> 
>>> I have also asked for a private gerrit for maintaining the 
>>> Icehouse patches after EOL.
>> 
>>> While I understand the later means some significant work, I
>>> don't understand why you have deleted the Icehouse branches.
>> 
>>> Effectively, under these conditions, I am giving up doing any
>>> kind of coordination between distros for security patches of
>>> Icehouse. :(
>> 
>> As far as I know, there was no real coordination on those
>> patches before, neither I saw any real steps from any side to get
>> it up.
> 
> Well... as far as I know, you were not there during the
> conversations we had at the summits about this. Neither you are on
> my list of Icehouse security persons. So I fail to see how you
> could be in the loop for this indeed.
> 

Indeed, in Openstack, people work in public, and publish details about
their (private?) talks on summits on the mailing list. This is the
place where decisions are made, not summits, and it's a pity that some
people see chats on summits as something defining the future.

If you don't think I (a member of stable-maint-core) should have been
in the loop, fine for me. Just don't complain when branches are
dropped. Note: infra got explicit approval from the team to drop those
branches. It could be avoided if 1) you were participating in stable
effort; 2) there were public discussion on the mailing list and
resolution that we want to extend branch life after end-of-CI.

>> That said, anyone can come up with an initiative to maintain
>> those branches under some 3party roof (just push -eol tag into
>> github and advertise), and if there is real (and not just
>> anticipated) collaboration going on around it, then the project
>> may reconsider getting it back in the big stadium.
> 
> I have a list of contacts for each and every downstream
> distributions.

Whom have you contacted on RDO side? Just curious.

> All of them agreed to work under this coordinated Git repo, so that
> we share the same patch. The only issue is that during embargo
> period, we can't discuss this type of patches in public. Which is
> why a private gerrit was the way to go. Though for not-embargoed
> stuff, we could well have used the already existing Gerrit
> infrastructure, without a CI (as all distro are running their own
> tests anyway).
> 

I am not sure RDO would be interested in consuming pieces of unclear
quality (no CI) thru rebase only to realize that half of those are not
valid. I would not dare to lower quality of 'after-eol' releases of
RDO by rebasing on top of unvalidated patches.

>> I am tired to say that again and again, but there should be some 
>> resource investment from interested parties, upfront, before
>> infra takes part of the burden on their shoulders. Asking won't
>> help.
> 
> How do you expect to see anything happening before Icehouse
> effectively gets EOL? By the way, I haven't asked anything but
> *not* doing something. I don't see how much "burden" I'm putting on
> infra here.
> 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJVpjgEAAoJEC5aWaUY1u57QmEIAIhJqqzzOK3K/iFuADHSlAu5
ydnwkMK5fnCIcuB7UesmB+SCg84/abYcumP08/AR0lPWYfbEfV/4WF821bLTmmg0
8JyjGyAARF/dxmkC1/Wf3P7O0g+lJAYU6dbEOwDKUX614oUHJCvXW1m9aSy5VVJ2
G8LmWgD6W34a6Kr+3wwHVc4FTSg+Jto4l8cbBqKnJQrhy9+8aXJAqlYHBRm4gNO6
xd2zlYHfPpXR6NTux4qRRYSZpjCjceBjfgEhaurDtuLUFy0AIS0UT0i+WJCuXZ29
dyouEDXhuDEvd2L78rz9XEWNJIwOd9XMyabC9oMNGh2bymo61WktQXid4hLTfuo=
=u3Vn
-----END PGP SIGNATURE-----



More information about the OpenStack-dev mailing list