Firefly EOL date - still Jan 2016?

Firefly EOL date - still Jan 2016?

[Ken Dreyer]

Hi folks,

This is mainly directed at the stable release team members
(http://tracker.ceph.com/projects/ceph-releases/wiki/HOWTO), since
they are the ones doing the work of backporting :)

On http://docs.ceph.com/docs/master/releases/, it says the estimated
EOL for Firefly is Jan 2016, which is coming up soon.

Does anyone on the stable release team have an interest in doing
releases beyond that date, or should we announce that as a firm date?

- Ken

Re: Firefly EOL date - still Jan 2016?

[Loic Dachary]

[-- Attachment #1: Type: text/plain, Size: 2518 bytes --]

Hi Ken,

On 13/11/2015 22:15, Ken Dreyer wrote:
> Hi folks,
> 
> This is mainly directed at the stable release team members
> (http://tracker.ceph.com/projects/ceph-releases/wiki/HOWTO), since
> they are the ones doing the work of backporting :)
> 
> On http://docs.ceph.com/docs/master/releases/, it says the estimated
> EOL for Firefly is Jan 2016, which is coming up soon.
> 
> Does anyone on the stable release team have an interest in doing
> releases beyond that date, or should we announce that as a firm date?

Although we're heading in this direction, publishing a point release still depends on resources that are not limited to the stable release team. Ideally, in the not too far future, someone from the stable release team could answer "I'll keep publishing releases" and really be able to do it all by her/himself.

As of today, running the integration tests can partially be done using an OpenStack tenant and no access to the sepia lab. But some of them (rgw in particular) still need work. There is no need for the gitbuilders because the teuthology-openstack creates the necessary package on demand, but corner cases were fixed this week and there probably are a few others.

The other blocker is the release process which still require significant manual intervention, privileged access and undocumented knowledge. This is improving quickly but we're not yet at a stage where an unprivileged third party is able to run it independently.

I'm under the impression that firefly will indeed be EOL early next year because:

 * it is currently impractical for community members to test and publish a release
 * the backport activity gradually slowed down as users migrate to hammer

I hope that my answer will be different by the time Jewel retires :-) The larger goal is to establish a light weight development process that requires as little resources as possible. Let say adding a patch to a stable release, running all the tests and publishing the release can be done within a week real time, a few hours of work and less than 1,000 USD to pay for the cloud resources. IMHO it is unlikely to expect stable releases from a community of developers if the effort/cost is significantly higher.

Cheers
> 
> - Ken
> --
> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

-- 
Loïc Dachary, Artisan Logiciel Libre


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: Firefly EOL date - still Jan 2016?

[Nathan Cutler]

> Does anyone on the stable release team have an interest in doing
> releases beyond that date, or should we announce that as a firm date?

For now my vote is to stick to the schedule and declare EOL on January
31, but I'm willing to negotiate :-)

Nathan