How to move storage OEMs to Samba 4.0 ?

Jeremy Allison jra at samba.org
Wed Jun 6 12:12:41 MDT 2012


On Wed, Jun 06, 2012 at 02:08:03PM -0400, Ira Cooper wrote:
> On Wed, Jun 6, 2012 at 12:59 PM, Jeremy Allison <jra at samba.org> wrote:
> > On Wed, Jun 06, 2012 at 06:52:17PM +0200, Kai Blin wrote:
> >> On 2012-06-06 18:02, Jeremy Allison wrote:
> >>
> >> > Actually, we've been good and getting better. OEMs have come to
> >> > really trust our stability within a release stream. It's moving
> >> > to a new release stream they *really* don't like.
> >>
> >> So aren't we risking that exact trust by putting a lot of new features
> >> into the stable release stream?
> >
> > Define "a lot" :-). I'm not proposing a lot, just the ones
> > we have already running at OEMs anyway. They're already
> > in production, we should give these things a home.
> 
> To debunk some things in this thread:
> 
> 1. Not all bug fixes are safe.  Sometimes your "bug fix" will blow up
> someone else.
> 
> 2. Not all features are unsafe.  If you never enable them.  They might
> as well not exist, in many cases, especially VFS changes.
> 
> So this is not an absolute issue.  Not even close.
> 
> I'm going to propose an in-between answer.  We allow back-porting of
> new functionality into 3.6.  But it should be under stricter review,
> to help with the QA issues.
> 
> I'd prefer 3 sets of eyes in addition to the original developer, from
> 2+ companies.  IE: Volker has a proposed change, he needs someone
> outside SerNET to sign off, the other 2 can be from SerNET.
> 
> In addition to more review, clearly this raises the level of effort
> needed to backport something.  Which should stop "frivolous"
> backports, and will provide a long term incentive to move forward to
> 4.0 and higher.

+1 on this. I really like this idea.

Thanks Ira.

Jeremy.


More information about the samba-technical mailing list