hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Andrew Purtell <apurt...@apache.org>
Subject Re: Minor release cadence for branch-1
Date Fri, 06 Mar 2015 22:27:53 GMT
I think we can manage. The nice thing about on demand minors is what
emerges will be driven by actual need. For example after we move up from
0.98 to 1.x, I might want to RM a minor line that corresponds to our
production. Someone working for a vendor might be RMing a minor line
corresponding to the latest release. The number of active lines will be
bounded by the level of available interest and effort in maintaining them.


On Thu, Mar 5, 2015 at 5:04 PM, Enis Söztutar <enis.soz@gmail.com> wrote:

> +1 doing patch releases more frequently, and doing minor releases on a
> per-need basis and if an RM volunteers to drive a minor version.
>
> Every minor release we do will come up with its own branch, and we might
> end up with a lot of active branches to accept bug fixes which might put
> some burden on the committers to check in to so many branches.
>
> Enis
>
> On Thu, Mar 5, 2015 at 1:04 PM, lars hofhansl <larsh@apache.org> wrote:
>
> > I'd say as long as there are folks who contribute patches we can do patch
> > releases for older minor releases. It's open source :)
> >
> > Keeping *all* minor releases alive seem onerous (there could be dozens -
> > unless we significantly drive down the time between major releases).Could
> > do the Linux model and support some minor versions a bit longer than
> others.
> > Putting myself in Salesforce's (where I work) shoes... We'd move from
> > minor or patch release to the next at our own pace. On a case-by-case
> basis
> > we'd decide whether to deploy a patch release, wait a bit, or move to the
> > next minor release.HBase is nicely managed in terms of stability and
> > compatability, so as along as minor releases are rolling upgradable (with
> > some versions skipped - i.e. we could go from (say) 1.1.2 to 1.2.7) we'd
> > likely be following the minor versions mostly.
> > We would *very* rarely switch major-version as client-server
> > incompatibilities are very hard to handle for us.I don't think as far as
> > big shops go we're very special...?
> >
> > Maybe we'd have to play this by ear... Start making new minor versions,
> > and see how much work it is to maintain the older ones and what our users
> > end up doing (staying on a minor version or adopting new minor versions).
> >
> > This is a very interesting topic.
> >
> > -- Lars
> >
> >       From: Sean Busbey <busbey@cloudera.com>
> >  To: dev <dev@hbase.apache.org>
> > Cc: lars hofhansl <larsh@apache.org>
> >  Sent: Thursday, March 5, 2015 12:48 PM
> >  Subject: Re: Minor release cadence for branch-1
> >
> > It sounds to me like we have consensus around monthly for patch releases
> > and minor releases on demand, provided we can find RMs.
> >
> > Would it be reasonable to keep all the minor release lines active until
> we
> > have a newer major release? At that point we could keep just the most
> > recent minor release going so long as there's demand.
> >
> > --
> > Sean
> >
> >
> > On Mar 5, 2015 2:23 PM, "lars hofhansl" <larsh@apache.org> wrote:
> >
> > > Any further comments on this?
> > > Seems important to get agreement at least generally.
> > > -- Lars
> > >      From: lars hofhansl <larsh@apache.org>
> > >  To: "dev@hbase.apache.org" <dev@hbase.apache.org>
> > >  Sent: Monday, March 2, 2015 10:26 PM
> > >  Subject: Re: Minor release cadence for branch-1
> > >
> > > Hmmm...
> > >
> > > I had only expect a monthly patch cadence for minor release (btw, we
> > > started monthly releases with 0.94.x).
> > >
> > > In 0.94 and 0.98 we had no clear distinction between patch and minor
> > > releases.
> > >
> > > For minor releases it seems an on-demand model is more what we want.
> I.e.
> > > we'd have a monthly 1.0.1, 1.0.2, etc. Then at some point we'd release
> > > 1.1.0... "when it's ready".
> > > Since that's a minor upgrade we can then have a few more 1.0.x releases
> > > (like 0.94 is now) and then tell folks to upgrade to 1.1.x.
> > > (in the end, though, patch releases should continue as long as folks
> are
> > > willing to contribute patches)
> > >
> > > I'd be happy to sign up to do a few minor (1.1, 1.2, or whatever)
> > releases
> > > - but I do think we should share the love not have the same folks do
> > > multiple releases simulataneously.
> > >
> > > -- Lars
> > >
> > >
> > >      From: Sean Busbey <busbey@cloudera.com>
> > >
> > >
> > >  To: dev <dev@hbase.apache.org>
> > >  Sent: Friday, February 27, 2015 10:06 AM
> > >  Subject: Minor release cadence for branch-1
> > >
> > > Hey folks!
> > >
> > > Apologies if I've overlooked this getting discussed already. Do we
> have a
> > > goal release cadence for minor versions out of branch-1?
> > >
> > > My first gut reaction is that it should essentially match the cadence
> > we've
> > > been aiming at for the 0.98 line. That would mean attempting to match
> > > monthly, I think?
> > >
> > > The obvious problem with this is that now that we have patch versions,
> it
> > > means essentially getting a new branch per month for backports. That's
> > > quickly going to get old, even if we presume most additions will move
> > onto
> > > branch-2 in a year or so.
> > >
> > > What do folks think about limiting which minor versions patch-level
> fixes
> > > go into? We could default to the most recent release + current minor
> dev
> > > and go back farther when requested by the issue filer?
> > >
> > > That means in ~3 months we'd expect branch-1 to be working on 1.4 and
> > most
> > > patch-level fixes to go into branch-1.3 and branch-1. If someone
> > reported a
> > > failure and they were on e.g. 1.1.z, we'd also do the fix in branch-1.1
> > and
> > > branch-1.2.
> > >
> > > Or should we just stick with hitting all of the branches on the
> > presumption
> > > that the cherry picks should be trivial?
> > >
> > > --
> > > Sean
> > >
> > >
> > >
> > >
> > >
> >
> >
> >
> >
>



-- 
Best regards,

   - Andy

Problems worthy of attack prove their worth by hitting back. - Piet Hein
(via Tom White)

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message