Date: Tue, 17 Jan 2012 18:57:19 +0000 From: Hugo Silva <hugo@barafranca.com> To: Ivan Voras <ivoras@freebsd.org> Cc: Tom Evans <tevans.uk@googlemail.com>, freebsd-hackers@freebsd.org Subject: Re: FreeBSD has serious problems with focus, longevity, and lifecycle Message-ID: <4F15C48F.7020302@barafranca.com> In-Reply-To: <CAF-QHFV8oj=ipwcsVo3e3P3kgGBPr%2Bz1gRzn3D3PT%2Bc0pHJtcQ@mail.gmail.com> References: <alpine.BSF.2.00.1112211415580.19710@kozubik.com> <jf3mps$is3$1@dough.gmane.org> <CAFHbX1%2Bi3JwCCBmqtOsW6m74VpDBSAmBOt7CPcCGAPCO2DBDkA@mail.gmail.com> <CAF-QHFV8oj=ipwcsVo3e3P3kgGBPr%2Bz1gRzn3D3PT%2Bc0pHJtcQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 01/17/12 12:42, Ivan Voras wrote:
> On 17 January 2012 13:02, Tom Evans<tevans.uk@googlemail.com> wrote:
> Almost certainly yes. The current release process involves src, ports
> and docs teams. Would you and other RELEASE users be happy with simple
> periodic snapshots off the STABLE branches, not much different from
> tracking STABLE? The only benefit I see would be a light-weight
> opportunity for testing which would probably end up being implemented
> by moving to date-based tags (e.g. if a critical bug is found and the
> fix MFC-ed, the "current" tag would be advanced to "$today")?
If (a cluster of) new features have made it to stable in the mean time
and have been tested and generally recognized as being in working
condition, my opinion is that a release should be up for consideration:
For starters, it's so much easier from a server management point of
view.. easier to memorize and compare 8.5-RELEASE vs 8.6-RELEASE ("now
with blah-ZFS features and cpuset added to rc.d, bind updated to blah,
openssh updated to version bleh, insert other usual suspects here") than
it is to track stable. I don't remember what was added to -STABLE 3
months ago. It doesn't appear to be as much of a good starting point
("good snapshot"), compared to a release either. I like to know where
I'm standing when setting up a new system, and a release is a convenient
starting point.
Also, one can always check what was added to a release on freebsd.org,
while for -stable some digging has to be performed.
For production servers, I'd rather run most servers with known good
release branches.
Sure, there were times when I needed something from the stable branch
and decided to use it, but when the next release comes out then it's
time to revert back to the release branch.
Come to think about it, those days are pretty much gone since 4.x
(incidentally, many of us who've stuck with FreeBSD for this long think
of 4.x as an epic series).
Running different -stable snapshots from different times results in
systems running different versions of the operating system, in my
opinion this is bound to bring problems (more stuff can go wrong).
It's so much easier to have a look at ganglia or cacti or clusterit or
anything similar and contrast the OS version and architecture. Running
-stable, not so much. Is it stable from 2 days ago or a year? And what
has changed since then?
Furthermore, releases get way more attention in freebsd's own webpage,
not to mention the cascading effect that has. You don't see a lot of
articles about new blah features on "9-stable snapshot dated Jan 17
2012", but there are many about 9.0-RELEASE being out in the street.
This also brings more visibility to the project.
Maybe I'm horribly mistaken about the releasographics of production
FreeBSD users, but I think most of us tend to run -release for the
reasons mentioned above, and perhaps some more that I'm neglecting right
now.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4F15C48F.7020302>