freebsd stable - Feb 2019 - Binary update to -STABLE? And if so, what do I get?

On Thu, Feb 14, 2019 at 3:10 AM Pete French <petefrench at ingresso.co.uk>
wrote:
>
>
> On 14/02/2019 01:43, Jason Tubnor wrote:
> > I also have hit this IPv6 issue (I thought I was going crazy until I
> worked
> > it out) and other iflib issues in 12.0, which have been fixed in
-STABLE
> > that really should be patched in 12.0 or bring forward an early 12.1
> > release. For our use case, 12.0 is just too buggy for production at
this
> > rate and we won't touch it, which is a shame because there is a
lot of
> good
> > work in there that we would like to use but it is trumped by the
> breakages.
>
> Any reason behind not running STBLE out of interest ? Yes, 12 has been
> buggy with regards to networking, but these things get fixed very fast
> and I now have all my machines on the lattest STABLE in production, as
> of yesterday.
>
> -pete.
>
Generally, not many.

Far and away the biggest is the requirement to build from sources. It's not
a big deal for me, but if I still had many systems to deal with, that would
be a pain.

I might also mention that just before the pre-release freeze and after a
release, STABLE and be a bit unstable as developers rush to get things in
before the freeze or to add things that they did not want to MFC with
little test time before a release. In recent years this issue has
significantly improved, though.

I run either HEAD or STABLE on my personal system and RELEASE on my only
server, the latter just so I can do binary updates. My server is on hold at
11.2 due to the IPv6 issue and I am amazed that this BIG oops did not
result in an EN and a patched release. Lack of IPv6 is not, for many
people, a minor issue.

The bottom line is that the only real reasons I see for not running stable
is the lack of binary updates, and issues with systems being slightly out
of sync if all are not updated to the same SVN revision at all times. Those
are very big reasons for many.
--
Kevin Oberman, Part time kid herder and retired Network Engineer
E-mail: rkoberman at gmail.com
PGP Fingerprint: D03FB98AFA78E3B78C1694B318AB39EF1B055683

On Thu, Feb 14, 2019 at 10:13 AM Kevin Oberman <rkoberman at gmail.com>
wrote:
> On Thu, Feb 14, 2019 at 3:10 AM Pete French <petefrench at
ingresso.co.uk>
> wrote:
> > On 14/02/2019 01:43, Jason Tubnor wrote:
> > > I also have hit this IPv6 issue (I thought I was going crazy
until I
> > worked
> > > it out) and other iflib issues in 12.0, which have been fixed in
> -STABLE
> > > that really should be patched in 12.0 or bring forward an early
12.1
> > > release. For our use case, 12.0 is just too buggy for production
at
> this
> > > rate and we won't touch it, which is a shame because there is
a lot of
> > good
> > > work in there that we would like to use but it is trumped by the
> > breakages.
> >
> > Any reason behind not running STBLE out of interest ? Yes, 12 has been
> > buggy with regards to networking, but these things get fixed very fast
> > and I now have all my machines on the lattest STABLE in production, as
> > of yesterday.
> >
> > -pete.
> >
>
> Generally, not many.
>
> Far and away the biggest is the requirement to build from sources. It's
not
> a big deal for me, but if I still had many systems to deal with, that would
> be a pain.
>
Just as one can setup a poudriere/synth system for building custom binary
package repositories (so one builds packages on one system for easy
installation on multiple systems using binary packages), one can also setup
a custom freebsd-update server (so one builds the OS on one system, for
easy installation on multiple servers using binary updates).  And that can
be done to track -STABLE or -CURRENT, I believe.

Granted, I have never done it, nor looked too deeply into the documentation
around it, but I do know it's possible. :)  At least in theory.  :D

IOW, the days of needing to compile everything on each individual machine
are behind us.

-- 
Freddie Cash
fjwcash at gmail.com

Good morning,
> Am 14.02.2019 um 19:11 schrieb Kevin Oberman <rkoberman at
gmail.com>:
> Far and away the biggest is the requirement to build from sources. It's
not
> a big deal for me, but if I still had many systems to deal with, that would
> be a pain.
> [?]
> The bottom line is that the only real reasons I see for not running stable
> is the lack of binary updates, and issues with systems being slightly out
> of sync if all are not updated to the same SVN revision at all times. Those
> are very big reasons for many.
We build from sources centrally, then zfs send/receive /usr/src and
/usr/obj to all of the machines, then just do the install(kernel|world)
part on all of them.

Kind regards
Patrick
-- 
punkt.de GmbH			Internet - Dienstleistungen - Beratung
Kaiserallee 13a			Tel.: 0721 9109-0 Fax: -100
76133 Karlsruhe			info at punkt.de	http://punkt.de
AG Mannheim 108285		Gf: Juergen Egeling