On Thu, 24 May 2018, Timothy Sipples wrote:
> Russ Herrold wrote:
> > It may turn out that we (ClefOS) need to fork and offer two
> > variants
>
> I guess I'd call them "streams" rather than "forks."
One might, but in Red Hat parlance the 'Z' [a namespace
collision here, NOT referring to Arch] extended support tail
is called a 'stream ;)
> For what it's worth, Red Hat seems to offer at least 3 major
> streams now: Fedora (their "community" release), RHEL
> Structure A, and RHEL
The argument might be made, actually that there is a fourth:
CentOS
in light of their take-over and re-casting of that project a
few years back. As one of its three founders, I was perfectly
happy to simply keep the community's source code resources
available in a binary form, simply replicating RHEL, deviation
for deviation, bug for bug. Post acquisition, RHT added a
plethora of distractions, and started using CentOS as a second
'testing bench' but without so much of that pesky community
'hoots and hollars' which the Fedoraproject developed
> The RHEL Structure A/RHEL pair of streams is a unique
> offering for the s390x architecture branch, at least for
> now. (Is it a one-time aberration or the start of something
> new? I have no idea, so ask Red Hat, I guess.)
I spent some time researching this yesterday. The kernel-alt
variant seems to be timed as to release mid RHEL 7, priced and
positioned to be a 'hard core' and high volume performance
engine for (particularly Z based) Docker / Open Shift / Atomic
transient / ephemeral common kernel instances [isolated by
Linux cgroups, and SElinux], rather than lots of 'each running
a potentially different kernel' KVM Open Stack instances,
which may be long-lived. "Open Shift" vs "Open Stack" is
another of those unfortunately chosen nameings, as the
initials 'O.S.' here as to two differing virtualization
technologies, as well as the commonly used: Operating System
;)
> In RHEL 7.5, Red Hat decided to offer kernel 3.10 (only) for
> all POWER processors prior to POWER9, and (only) kernel 4.14
> for POWER9. For X86-64 it's only 3.10, and for ARM64 it's
> only 4.14.
Look again, but think about shared kernel spaces, and it makes
more sense. This gets RHT an uplift on the kernel (and the
nice additions for Z in 4.14) and 'buys some time' as to the
fact that the 'long lived Enterprise maintainability' model
that RHEL offers, is really _too long_ for people wanting to
respond in a more _agile_ fashion, and to also provide a
platform capable of running the 'latest and greatest'
RHEL-next (externally, and for partially historical reasons,
RHT employees do not call formally call the next Major release
'8' yet, as it has not shipped yet / some 'leaks' by others
in bug trackers call it '8') has the longest ever interval
after its prior Major. Up to 1600 days since the RHEL 7 drop
as of Apr 30 2018. Previous intervals were:
574 days after RHL EE 6.2E
637 days after RHEL 2.1
476 days after RHEL 3
506 days after RHEL 4
1309 days after RHEL 5
1302 days after RHEL 6
> There are certain newer capabilities that RHEL 7.5 doesn't
> support on s390x that RHEL 7.5 Structure A does. Red Hat's
> release notes explain all that.
Lots of new nuggets in those Release Notes which were not in
the Beta copy. I'll miss sendmail, but not XFS
Actually there was a lightly publicized 'stalking horse'
'testing candidate' as well:
RHEL 7.4 variant with an -alt kernel
I again use 'variant', as it varied, and will not use the
over-loaded word: fork here ;)
> But it's possible to mix RHEL and RHEL Structure A instances
> on the same machine and in a Red Hat supported way. (And,
> for that matter, other supported RHEL releases.)
You have the benefit of potential access in light of IBM's
long time and current collaboration with RHT, and I
hope so, but building a clean testing candidate has been
tricky
It turns out there is also a difference in how an 'El Torito'
capable [needed for one deployment method on Z] CD vs a DVD
image is spun, as there are s390x specific size requirements
that differ from other Arches, and some tooling changes were
needed
It's no big deal to have this kind of delay building a
community rebuild of sources. The test 7.5 binary packages
will 'yum update' over a base install of an earlier ClefOS 7
variant just fine. The ability to move the absolute
underlying base kernel back and forth between the 3.10 and the
4.14 series seems unlikely, and really, not something that
might not be better done with a fresh install. There is the
open question, which I have not tested yet, about trying a
kernel-3 to kernel[-alt]-4 transition, but again, probably
better done with a fresh install. It will probably end up
with a minimal installer image based on each
Thank you for your insights
> --------------------------------------------------------------------------------------------------------
> Timothy Sipples
> IT Architect Executive, Industry Solutions, IBM Z & LinuxONE,
> Multi-Geography
> E-Mail: [email protected]
-- Russ herrold
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
----------------------------------------------------------------------
For more information on Linux on System z, visit
http://wiki.linuxvm.org/
