[PLUG] Linux distributions

[Nathan Williams]

On Wed, 2015-08-19 at 22:17 -0700, King Beowulf wrote:
> On 08/19/2015 05:51 PM, Nathan Williams wrote:
> > I'll echo the endorsement for CentOS 7. It's reasonably new enough
> > to offer
> > new-ish packages for desktop use, while also being the "gold
> > standard" for
> > stability and long term support (EOL in June 2024). Add in EPEL and
> > elrepo
> > for "extras", and you've got a pretty solid system that you won't
> > have to
> > totally replace every 6 months.
> 
>  CentOS is no longer an independent project, more like a test bed for
> Redhat.

I'd dispute this phrasing (specifically in re: independence). Red Hat
is indeed a major contributor/supporter, with a majority of seats on
the governance board, but CentOS is and remains a community
led/developed distribution, certainly to a greater extent than e.g.
Ubuntu is independent of Canonical. I have yet to see Ubuntus lack of
independence mentioned as a reason to avoid Ubuntu as an end-user.

As for being a test-bed, it's certainly a more innovative CentOS
community than in the past, but most of that work is going on in SIGs,
so doesn't have a lot of impact for someone using core CentOS and not
one of the SIG-derived products like RDO, Atomic or the Cloud images.

>   Being supported for 10 years is great for a server, not so much
> for end-user desktops.  For a desktop, some system files just can't
> be
> upgraded past a certain point without compromising system stability.
> 

Agreed. I'd expect a desktop user to upgrade distros on average every 3
-5 years, basically following typical hardware upgrade lifecycles.

> > 
> > In addition, it's got that new-fangled systemd init system that all
> > the
> > major distros are going to be on within the next year or so, so you
> > won't
> > end up having a really core piece of the system be different from
> > what
> > everyone else is running (better supportability if you need help
> > down the
> > road).
> > 
> > Cheers,
> > 
> > Nathan W
> > 
> 
> the OP may want to hold off on any distro jumping on systemd.  While
> the
> issues and controversy concerning systemd may not mean much to end
> users, it does to DEVELOPERS.  You know, the guys and gals that
> create
> the software that creates a linux distribution and the various
> programs
> we want to use.

I knew I'd regret mentioning this as an advantage the moment I sent it,
and I'm guessing from the tone of the reply that it struck a nerve. If
so, I'm sorry, and I sympathize.

But, given the broad range of distros who are consolidating around
systemd, it seems reasonable to conclude that the majority of the
developers who do build linux distributions are in favor. As you noted,
it probably doesn't matter much to an end-user either way.

What does typically matter quite a lot to end-users, and the reason I
mentioned it at all, is the availability of support for issues. Since
an undeniable majority of Linux users either already are, or shortly
will be using systemd-based distros, I figured I'd mention it as a
point in favor (sticking with the herd), but that's truly the limit of
the degree of the intended endorsement.

Personally, I'm pretty ambiguous about it overall (migrating was rather
unexciting, both for workstation and servers), though I'm happy for
many of the new resource-control capabilities that the integration with
cgroups has made available to me as an administrator.

Regards,

Nathan W


> 
> -Ed
> 
> _______________________________________________
> PLUG mailing list
> PLUG at lists.pdxlinux.org
> http://lists.pdxlinux.org/mailman/listinfo/plug