[BlueOnyx:23202] Re: Upcoming BlueOnyx 5210R additions

[Ernie]

Hi Michael,
Lol at the WoW binge.

Sounds like you have had a good think about the CentOS situation.

Debian is my favourite NIX, I use Debian 10 all over the place, upgrades just work, 
the binary package management is better than FreeBSD, though FreeBSD is great for those that
like to work from source and hate systemd.

The change from rpm to .deb is major, but mostly at the package management
level, once unpacked BX would be fine.

>From a credibility perspective, Ubuntu 18.04 LTS would be an obvious choice, a major
upgrade every 2 years with 5 years support, that would be the closest thing to RHEL in the debian
world, but Debian itself runs some heavy duty stuff eg. Proxmox which is
pretty stable and can take a hammering in data centers.

I too have been thinking about how to decouple the base layer of BX and use
containers to deliver the services. I think it's totally viable, it
partially addresses HA with things like swarm, and containerization seems to be the direction
enterprises want to go, rather than the monolithic setups we are use to.

Kubernetes, the container leader,  is irrelevant complexity, Swarm is more than enough to allow BX to geo-replicate and have HA.
cced would need to talk to a distributed database backend.

Containerization a big project, a port to Debian 10 would be an order of magnitude easier.

Introducing containerization could be done incremenally over relases, serivce at a time sort of thing.


- Ernie.






> Hi Ernie,
> 
> > It's pretty sad that Centos can't keep up with RHEL releases.
> > RedHat changed their release schedule to 6monthly but they are suppoting
> > multiple versions at the same time. 8.1 has been available in beta since
> > July 24th. and Centos havent done any work on 8.0 since the 7.7 code became
> > available. It's a mess.
> 
> Yeah, and (according to a Reddit post) CentOS also had one of their lead
> developers disappear for a few weeks because he went on a WoW binge. Go
> figure. "Mess" doesn't start to describe that kind of cringe-drama.
> 
> > What's the fallback, use Fedora like BlueQuartz originally did?
> > Fedora support cycle isn't very long, but they are the upstream for RHEL.
> Fedora isn't really an option either, isn't it? It's lifetime cycle is
> really uncomfortably short and you still can't do in-situ upgrades from
> one major version to another without burning the barn down. If that were
> possible it might be an option (not a good one, though), but it isn't.
> 
> When we start to look outside of the RedHat sithspawn-of-hell family the
> list of available options for a suitable enterprise Linux or *NIX based
> OS quickly collapses to less than a handful of options:
> 
> - Debian
> - Ubuntu
> - OpenBSD
> - FreeBSD
> 
> None of them is RPM based, so our build architecture would need to be
> adapted to spit out something other than RPMs. Which is doable, but
> quite a pain.
> 
> If we say it must be RPM based, then these here might be options:
> 
> https://en.wikipedia.org/wiki/Category:RPM-based_Linux_distributions
> 
> But when you look at them in detail you'll see that it really boils down
> to only three serious contenders and we already know we're not touching
> the third on this list with a 10 foot pole:
> 
> - Mageia (fork of Mandriva)
> - OpenMandriva (fork of Mandriva)
> - Oracle (nope!)
> 
> Even Mageia and OpenMandriva aren't that suitable to us. Because the
> thing is that we can't base BlueOnyx on some kind of exotic "I've never
> heard of that!"-OS with mostly unknown, possibly dubious or short lived
> history. Because we don't know if it'll still be around tomorrow, next
> month or next year. Likewise many of us use some kind of third party
> stuff here and there and if our favorite toys aren't available on
> "ExoticOS 08/15", then we'll not be happy. It's also a hard sell if you
> have to tell your clients the (dubious) benefits of some unheard of
> "ExoticOS 08/15" if you yourself hardly know it well enough.
> 
> So it has to be one of the big names and we're back at the first list
> shown above.
> 
> >From that list I'd directly ditch Ubuntu like a hot potato, though. It's
> a great desktop OS and I've been using it for years on my office PC and
> the Laptops. But for heavy duty datacenter stuff with preference on long
> uptimes it not suitable.
> 
> The *BSD flavors are all great and I like them. Yet the switch from
> CentOS to either OpenBSD or FreeBSD would be a hard sell, too. They're
> different enough to cause you grief during the transition until you've
> learned all the intricacies. And some of us use forms of virtualization
> that wouldn't support a *BSD well enough to allow me to choose it. So
> ... that's not really an option either.
> 
> Long story short: I'll be using the downtime while we wait for CentOS 8
> to explore doing a BlueOnyx port to Debian 10. I'll not say that there
> *will* eventually be a BlueOnyx for Debian, but I'll give it a honest
> and serious try as time permits it. Just to see if I can adapt our build
> architecture to spit out *.deb's instead of RPM's.
> 
> Due to an ongoing support contract I've spent enough time around Debian
> 8, 9 and 10 to start to appreciate it. And the best part of it is
> certainly that you *can* do in-situ upgrades. Did some Debian 8 -> 9 ->
> 10 upgrades recently and almost all went without a hitch for as long as
> you hadn't tied them into too many exotic third party repositories.
> Which is the other benefit of Debian: You can get sort of everything for
> it from third party repositories.
> 
> However: As is CentOS 8 will still get priority treatment because I
> don't want to rock the boat too much. But pursuing to solve the
> technical obstacles that currently prevent us to do a Debian port might
> be useful in the long run.
> 
> There is also another option which keeps floating around. At least on an
> esoteric level. Call it a brain fart, if you will: With the prevalence
> of virtualization we do have the option to decouple BlueOnyx from the
> actual base OS, making it more or less OS agnostic (within limits). Like
> stuffing it into a Docker image, a Snap or use Kubernetes. But: Docker
> images or Snap images are like the box of chocolates that Forrest Gump
> has on his lap on the park bench: "You never know what you get."
> 
> While there could be some merits to this approach: The whole idea
> doesn't convince me.
> 
> TL;DR: We wait for CentOS 8 as it's still our prime contender. Meanwhile
> I'll put up a Debian 10 "playground" to see if I can adapt the BlueOnyx
> build environment to produce packages for Debian. If RedHat's hellspawn
> eventually starts to stink much more than it already does? In that case
> we then have the long term option to go for Debian instead.
> 
> -- 
> With best regards
> 
> Michael Stauber
> _______________________________________________
> Blueonyx mailing list
> Blueonyx at mail.blueonyx.it
> http://mail.blueonyx.it/mailman/listinfo/blueonyx