s/Red Hat/IBM/g
As we've already seen with CentOS, anything relying on a Red Hat code base is doomed when IBM start meddling.
Fedora, the community Linux distro used by Red Hat for early implementation of new technology, is not just for experimentation, project leader Matthew Miller tells us. Miller is "ultimately accountable for everything that happens within Fedora and in particular is responsible for maintaining Red Hat's relationship with Fedora …
"IBM/Red Hat might start to think Fedora excess to requirements."
My (limited) understanding is that Fedora releases act as the source of future RHEL/Stream builds, so that a given Fedora release is identified as being a suitable basis for future RHEL, then the packages are forked and the RHEL bug squashing and testing process begins.
If I have that right, a Fedora like activity resulting in an actual 'release' with RPM style packages and a coherent set of libraries &c would need to continue to feed into the RHEL/Stream refinement process whether or not the outcome of the Fedora like activity was publicly released.
As the Fedora project gets a fair amount of support from outside of RedHat staff, IBM/RedHat might well decide that the free work was a good trade for server space and some publicity and funding. We'll see.
The point of RHEL is that it’s a hardened version of an already tested-in-production system. This is what separates it from other enterprise distros like SUSE or Ubuntu LTS, which only see limited external testing before release. Without Fedora there would not be constant production-grade testing of each and every changeset nor would there be a ton of free labour doing the key integration work for the riskier new technologies which may take off.
CentOS Stream only provides early testing of what will be minor point releases and those are glorified Service Packs at most. It’s just not feasible for Red Hat to kill Fedora without giving their competitors an edge in mindshare and willing testers; especially when openSUSE Leap is the same as SUSE Enterprise under the hood and the more Ubuntu LTS adoption there is, the more free advertising Canonical Advantage gets!
"The point of RHEL is that it’s a hardened version of an already tested-in-production system."
I'd argue that today it's solely for licensed support. I still remember the day I logged in and BAM!... no more updates, commercial only. Now it appears IBM is going to swing some more doors shut.
People should just get out while they can, unless they like dancing to "Oracle'tral Manoeuvres in the Dark".
Sometimes it's bad enough trying to keep an estate of Enterprise Linux up to date in production. I can only assume some of those mentioned *really* want new kernels/packages. I guess if you frequently pull down and rebuild using a suitably large amount of automation then it could be viable.
I think I've got to the age where if it works I'm less inclined to start messing with it (life is too short). So on that point I'd have a hard time even using Fedora on the desktop for a prolonged period of time. Though I do download it and throw it on a VM every few releases just to see what it's up to.
I'd say overall it's a trade-off between having the stability of a more long-term distro or having the time to mess with "bleeding edge". Yes I deliberately used that phrase, because Fedora is. It pretty much follows the "move fast and break things" mantra.
The single code line maintained by Ubuntu makes a lot of sense. Dip into the interim releases as required, otherwise stay with the LTS drops. My only reservation is the insistence on a fixed date for the release. If some things are not ready then I would prefer to see a delay in the schedule than produce something half-cocked. But that's why many people never go for a 1.0.0 release, and wait for 1.1.0.
I've been using Fedora on the desktop at home since it started and am pretty happy with it, one or two things that were not well-maintained have eventually dropped of, but there's generally been some replacement (I suppose that's life). I'm less interested in stuff like Btrfs though, and have been using KDE rather than Gnome for years now. It's not really clear how Fedora can be it's own thing, but also a test ground for RH. Additionally if the squeeze is being put on CentOS in production, then who would now consider Fedora for production and believe the same wont happen?
Same here. I've been using Redhat Linux since version 6.2 and Fedora since it first appeared.
I wasn't happy with Gnome 2's half-finished release and then, still unfinished, its abandonment in favour of Gnome 3, which I hated on sight. At that point I switched to XFCE and have been there ever since. Why not? It does what I want and doesn't annoy me with excessive decoration and fiddly bits.
I know RedHat in general has been the go-to for business for a long time, but between Fedora and Centos there's no question they are doing their damndest to squeeze their grip.
Time to be thinking about moving on to something else. Plenty of good options available with enterprise "support" that don't involve being shafted by a failing IBM. Also - do you want to be dragged down with IBM's failings?
Oracle has run a fully-supported clone of RHEL since the JBoss purchase, and Oracle does provide some very useful additions (the UEK, and their EPEL mirror are my favorites).
The reaction to Red Hat originally dropping their free product first saw WhiteBox Linux, then Tao Linux as free rerolls of RHEL, finally evolving to CentOS, while maintaining Scientific Linux for v7.
Rocky Linux and AlmaLinux are two new CentOS 8 clones. There is already a simple migration for CentOS to the Oracle yum repos, and migrations are likely to emerge to these new distros as well. Any CentOS user who wants a supported platform today can have it (even Red Hat has a converter).
IBM may have an agenda for Fedora, but it is unlikely as extreme as CentOS. For those who have truly had enough, but prefer RPM, SUSE is also an option.
FFS, Gnome 3 was broken. I've recently read the article about Gnome 4, and that looks just as shite. I keep saying this but:
* No hot corners
* No top bar
* No silly thing on the side
* Put ticks on the slide bars (then you can actually use it with a 30,000 line output to hunt for text)
* Stop the slide bars from disappearing (and make them wider, FFS). The mouse is a device to let you click on things. How can you do that when the thing you click on is painfully thin and keeps disappearing when you move your mouse away. Who thought this was a good idea?
* I like the menu at the top of the window -- STOP putting stupid non-standard buttons into the title bar. I'm looking at you Gedit. (Put the buttons together in one place, lets call it a menu, instead of any spare corner space you can find -- all over the window rather than conveniently together.) Someone please go on a UI course.
Just stop fucking it up.
I love Linux but hate Gnome. Can we do a Nominet on them and kick out their architects and get someone who knows what a desktop should do?
Please.
"Although Red Hat sponsors it, Red Hat doesn't own it in a meaningful way. The community makes decisions about what's going to happen."
Beat me to it.
I was wondering if he actually kept a straight face while saying that.
And if he thought his interviewer was stupid.
O.