If you want a drop-in replacement for CentOS and you’re using RHEL-things like oVirt, FreeIPA, OpenShift, etc, these are the only options that are remotely close IMO.
Try CentOS Stream and see how it goes
Try Oracle Linux and see if you can still look at yourself in the mirror
There was some talk of a free RHEL?
And yeah we’ll see how Rocky goes…
Distant 2nds if you’re simply using CentOS to do generic server things.
OpenSuse Leap
Ubuntu LTS
Debian
I wouldn’t consider Fedora a suitable alternative as the upgrade cycle is too frequent and stability isn’t there. FreeBSD and other linux distros are fine, but don’t really resemble CentOS in any meaningful way, so I wouldn’t consider them alternatives specifically to CentOS.
Right, I kinda forgot about this. I would personally arrange them in the following order:
Oracle Linux
OpenSUSE Leap*
Ubuntu LTS
Debian
Rocky sounds cool, but it is an uncertain project at this point in time.
*I put OpenSUSE in the second place because I know of its stability and praise in the enterprise, but as always, if I haven’t used it, I cannot vouch for it.
Has anyone tried Amazon Linux 2? How about Clear Linux as a replacement for CentOS in some cases?
I can. its praise is well earned and often highly understated. They have a strong leadership in the community and a couple really really really good devs in charge of most of their distro packaging and testing teams
Personally, replacing CentOS with Oracle Linux had been my approach at work for quite a while before even any idea of a rolling CentOS release. For the simple reason, that I can suggest clients to use Oracle Linux in production if they don’t feel like buying a RHEL license outright. With OEL, they have the ready option to upgrade to paid support.
I’m not a fan of Oracle the corporation but from a purely business perspective, it is the next best alternative. Feel like RHEL/CentOS is shooting themselves in the foot with this move.
We’re (well my dad; I’m tagging along ) shifting from CentOS to openSUSE Leap for servers. Since the software my dad works on is made to run on lots of different *nix platforms, it’s easy to switch over and if SUSE does some shenanigans, then switching again will be no big deal either.
Well the base OS has to run on the bare metal though? Most of his stuff he puts VMs for dev/testing, but it’s not the kind of stuff you put in containers; it’s made to run on big single-purpse computers. I don’t think I can say more though, partly because its way over my head.
Thats cool; but how can we verify containers so there aren’t shenanigans? The only Docker container we are running is a GitLab one that is from Synology that we run on one of our NAS. I kinda think I’d be happier with a playbook that installed the parts…