The "New" great debate: Centos "alternative"

seriesn

As most of you already read/heard/been told that centos is "Dead" (or not, depends on your perspective), I hope this thread can turn into a great resource for those, who in the near future will be googling "Centos alternative" or "Best centos alternative".

So far, here are the options for those who are looking to migrate without reinstall/rebuild :

• Oracle Linux: https://linux.oracle.com/switch/centos/
• Cloud Linux "Community edition": https://blog.cloudlinux.com/announcing-open-sourced-community-driven-rhel-fork-by-cloudlinux
• Rocky Linux: Created by the original centos founder (https://rockylinux.org/)

Point to be noted, both OL and CL are for profit companies and may or may not pull an IBM.

The followings are also an option, if you are looking to forget about Rhel, like it was that EX, you had great time with, but don't want anything to do with anymore.

• Debian (supported by a decent amount of current softwares) and
• Ubuntu (Supported by decent amount of commercial control panels, including Cpanel, announcing their future plan

---

What do you think? Share your expert/non expert opinions below. Feel free to share your pushup videos ( @yoursunny ) as well. Looking forward towards some exciting and interesting conversation/healthy debate.

If in doubt, simply blame and sue @deank and only sue @deank cause @deank has been saying since last year, End is neigh. Look at 2020.

ehab

debian (when k8s support is needed - iscsi - beats)
alpine ( when the vps spec is low and possibly used for docker)

yoursunny

Everyone should switch to Ubuntu LTS or Debian Stable.

I do not use RHEL derivatives because, as a geocacher, the word "DNF" has a negative meaning.

---

Push-ups solve all the problems. Serve push-up videos from all your VPS. Never leave a VPS idle again!

lentro

Staying away from Oracle right now with their Java garbage. Don't trust them right now. Same with CL - they are for profit, so I'm afraid they will pull the plug some day (or an investor will).

Rocky Linux isn't ready yet, but it's the most promising. I'm in their Slack, and there's a ton of momentum.

In the end, I voted for Ubuntu because 95% of my servers use that. It's my fav

angstrom

Debian and Ubuntu are all fine and well, but they won't satisfy those who are used to the CentOS ecosystem

rcy026

Just to be clear, the main problem with CentOS dying/changing is finding a distribution with support from various panels?
I've never used such a panel in my life so I find this "which distro is best?" a bit confusing, since I simply run whatever works best for whatever purpose the server has. But I can see that as a provider you want/need to slimline the environment and keep it as homogeneous as possible.

I suppose that the developers of the various panels are aware of this, and will focus their effort on some other dist in the near future. Supporting a dist that nobody wants to use seems like bad business so they will have to go some other way pretty fast, and running down a completely unknown road like Rocky Linux is very risky so my bet is that Debian and/or Ubuntu is the path most will take.

seriesn

@rcy026 said: Just to be clear, the main problem with CentOS dying/changing is finding a distribution with support from various panels?

Not necessarily the case. However, a decent percentage of userbase usually picks an OS or even VPS spec, based on their preferred control panel/software stack support.

Obviously, the other concern, that is, do you really want to be run production server on a "beta" software/'s. If you are a power user, this may not be as relevant to you vs someone, who runs yum update -y blindly.

InceptionHosting

For me for anything that needs to be RHEL based I will likely lean towards rocky. I just hope someone forks rocky and calls it adrian.

but really I hope generally debian is just seen as the sable base from now on, it really should have been 10 years ago.

jureve

There are several alternatives. However I use Debian and Ubuntu on my servers.

Joshua

Debian is my typical go to

Iroshan464

@seriesn said: RL founder did initially sell Centos to RHEL.

I read somewhere that Gregory Kurtzer was not on the board when centos was handed over to RHEL. Maybe this time he will stay.

PHP_Backend

@seriesn said: RL founder did initially sell Centos to RHEL

Partly misleading, Greg Kurtzer didn't sell CentOS to RH, he left before the deal.

I was forced to relinquish leadership of the project to a party in the UK

source: https://gmkurtzer.github.io/

From there, due to a number of situations not really germane to this article, Greg moved on

source: https://blog.centos.org/2019/03/greg-kurtzer-centos-founder/

seriesn

@PHP_Backend said:

@seriesn said: RL founder did initially sell Centos to RHEL

Partly misleading, Greg Kurtzer didn't sell CentOS to RH, he left before the deal.

I was forced to relinquish leadership of the project to a party in the UK

source: https://gmkurtzer.github.io/

From there, due to a number of situations not really germane to this article, Greg moved on

source: https://blog.centos.org/2019/03/greg-kurtzer-centos-founder/

Ahh! Did update on the OGF. Forgot to do so here. I do stand corrected.

PHP_Backend

To me, it is too early to take a decision.

CL is another company, but I am curious with their "we will make our build system & tools opensource" promise. If it is done correctly, even other fork can take advantage of it. Cross-contribution can definitely improve both side. A win-win situation.

Yes, Rocky linux will be "Beta" for a certain time. But remember, when CentOS first emerged, they didn't have this many userbase like now they have. Lots of people already ready to jump the ship and test the hell out of it. After the release day, you providers can be sure that you will receive lots of request to setup template for Rocky.
And, this readymade userbase will make the power/enterprise users to take notice of it. One step at a time, man. Word of mouth has been always effective in FOSS world.

And, moving to non-rpm system is out of question (for now). There are huge amount of codebase that targets rpm-based system. Changing those codebase is a big no no. That amount of work will never justify the reason to move out.

So, sit back and relax. Support any party that comes with a good plan. And when the day to take a decision comes, make sure you choose the most flexible one. Keep calm and trust in FOSS.

terrorgen

I have been used to debian for my server needs. Recently deployed a few alpine Linux machines, those are prem.

Never see a need for CentOS.

nem

Still too early to see. Give it a few months to see how traction develops and who's committed to what ideologies. Any drastic action now is akin to your landlord telling you your apartment won't be renewable at the end of the term, then FUD'ing around looking for a new place to live 12 months in advance.

It's something that bears caution, but there's a lot of time between now and Summer when firms need to begin to decide how they want to respond to this.

DataRecovery

Just to make sure I have a proper understanding of the situation: am I correct to assume that the primary reason people started looking for an alternative is that we won't have a "stable" CentOS, which would be safe for production use?

@seriesn said: Share your expert/non expert opinions

I always been using CentOS for my hobbyist and work-related needs. I will probably keep using CentOS 7 for as long as possible and will switch to CloudLinux afterwards.
These Russian guys seem to be the best prepared to take over the game.

InceptionHosting

@seriesn said:

@PHP_Backend said:

@seriesn said: RL founder did initially sell Centos to RHEL

Partly misleading, Greg Kurtzer didn't sell CentOS to RH, he left before the deal.

I was forced to relinquish leadership of the project to a party in the UK

source: https://gmkurtzer.github.io/

From there, due to a number of situations not really germane to this article, Greg moved on

source: https://blog.centos.org/2019/03/greg-kurtzer-centos-founder/

Ahh! Did update on the OGF. Forgot to do so here. I do stand corrected.

Rude

elliotc

Been migrating to Debian/Ubuntu for a while.

Harambe

I think CloudLinux community edition will probably get the most market share in the hosting world, considering that's their primary focus and it'll be easy for DA/cPanel/etc to sign off on it as an approved OS.

I'll personally be sticking with Debian. Would love to see a serious Debian LTS project with 5+ year support cycles so I never have to run Ubuntu on a server ever again.

yoursunny

@Harambe said:
I'll personally be sticking with Debian. Would love to see a serious Debian LTS project with 5+ year support cycles so I never have to run Ubuntu on a server ever again.

I use Ubuntu LTS but upgrade every two years, typically after XX.04.1 version comes out.
I have Debian on smaller machines, but haven't experienced an upgrade so far.

I like to live close to the edge.
I'm definitely not waiting 5 years to take advantage of the latest features.

Quote on legacy code:

The stallion coder doesn't deal with legacy code. I rewrite all of them every year.
Dependants? It's your problem. I gave you 30-day notice on deprecations, and then it's removed. You are supposed to rewrite dependant code every year, too.

Harambe

@yoursunny said: I like to live close to the edge.
I'm definitely not waiting 5 years to take advantage of the latest features.

I'm getting old, I just want stuff to work man.. lol.

Stability is more important than new features in 95% of stuff I do, just give me a nice foundation with security updates and I can sleep well.

yoursunny

@Harambe said:

@yoursunny said: I like to live close to the edge.
I'm definitely not waiting 5 years to take advantage of the latest features.

I'm getting old, I just want stuff to work man.. lol.

Stability is more important than new features in 95% of stuff I do, just give me a nice foundation with security updates and I can sleep well.

Don't you want:

• AF_XDP sockets so you can transmit more than 10 Gbps in one core?
• kernel drivers for the shiny new 5G modem?
• acceleration for post quantum crypto?
• compiler optimization for next year's CPU model?

Harambe

@yoursunny said: Don't you want:

Nope. I just need my servers to keep brrr'ing along at a reasonable speed. If I have new hardware then I obviously need the newer kernel, but that doesn't happen often.

AlwaysSkint

I've been running CL for a few years, to support older PHP on cPanel. Gotta say that I ain't keen - there's a big overhead involved, in file structure/space and running resources.

yokowasis

I still don't get it. Ubuntu has butt load of derivatives. Can't centos do the same? Can't just people fork centos and only allowed battle tested package / repo?

You know creating something like CENTOS LTS

havoc

Started standardizing everything on debian. Compatible with most ubuntu heavy tutorials but without canonical's bullshit (snap etc)

PHP_Backend

@yokowasis said:
only allowed battle tested package / repo?

The challenge is, who will fight the battle?

Ubuntu has butt load of derivatives

and, none of them are that much stable to use in server env. (correct me if I am wrong)

seriesn

@DataRecovery said:
Just to make sure I have a proper understanding of the situation: am I correct to assume that the primary reason people started looking for an alternative is that we won't have a "stable" CentOS, which would be safe for production use?

That’s correct.

@AnthonySmith said:

@seriesn said:

@PHP_Backend said:

@seriesn said: RL founder did initially sell Centos to RHEL

Partly misleading, Greg Kurtzer didn't sell CentOS to RH, he left before the deal.

I was forced to relinquish leadership of the project to a party in the UK

source: https://gmkurtzer.github.io/

From there, due to a number of situations not really germane to this article, Greg moved on

source: https://blog.centos.org/2019/03/greg-kurtzer-centos-founder/

Ahh! Did update on the OGF. Forgot to do so here. I do stand corrected.

Rude

That’s whatcha get for not sending me good morning and good night texts.

@AlwaysSkint said:
I've been running CL for a few years, to support older PHP on cPanel. Gotta say that I ain't keen - there's a big overhead involved, in file structure/space and running resources.

The overhead never felt that major on this tbh. Pretty sure once you remove LVE and all the fancy stuff that Cl Introduced, the OS would and should feel lean. Doubt they will be throwing in their signature products in the free version.

rcy026

@yoursunny said:
Don't you want:

• AF_XDP sockets so you can transmit more than 10 Gbps in one core?
• kernel drivers for the shiny new 5G modem?
• acceleration for post quantum crypto?
• compiler optimization for next year's CPU model?

Actually no.
Never had a need for 10Gbps in one core, I don't run servers on 5G and rarely compile anything.
Developers might need that kind of cutting edge, but developers don't need stability in the same way providers do.

rcy026

@seriesn said:

@rcy026 said: Just to be clear, the main problem with CentOS dying/changing is finding a distribution with support from various panels?

Not necessarily the case. However, a decent percentage of userbase usually picks an OS or even VPS spec, based on their preferred control panel/software stack support.

Obviously, the other concern, that is, do you really want to be run production server on a "beta" software/'s. If you are a power user, this may not be as relevant to you vs someone, who runs yum update -y blindly.

This is basically what I meant. The reason people (atleast to some extent) run CentOS is that it's supported by their preferred control panel/software. If the panels change track, people will follow, and hence they will not have to run "beta" software on their servers.