Please file in a calm and orderly fashion toward the exits and remember: IBM killed Centos for profit.

Terraform users can pick up their new alternative here:

https://opentofu.org/

and for those of you with Vault, you can find your new alternative here:

https://openbao.org/

>IBM killed Centos for profit

The community is better off with the combination of AlmaLinux + CentOS Stream.

AlmaLinux provides differentiation by supporting hardware and patches that Red Hat doesn't necessarily want to.

CentOS Stream provides a way for all downstream distros (Red Hat, Alma, Oracle, Rocky, whatever) to contribute back to the ecosystem in a way that is deeper than submitting bug reports to bugzilla.redhat.com

The latter point is something people really miss. CentOS was not much of an "open source" project in the same way that Android isn't much of an "open source" project. There was few ways to contribute directly beyond filing a bug report and waiting for Red Hat / Google to handle it. CentOS Stream is vastly more open as a project.

Wasn't the value of centos that it was bug-yor-bug compatible with RHEL but without the license fee requirement?

>bug-yor-bug compatible with RHEL

It was never quite that, but that was the theory.

The problem is because it was an exact clone of RHEL, it's somewhere between difficult and impossible to have a proper community around it. Like I said it kind of revolved around "file a bug to Red Hat, wait for someone to pick it up, and then wait more for it to be shipped, which will be months away at best and could be years if it's not a paying customer's priority"

What I've seen people use CentOS for is testing a product intended to run on RHEL. For that, CentOS didn't need a separate community, and bugfixing at a pace different from RHEL would have been actively undesirable.

For that purpose, RHEL is completely free to begin with

Yeah RHEL has had a gratis developer license from their website for maybe a decade now?

Before that I would pay $49 a year for it.

As far as I understand, you still need to connect to the Red Hat license server, and you're limited to 16 machines. That's just stupidly unsuitable for a CI worker pool of hundreds of VMs being created from scratch for every run.

Yeah that’s something they’d want you to pay for.

I totally get why CentOS was useful. I was briefly involved with Rocky when it first was getting started.

If you do pay for RHEL you can have free non production subscriptions. Of course it's probably suspicious to pay for 1 production and 20000 testing servers, but it's pretty good if you are too small to negotiate an unlimited-machine subscription and you want to use RHEL in production.

https://developers.redhat.com/articles/2022/07/06/what-quali...

Not for production purposes.

OP was talking about testing purposes.

But RHEL is free for limited production purposes also, to a maximum of 16 systems.

https://www.zdnet.com/article/red-hat-introduces-free-rhel-f...

> IBM killed Centos for profit.

That process began happening long before the IBM acquisition when Red Hat hired Karanbir Singh (CentOS's maintainer) in 2014.

What's the story behind that?

IIRC, I saw something about it on HN at the time, but did not read it, being busy.

Red Hat had open source/community versions of their products, for example OpenShift Origin (now OKD) and RDO (for OpenStack). They wanted to use CentOS but it was not working well because release dates of CentOS were not predictable.

So they acquihired CentOS, but in doing that they changed its target from "community of users [that don't want to pay for RHEL]" to "community of developers that target RHEL". CentOS Stream is much better (though not perfect) for the latter population.

That said, there are many people using CentOS Stream in production (including myself).

Got it, thanks.

What about Packer? Vagrant? Consul?

Do you even realize OpenBao is an IBM initiative?

It's not, it's a Linux Foundation initiative with some IBM developers involved. IBM cannot kill it.

Do you know that OpenBao is actually funded by IBM?

I'm biased (co-founder) but you should take a look at Infisical for secret management: https://infisical.com

As I said elsewhere, it's not. It's an open source project, under the Linux Foundation umbrella, with all of the good licensing that goes along with that. It has some contributors who are also IBM employees, but to say they're "funded by IBM" is a bit of a stretch.

FWIW, most of the code and docs contributions have come from non-IBMers [0]. That said, IBM has done a lot of great work building the foundation and initial community and without them, OpenBao wouldn't be here. :-)

Speaking for myself, but I do not get any monetary compensation from IBM and I suspect this is true for all of the other non-IBM contributors.

[0]: https://github.com/openbao/openbao/releases/tag/v2.0.0-alpha...

Is someone working on a Nomad fork?

I don't have any further insight, but looking at <https://github.com/hashicorp/nomad/forks?include=active&page...> coughed up https://github.com/atlassian/nomad/branches although confusingly it says "updated last week" but browsing any one of the branches seems to be stupid old so I got nothing

Finding conceptual forks, e.g. $(git push --mirror ...) would be trickier but I bet sourcegraph could do it

Ultimately, the question boils down to: what risk are you driving down: hitching your wagon to a dead stack, not getting security updates, not getting PRs merged, $other?

Off topic but how is Nomad for running production workloads? My org has been wanting to push into containerization, but we're finding the complexity of k8s a high barrier to entry given our small team and lack of in house knowledge.

Is there a specific aspect of k8s that's causing friction along the learning curve? I'm assuming your referring to the developer side of the equation, not the ops/admin workflows.

Before jumping straight into production k8s, something you can mess around with is using Podman to generate[0] and run[1] Kubernetes resources that you can parse through and familiarize yourself with. Paired with Podman Desktop[2] can produce a nice graphical environment. After that you could take a look at more production-simulating environments with minikube[3], OpenShift Local[4] (very much recommend if you have the resources to run it), and the no-cost OpenShift Sandbox[5].

In general, the Red Hat Developer[6] site has a lot of good resources to learn from, both passive and interactive. I highly recommend going through the courses and tutorials available if it can help your team skill up (assuming k8s is the direction you want to go in).

[0] https://docs.podman.io/en/latest/markdown/podman-generate.1....

[1] https://docs.podman.io/en/latest/markdown/podman-kube.1.html

[2] https://podman.io/features | https://podman-desktop.io/

[3] https://minikube.sigs.k8s.io/docs/

[4] https://developers.redhat.com/products/openshift-local/overv...

[5] https://developers.redhat.com/developer-sandbox

[6] https://developers.redhat.com/

Been running both nomad and k8s in prod for the last few years. Nomad has been stable and reliable to the point of replacing k8s in some projects because of how easy it is to work with and onboard developers.

If that is the driving force use a managed provider the AWS elastic container service or Google's app engine.

Maybe Incus[0] as an alternative instead of fork.

- [0] https://linuxcontainers.org/incus/docs/main/

I’m not happy about this but my silver linings are:

1) At least Ellison doesn’t end up with it

2) Consul should tend to bring down the aggregate complexity of IBM solutions just a little bit.

I would predict that IBM will cancel Consul and Nomad completely to focus on k8s.

If so, Consul → OpenCabbage, Nomad → OpenNoodle

If the community hasn't already forked it implies that there isn't enough community interest to continue development. We should be skeptical of any "pity forks" that happen later.

If opencabbage is real, it is very hard to find.

Maybe. Or they could just completely opensource the projects.

It’ll be about terraform won’t it?

True story...At one point in time, IBM bought so many companies, they produced internally an audio CD, sales reps were supposed to listen to on their way to customers visits. Reminding the name of the company, and explaining what each company was about...

The internet deserves a ripped copy of that CD.

I will see if a friend of a friend still has it's copy ;-)

I find it hilarious how much this resonates with my time at AWS. Except it was for new service announcements, and video rather than a CD. History repeats.

Jumped on the Java boat alongside Sun and Oracle, pivoting from Smalltalk into Java, contributions to the Linux kernel since 2000's, collaborations with Apple and Sony, helped porting Go toolchain into their micro and mainframes, acquired Red-Hat thus indirect contributions to all major Linux projects that Red-Hat works on,...

Back in the day I read that Eclipse was based on an IBM IDE for Smalltalk called VisualAge, which they ported to VisualAge for Java, which became Smalltalk.

A while ago, so some details may be wrong.

"… IBM VisualAge for Smalltalk and IBM VisualAge for Java. Both of these products were written in Smalltalk. The IBM VisualAge Micro Edition product was the first serious and actually quite successful experiment with writing the entire IDE in Java. Many concepts found in Eclipse have been tried out in that product already. However, for third parties, it proved difficult to extend the product with new components …

A small team of experts set out to take the experiences of the previous years of designing and implementing development environments. The result was Eclipse, a platform designed from the ground up as an integration platform for development tools."

FAQ Where did Eclipse come from?

https://wiki.eclipse.org/FAQ_Where_did_Eclipse_come_from%3F

One of those experts, Erich Gamma, is the main VSCode architect.

Wow.

One of the GoF guys, too.

https://en.m.wikipedia.org/wiki/Erich_Gamma

>VisualAge for Java, which became Smalltalk.

Whoops! What a brain fart, or cosmic ray effect, or whatever :)

I meant to write:

>VisualAge for Java, which became Eclipse.

It was still possible to edit and correct my previous comment, but since people had replied already, I decided to correct myself in a reply to my own comment above.

Thanks, guys, for pointing out the mistake.

VisualAge for Smalltalk came first.

Smalltalk (70s) came long before both Java (90s) and Visual Age (80s).

Yes, I noticed my error above, and corrected it, thanks to you, in a reply to myself.

They did a lot of different experiments with JIT as well, though I don’t know how much made it into other toolchains. In particular J9 was a sort of moral equivalent of PyPy.

On the downside they are materially responsible/accomplices to the crimes that are J2EE and XML ALL THE THINGS.

I don't begrudge them XML even if I don't like how it looks. Json still doesn't have a matching schema story.

J2EE was created by Sun, based on a Objective-C framework.

https://en.m.wikipedia.org/wiki/Distributed_Objects_Everywhe...

JSON and YAML keep trying to re-invent the nice tooling ecosystem from XML, badly.

Sun Proposed J2EE (and made a mess of it) but what it is and what it has been had a lot of chefs.

IBM had a lot to do with both XML and J2EE being used in the most Rube Goldbergian senses possible. They didn't invent XML either.

Lots of companies had a lot to to with J2EE and XML, they are both standards, designed by a consortium of companies.

Sun, Oracle, Weblogic, IBM, Red-Hat, Rational, Microsoft, Pivotal, Apache Foundation,....

And if I have to chose between XML, endless annotations, anemic JSON, YAML spaghetti, I will take XML, thank you very much.

Well, they also have Red Hat, HashiCorp is right up the same alley.

Which alley do you mean?

They are almost 100 years old which is an accomplishment. I wonder how many they have left. I bet a big chunk of that depends on companies still running AS/400 systems.

"Infrastructure", which includes z mainframes (called out by name) and I assume iSeries also (though not called out), is the smallest segment of IBM's revenue.

See page 17: https://www.ibm.com/annualreport/assets/downloads/IBM_Annual...

Much, if not most, of their consulting revenue is for mainframe support and much of their "hybrid cloud" is just hosting mainframes.

My semi-educated guess is that mainframes are a much bigger chunk still.

(AS/400 is midrange. They are called something else now, I heard. Maybe I-series.)

A lot of those indispensable infrastructure systems we all depend on, such as insurance, income tax, etc., run on mainframes, as they have for decades now.

So don't diss big iron / Big Blue.

I, for one, hope they have many years left, though I guess they could do with improving their corporate culture and business practices some.

Those z-boxes can comfortably run 5000 instances of Linux as your very own virtual data center. Makes you wonder what inference on mainframe would look like, though you don't have GPUs.

Do people use them for that, and is it cost-effective?

There was an interesting point in the recent Acquired podcast about Microsoft vs IBM. Supposedly IBM revenue only decreased after 2012 and Microsoft only passed it in revenue in 2015.

Aren’t they always fighting the last war?

IBM is good at selling to large organizations. Large organizations are risk-averse, and prefer not to buy bleeding edge. IBM is pretty much institutionally specialized to sell what was hot last year (or last decade).

People might make fun of this, but I'd like to note that there is a lot of very stable revenue in their business.

I'd also point out that given just how bloody the bleeding edge of software development tooling is, only investing in things that have managed to stick around for a decade sounds pretty damn smart.

Hell, React alone has been through 3 or 4 different paradigms in the last 10 years.

Makes sense when you put it like this. In which case HCP may even be too new for them.

Given the stock price… it had to be someone. Same for Gitlab soon I guess

Anyone want to guess who?

I'll start: Jetbrains?

Still don't get why JetBrains hasn't been folded into the Alphabet family yet.

I’m failing to identify the benefit for the buyer here. What’s your reasoning?

Kotlin and Android Studio are intrinsic to the Android development experience.

Why pay for what they get for free, then?

That's a fair question. Still, you often see companies snap up those who build premier developer tools for their ecosystems.

What if Apple bought them?

Like when Google bought Fastlane.

Probably were waiting to see how the company dealt with the sanctions on Russia first.

Brains are twisted: JetBrains closed its last legal entity in St. Petersburg - https://www-dp-ru.translate.goog/a/2023/03/01/Mozgi_svernuli...

True, but I was expecting Google to have bought them like a decade ago!

Atlassian

Please no.

I wonder if Microsoft grabs them.

Sad, because GitLab was great for awhile. Then they stopped fixing bugs and started adding unneeded features, and made their pricing very expensive.