Hacker News Reader

Comment by solatic | Hacker News Reader

It's like Vader asked Luke to join him in defeating the Emperor (Microsoft) and ruling the galaxy as father and son - and instead of the audience hearing "I'll never join you!!", we hear "sure, let's team up."

The cognitive dissonance is so strong here. WTF just happened? If you asked me a month ago to put down serious money in Vegas on this never happening, I'd have happily done so. What on Earth were they thinking?

IBM is basically taking it's failing Kubernetes distribution, saying "why lose when we can just buy the winners?", and went ahead and bought Red Hat and OpenShift instead. A year from now, we'll start to see heavy IBM integrations into OpenShift, radically increased licensing fees for RHEL to squeeze every penny out of enterprises which bought RHEL specifically because they need the support guarantees because they can't migrate away quickly, and every other Red Hat project - Ansible, Cockpit, Fedora, CentOS, etc., will get torn to pieces by IBM bean counters.

Red Hat shareholders just sold out. Goddamnit.

View on HN

This announcement made me really sad. Red Hat was a company I admired.

I guess the only silver lining is that the buyer wasn't Oracle.

Obligatory link to one of the best rants ever presented against Oracle as an acquirer: https://www.youtube.com/watch?v=-zRN7XLCRhc&t=34m

"Do not fall into the trap... of anthropomorphizing Larry Ellison... If you put your hand inside a lawnmower, it will get chopped off. The lawnmower doesn't care... The lawnmower doesn't want to kill open source. The lawnmower just can't think about open source. The lawnmower can't have empathy."

Ditto, I can almost guarantee I'll be in a meeting in the week or two where my directors will want to discuss the possibility of moving off of Redhat, at least as a contingency plan.

Or Microsoft (who, after all, <heart>s Linux).

I share the sentiment.

Gawd. Freaking. Dammit.

I've been using RHEL-derived systems for like almost 20 years. This actually feels like a betrayal of the Open Source community.

Any bets on whether Fedora and CentOS will exist in November 2019?

Where’s the history of IBM buying and sun setting companies? I don’t have the same prejudice towards IBM that I do for Oracle. But I’m not in a position to know.

I’m thinking RHEL’s support contracts will keep IBM from shuttering RHEL. An IBM branded RHEL would represent plenty of income.

SPSS comes to mind.

Can you expand?

ISS?

They can cancel many projects that RedHat do as a general giving back to Linux. IBM wouldn't care.

Rational Rose, Lotus

Can you expand?

WeatherUnderground was effectively shut down post-purchase. API signup is gone, etc. Terrible transition and no support for prior products.

This is 0% true...

Weather underground is a shell of what it used to be. They got rid of a ton of features and cluttered everything with ads

This is 100% true. They abandoned so many good parts of WU and, today, it seems to continue to operate as data collection (mobile apps / home weather stations) and ad revenue generation.

Sequent, though I can't see them doing the same things to RH.

zOS reboot, duh!

Kidding aside, nothing about this is good for the open source movement.

Donated to Apache...

The boilerplate and vague statement of Red Hat remaining a distinct unit doesn't really tell us anything. More relevant is what the CentOS and Fedora communities think of this acquisition, because no matter what they are community driven projects.

There are two coins to toss: whether IBM reaches into Red Hat in a way that kills off either project; and whether enough of the community steps out.

I'm curious what openSUSE folks think of SUSE having been acquired by Novel, then Attachmate, and then the Micro Focus merger. They've been through a lot, and openSUSE is still here.

Conversely, I wonder what this does to the Linux ecosystem. It looks like at this point, Debian and Arch are the only major self-sufficient distros (i.e. not built on top of another distro) that are still community-owned and community-driven; and of the two, Debian is clearly more broadly popular. So, will this result in Debian becoming the de facto standard of "open Linux"? This could make things interesting when it comes to packaging etc.

There are others, puppy Linux, Alpine, and slackware to name a few.

What about PCLinuxOS, openSUSE, Mageia, Void, Nixos, Solus ?

None of these qualify as "mainstream" compared to RedHat/CentOS/Fedora and Debian/Ubuntu. Even SUSE, which used to be in the big league, is a shadow of its former glory.

If you look at what runs on production servers, it's virtually always RHEL/CentOS, or Debian/Ubuntu. Everybody else isn't even above 5%, and most of those you've listed are in fractional digits.

For another data point, if you look at websites, Debian+Ubuntu is already >50%. At this point, I think it's well on its way to becoming the Linux distro, with everything else being relegated to the hacker/boutique niche. And I think that this announcement, and what IBM is likely to do with RHEL afterwards, will accelerate that trend substantially.

Shame. As far as I can tell OpenSUSE & Fedora are on the same level. I often struggle deciding which one to install on new servers because I love them both so much.

well, that’s basically already happened. with its strict adherence to GNU, and with the truly phenomenal number of distros based on debian (especially if you include ubuntu derived distros, but ubuntu is so different these days that idk if it’s really the same any more?)

Is it really so different though? It feels like it's been converging if anything, what with Ubuntu on systemd these days, and moving from Unity to Gnome Shell. What are the substantial differences at this point?

Actually, these days Ubuntu is closer to Debian than previously.

Ubuntu is Debian... repackaged

I wonder if the Fedora community could immediately fork off their own and essentially "leave" IBM/RH hanging? Not even sure if that is legally possible now, as the terms of the sale could have potentially included looming/upcoming license changes that might prevent that. I'd say they'd have to act quick and with their general feeling on whether they'd want to go that route (the Fedora community), and I'm sure their would still be legal challenges from IBM in any case if they attempted something like that.

There are no legal challenges and there is no hurry. You wouldn't be able to call the fork Fedora but other than that it's absolutely impossible for any shady backroom deal to affect your future ability of forking Fedora. The licensing of the collection as well as the individual components grants you that right and there is no way to take it back retroactively.

With the current state of things, forking Fedora seems unlikely to be a wise decision. RedHat is the major contributor, paying the salaries of many Fedora developers. So far, nothing has changed here with that acquisition.

All the buildings tools, backends, many devs are provided by Redhat. Without them I don't know how we would operate.

OpenSuse still rocks it.

> I'm curious what openSUSE folks think of SUSE having been acquired by Novel, then Attachmate, and then the Micro Focus merger. They've been through a lot, and openSUSE is still here.

OpenSuSE has been more like Fedora over their years; they historically never had a CentOS equivalent (although the newer OpenSuSE is moving there).

Regarding openSUSE (I work for SUSE so obviously don't speak for the wider openSUSE community, just my 2¢ as a contributor):

While people do have a reasonable level of hostility over the Novell acquisition (which has left some deep cultural scar tissue within SUSE), they did give us openSUSE.

Overall there is often worry when we have an acquisition (since a very large portion of openSUSE maintainers are employed by SUSE). With EQT quite a few folks were worried about how separated the finances were between openSUSE and SUSE and I believe Richard Brown commented on how exactly he's pushing for better financial and trademark separation (the only two things that they really share anymore).

So while people do get worried every one in a while, I get the impression that overall things are going okay despite the series of acquisitions in recent years.

However, Fedora/RedHat have a different structure and relationship and I wouldn't use the openSUSE/SUSE model to predict how things will work out.

Not a huge fan of Redhat, but I am a big user. I'm willing to wait and see.

Maybe we'll see more community offerings?

What is the relationship between RHEL and CentOS?

Never actually used RHEL, but I've heard that they're basically the same - one comes with the support, the other obviously doesn't.

Is CentOS supported by Red Hat, so IBM in theory could shut them down?

CentOS is the main competition to RHEL licenses. I have been in many meetings where the "use CentOS so if we have to buy RHEL its an easy conversion " was made

I assumed Redhat was going to capture lost revenue by somehow making CentOS less viable.

IBMs business nature makes me think its even more likely now.

From my perspective, it always seemed that RHEL's big problem was that CentOS was easier to use. Whenever I set up a machine, I always reached for CentOS because to me it was the same thing, except I didn't have to ask for a license, I didn't have to securely store the license, I didn't have to enter the license, and I didn't have to document my strategies for handling those things.

It's generally also more reliable. In any sort of an org like setup, RHEL will be managed through Satellite. The new Satellite seems like a massive pain. The old one also had issues. Running centos compared to registering clients to satellite is a breadth of fresh air. Sometimes I wonder if RH creates complexity, and sells these overly complicated solutions to problems they create.

I can't find the reference, but Red Hat recently hugged CentOS after sort of passively ignoring it for a while. Now, as I understand it, Red Hat has committed to actively supporting CentOS.

They're different products for different customers. Folks who want to work cheap and hack their own stuff together will use CentOS and would never have bought a RHEL license anyway.

Meanwhile, folks who need a "we've got support" answer for every question will use RHEL and wouldn't be tempted by CentOS anyway; the cost savings is not worth the risk.

redhat owns the centos trademark and employs most of its developers.

>Any bets on whether Fedora and CentOS will exist in November 2019?

I would say that Fedora and CentOS aren't going away anywhere. Not because of this anyway. There were similar concerns around RH's acquisition (if that's what you call it) of CentOS a couple of years ago, but things have largely been the same. And it's mostly for selfish reasons. The overall dev mindshare of RH based systems has shrunk compared to Ubuntu. So anything that moves people away from Ubuntu to the RH ecosystem is net win because eventually some corp will write a check when they need support. It's the same idea as MS not going after pirates just to increase MS's overall market share.

> I would say that Fedora and CentOS aren't going away anywhere.

That was the sentiment regarding OpenSolaris when Oracle bought Sun... (And I can't believe no one's mentioned this in 800+ comments so far.)

OpenSolaris came after regular Solaris. It was a retroactive opening of code, and wasn't as essential to regular Solaris. In contrast, Fedora is quite instrumental to RHEL's existence. CentOS is not that instrumental, but it's not a major cost sink either. It used to be community driven earlier and can become so again. i.e., Unless they go out of their way to do something malicious, which would have little payoff anyway.

> Fedora is quite instrumental to RHEL's existence

Could you expand on that? I'm curious.

Fedora is upstream for RHEL. RHEL is made up of several Fedora releases. Fedora is also the testing ground for RHEL.

Fedora and CentOS will both survive -- but maybe under a different name.

This is possible because everything in Redhat/Fedora/Centos is open source.

I'd be willing to bet numerous people are working on a non IBM/Redhat version of Centos.

That's the thing with an acquisition like this: what did IBM buy exactly? The code is open source, the people can leave and form a new company. The thing IBM really owns are Red Hat's contracts. But when those expire, the other party could sign their new contract with a company of former Red Hatters, if they want.

Buying an open source company only makes sense if you give the employees of that company exactly what they want. They are the real value.

It takes a lot more than a few to provide support. It would take years to get near the support that Red Hat provide now.

Theoretically, it seems all Red Hatters could quit simultaneously and form a new company according to the same structure they had before.

Of course in practice that requires a lot of coordination, there's a ton of legal hoops to jump through and you end up with a big company built around customers it doesn't currently have.

> Red Hat shareholders just sold out. Goddamnit.

RedHat is a public company, in what fantasy land do you exist where this isn't expected?

Furthermore, the deal is still subject to shareholder approval:

> The acquisition has been approved by the boards of directors of both IBM and Red Hat. It is subject to Red Hat shareholder approval. It also is subject to regulatory approvals and other customary closing conditions. It is expected to close in the latter half of 2019.

So you really should be complaining that RedHat's board of directors just sold out, and that's their fiduciary duty.

Selling out is not necessarily the “duty” of the board of directors. It could definitely be in the interest of shareholders to stay independent.

Agreed. The board's fiduciary duty is to do what's best for the stockholders.

> Red Hat shareholders just sold out.

The Red Hat board were offered a massive premium on an already generously-priced stock. If they hadn't sold there probably would've been lawsuits left and right.

The deal is subject to shareholder approval. I think based on that approval, you'll be able to estimate how right or wrong you are about lawsuits had the board rejected the offer.

RHT is $20 billion in market cap as of Friday, and the offer is $34 billion.

As soon as I heard this news, I was looking to invest in SUSE until I found out they were just bought by a private equity firm earlier this year.

What's going on with Linux providers? Is Amazon really just dominating the space?

no, but most cloud providers are not opensourcing their contributions to linux.

Is that really true? The big three cloud providers all have commits to the Linux kernel:

~/linux   master ● $ git log --author=amazon --format='%h %s%n %ad, %an <%ae>' --date=short | grep @amazon | wc -l 206

~/linux   master ● $ git log --author=microsoft --format='%h %s%n %ad, %an <%ae>' --date=short | grep @microsoft | wc -l 1825

~/linux   master ● $ git log --author=google --format='%h %s%n %ad, %an <%ae>' --date=short | grep @google | wc -l 11283

Good on Google.

Besides the kernel contributions, Google has a long list of open source projects: https://opensource.google.com/projects/list/featured

For all the hates Google gets on HN, they have been at the forefront of supporting OSS.

Google develops Android. It's in their best interest that their changes to the kernel get mainlined... The internal APIs of the kernel tend to change a lot, and Linus doesn't really care about out of tree breakage.

That's how userland Android is.

https://opensource.microsoft.com/

https://aws.github.io/

https://opensource.fb.com/

https://netflix.github.io/

https://uber.github.io/

https://opensource.apple.com/

This list is not nearly exhaustive, but my main point is Google is hardly the outlier here.

Good on all the developers pushing for open source so much. The industry is in a much better place than it was 25 years ago.

SUSE has been acquired many times in the past several years (Novell, Attachmate, MicroFocus). In theory, EQT will help us "get back on our feet" in terms of self-sufficiency but how things will actually pan out is obviously still unclear.

[Disclaimer: I work at SUSE.]

Ubuntu (Canonical) just stared looking more attractive as they were trying to be an independent commercially-supported Linux offering.

Or, if you don’t like Canonical (and to be fair they do a lot less than Red Hat do), encouraging corporate users to sponsor Debian directly would be amazing.

I'd rather corporate users sponsor Arch. My guess is Canonical will be announcing they are getting acquired by Microsoft soon.

Except corporate users usually want support.

It does make me wonder though if it would ever be possible for a bunch of business savy open source developers could ever get around to creating an open source co-op. Something like the commercial version of the FSF. Build open source products with solid support contracts, and build/contribute to open source that way.

The organization would be owned by the very people building and contributing the code.

RedHat is the only company that I can see that really did everything in the open.

I'm not sure which one I feel worse about, Oracle buying Sun, or IBM buying RedHat? I feel that Oracle did some major missteps in their acquisition (for this I look squarely at OpenOffice, and their misshandling of it, although, the OOo community hated the Oracle acquisition from day one, which I guess might have made it a little like poison berries - no one would want to go near it).

Oracle completely ruined MySQL during the acquisition too.

> Except corporate users usually want support.

What prevents anyone from offering an enterprise support program for Arch?

Corporate users can barely get around to patching windows desktops and are happy with Redhat being so slow moving, they're not going to jump on the arch constant upgrade cycle anytime soon.

This so much.

It is not ideal for managing servers without better supporting documentation, but engineering efforts would be pretty minimal to make it work well.

What's it take to build an excellent Arch enterprise support team (because that's the only way I'd ever want to do it)?

Taking on the big ones on their own turf isn't going to work, I think. By that turf I mean supporting years old versions running in dusty custom data centers.

I'd suggest building an auto-upgrade system on top of Arch (or Alpine), and go for immutable infrastructure as the selling point. That's stepping on CoreOS's toes a bit, but I haven't seen any progress from that crowd ever since Red Hat bought them, so it'll probably get even worse now.

That way you can target AWS/Azure/$OTHER_MODERN_STUFF in a more focused way, and you won't be stuck on supporting months/years old versions of the OS.

> I'd suggest building an auto-upgrade system on top of Arch (or Alpine), and go for immutable infrastructure as the selling point. That's stepping on CoreOS's toes a bit, but I haven't seen any progress from that crowd ever since Red Hat bought them, so it'll probably get even worse now.

In that case, why not go all the way over to NixOS? They already have a more or less complete cloud stack with NixOps, the only problem is hardly anyone knows how to use it.

>only problem is hardly anyone knows how to use it.

That's a massive, show-stopping problem.

> That's a massive, show-stopping problem.

How so?

I agree that it's a nontrivial problem to learn a piece of complex software from scratch to the point that you can offer comprehensive enterprise support for it, but how is that show-stopping?

No, I mean NixOS/NixOPS being incomprehensible to a large amount of us even after several years of existing. That's a huge problem, and it's their problem.

I'm not sure what the cause is. If it's a fundamental technological problem, then it's insurmountable, but if it's just a documentation problem, then it might be fixable. It doesn't look good though, since, like I said, the whole stack has existed for a long time. You would think that something would've been done about it.

All that said, perhaps I'm just dumb. I'll be happy to see someone prove me wrong and succeed with that combo. The underlying technology is certainly interesting and powerful. But I won't be trying it.

Well, at least my shares of DEBN just shot up.

Which went up because.... it's now seen as an acquisition target as well?

Edit; ah sorry I didn't know...

I think mindslight is joking; Debian is not a company, it can't be bought.

Hopefully this will cause an uptick in Gentoo (and OpenBSD)!

Yup, since I switched to FreeBSD year or more ago it makes watching news like this much less stressful.

Curious why Free>Open, for your usecase, as I'd like to learn more about them.

Very simple reason, zfs, checksums and scrub. I would love to use open but I was already bitten by bit rot on ext4 on linux.

And Arch Linux.

I love arch, I was using it for over 10 years in a row, but it uses systemd now 'owned' by IBM.

May be worth looking at:

https://artixlinux.org/

There is also Gentoo, Alpine, and Void.

Yeah, 10 years down the line RHEL is pretty much guaranteed dead. Probably say goodbye to Fedora as well.

This is sad, Fedora has been my favourite distribution for a little while. Red Hat have been a good community player and it is a big loss for them to be gobbled up by IBM. What is the best alternative to Fedora and CentOS?

Debian is solidly community based. Ubuntu depends on Debian in order to add its own value on top.

The obvious alternative would be Ubuntu. The LTS releases are alternatives to RHEL/CentOS. The others are Fedora alternatives.

However, Red Hat is _far_ more than RHEL. They actually build a ton of OSS that's used in RHEL and elsewhere. Ubuntu does little more than repackage Debian. That's not a criticism. They package well, and they know how to polish. But there's no replacement for Red Hat as a company.

It's so weird because as a long time fedora user I absolutely detest Ubuntu. Can't stand how bloated it's become.

Try vanilla Debian. I really don't get the fascination with Ubuntu - modern Debian does all the same things, and in many cases it does it better (saner defaults, less opinionated).

And if one really needs a "one click out of the box" desktop distro, then Mint.

My guess is that people tend to favour the OS they ‘grew up’ with.

Ubuntu made big inroads on the desktop in the early 2000’s, and now those people are running their own infrastructure.

Mint is based on Ubuntu. There is a version of Mint based on Debian but that is experimental.

LMDE has been around for several years now. IMO, it's "experimental" in the same way many Google services are "beta". Which is to say, in practice, it works just fine.

you say opionated like it's a bad thing. Opinionated is exactly why I use Fedora.

It can be a good thing or a bad thing, depending on how closely it matches your opinion.

Conversely, non-opinionated might not be good, but it's guaranteed to be not bad.

I agree with the first, but on the second, I'd rather switch to a different system that more closely matches my opinion.

I've been doing this crap way too long to want to configure everything myself.

That's why we have Debian!

Obviously Debian always will be there...

OpenSuse, they've shown an ability to deal with corporate buyouts and survive. They also continue to support the open source community.

NixOS due to its excellent community and quality.

DragonFly BSD

I'm sticking with Fedora - too early to tell.

I only used tried-and-tested distros like Yggdrasil.

Too early to say what will happens with Fedora and other RH products.

But Red Hat/Fedora has been a huge contributor to very basic Linux ecosystem things which everyone benefits from, like the freedesktop standard stuff and more.

I wonder how this change will pan out across the entire Linux-sphere. I’m wondering and I’m trying not to be pessimistic.

public companies are beholden to their shareholders. Enough shareholders liked the premium they were being offered...

It's not good, but it's the peril of most public quoted companies.

Public companies can have multiple classes of stock with different voting rights. Private company can be sold.

Sure they can, but most public companies (absent places like Facebook, which IIRC has a setup which favours Zuckerberg heavily) are more susceptible to being forced to sell whether they want to or not.

Private companies (especially one's who aren't beholden to VCs or similar) have the option to say no, regardless of how much money they're offered.

I always find this language kind of interesting. Companies are inanimate entities, they can't want or not want something. When you say "they want to" or "have the option to" who do you really mean?

Pretty sure the gp is referring to employees and top management, rather than shareholders. I.e. the actual daily functioning portion of the company.

I think it is more like Frontier buying phone lines from Verizon so they could temporarily reduce their free-fall in revenue.

Two or three years from now IBM will be the next GE.

> Red Hat shareholders just sold out.

Might be going out on a limb here, but maybe that’s why they bought the stock, so they could sell out at a profit?

Defeting aws not ms. And projects like ansible and cloud formation / manage iq are the weapons

> The cognitive dissonance is so strong here.

Where? Can you give an example of an expression of cognitive dissonance here?

Reply on news.ycombinator.com