Red Hat's open source rot took root when IBM walked in What is Red Hat thinking? It's not Red Hat's fault, I blame IBM for the company's recent Linux code licensing changes and all the misery it unleashed. While I know about Red Hat as well as any outsider can, I'm not an insider. I should have taken co-founder Bob Young up on his Red Hat IPO stock offer in 1999 instead of …
I always understood that RedHat's real value wasn't the source or the binaries, it was the support that they offered and where they get their revenue from.
Which is why this move makes no sense to me. A company that I worked for used CentOS as a testing ground for products to be deployed to customer's RedHat instances and I suspect that a lot of people do that also. I know that RedHat allows free development licenses but licensing is a royal pain in the ass when you are developing, spinning up VMs and the like. But then we didn't require the support. From that perspective, CentOS was an advertising tool for RedHat itself.
Makes no sense.
> and I suspect that a lot of people do that also.
All our prod and even UAT Boxes and VMs ran on licensed RHEL for 10 years straight, while developement, test environment and container images were based on CentOS.
Moved Dev/Test/Container stuff to Rocky last year to keep RHEL going in prod.
Now we are working actively on an exit strategy.
RedHat/IBM shot their own foot with this IMHO pretty dumb move...
The same here. We run Red Hat (and thus contribute to Red Hat's coffers) for all our production servers. We did that because of the support. If there is a problem, Red Hat would (probably) be able to help.
Pre-production stuff we're happy to use CentOS purely on the basis that we could spin stuff up in virtual environments, make a mess, roll back, etc etc etc without having to worry about the licencing given that we would only need to take care of that once we went to production. But - as I said elsewhere, for us the parity between the pre-production distie and the production distie are paramount in importance!
Yes, it now sucks, and it means I now have to migrate to something like Rocky for the time being, but hey... that's life in pre-production land.
I would be happy to contribute dosh to Rocky or others, but given that when it says 'Free' on the gubbin, our management go 'we're not in the business of giving away money' and definitely put the kibosh on contributing a penny or two to cover Rocky's costs.
Same for us. We develop special software for embedded/rugged use, and we used to use CentOS for development and licensed RHEL for customer installations. Our own DC was approx 70% RHEL and 30% SEL, and a number of CentOS workstations and servers for development work.
Today we're almost 100% SEL and openSUSE. There are a few remaining systems on RHEL but once support runs out they'll be migrated to SEL or retired. And the purge for RH also includes other RH products than just RHEL.
We went with SUSE because they are the #2 enterprise Linux vendor after RH and since we were already using it we had experience with it. Also, for the community clone openSUSE, SUSE actually went the opposite way to RH and some time ago moved the openSUSE Leap (LTS) train to the SUSE Linux Enterprise repositories. Like RH, SUSE is also a major contributor to Linux and many FOSS projects, and while that's not a decisive factor it helped us to make the decision.
I think it's more about Oracle than anything else. Oracle has leached lon the downstairs work, not contributing and offering undercut support for RedHat's own services. Which is really sleazy.
Being IBM, they're cutting off their nose to spite their face and limping all non paying users together.
That said, I avoid anything Oracle or IBM touches as a general rule.. I've been very happy with Debian and Ubuntu Server.
As a contrast, Canonical (Ubuntu) has played much nicer with the community and is both more and less successful than RedHat as a result. Outside of big business circles they are fast more broadly recognized and used. They are also much less profitable..
>> As a contrast, Canonical (Ubuntu) has played much nicer with the community
No, they are not. In fact, Canonical even gives back less to the Linux and FOSS community than even Microsoft does.
>> and is both more and less successful than RedHat as a result.
It's successful with consumers because they had great marketing (the "Linux for humans" schtick) and because they are based on Debian (everyone knows Debian) - for whatever that's worth.
Canonical also made some inroads with businesses which didn't need the level of stability (in terms of packages, not system stability) and who didn't want to pay what RHEL and SUSE were asking for support - although even with Canonical the old adage "you get what you pay for" still holds true.
Oracle actually contributes quite a bit to the Linux kernel. They are very active in XFS and btrfs development, and (briefly) held the role as top contributor.
https://blogs.oracle.com/linux/post/oracle-is-the-1-contributor-to-the-linux-kernel
Oracle is actually more expensive than Red Hat's lowest support offerings, charging $499/year for basic support, and $1,399/year for premier. Both tiers allow 24x7 access to file service requests (SRs).
https://www.oracle.com/linux/support/
Red Hat has a more complicated support structure, starting with workstation-self support: $179, workstation-8x5 support: $299, server-self support: $349, server-8x5 support: $799, server-24x7 support: $1,299.
https://www.redhat.com/en/store/linux-platforms
Oracle is substantially undercutting both of the server supported tiers, but the question is somewhat more nuanced than at first glance.
Oracle has also indicated that they will never conduct hostile software audits over Linux licensing. That is a big reason to avoid Red Hat, as those audits are unpleasant.
What Oracle is doing to RedHat might be "sleazy", but isn't it the essence of open source - and the core problem with open source as a business model?
And in any case, is what Oracle does to RedHat any different from what RedHat does to upstream Linux? RedHat's basic position is "everybody upstream from us provides their work for free, nobody downstream from us gets our work for free".
We're an academic site and have access to a RHEL site license, we have RHEL installs, some other groups we work with within our institution also have RHEL installs. Others have CentOS (they could equally have RHEL). RHEL is to some degree a pain dealing with channels in this environment (hey, is that compatibility package we need in a server or a desktop channel? is it in a developer one?), making CentOS easier for some cases. RedHat support itself has rarely proved itself useful to us, access to some of their customer only solutions has sometimes been helpful and sometimes not moved us forward with a problem. Generally we've spent our on time hunting the internet and figuring out solutions if needed. The subscription is not our choice as a department, I suspect we have it for institutional infrastructure reasons.
What we also have are deployments like RocksOS for HPC use. This is CentOS based. Do I see them going to Stream or back to a directly RHEL-derived basis (bear in mind package manage and channel issues above)? No I don't. We have plenty of packages we use created by the rest of the community. They target Debian and CentOS; conveniently we use RHEL because it's CentOS compatible, not CentOS because it's RHEL compatible. It's becoming ever less of a useful desktop system, I've written here before about how they are dropping the components that support remote use without any suitable replacement (sorry, VNC doesn't cut it). Occasionally in normal use a component will just go to 100% CPU and the bugzilla solutions for this tend to be along the lines of kill it or delete a certain file every so often.
My own journey to using Linux (and partly to doing what I do now) started with a brief flirtation with Mandriva as a student. This was quickly discontinued, but later I found myself using RedHat Linux (before it was RHEL) in the lab for computing projects. I thought it was convenient to set up my own dual boot system. Eventually they dropped RedHat linux and replaced the "community" version with Fedora, I moved to that (also now being reconsidered). When I started with my current employer having experience of linux was an advantage. They were using slackware at the time (this isn't quite as long ago as it sounds...) and then moved to RedHa... no, they moved to Ubuntu. Later on I encouraged taking up RedHat which is were we are now. I've always liked developing on Linux systems because access to tools is so easy, I remember experimenting with programming when i was at school and being hampered in part by the cost of tools. So how do I feel about starting new students on an environment controlled by an organisation that sees them as hobbyists and freeloaders? Not great really.
What if a bunch of RHEL customers would be willing to pay a REALLY CHEAP license, either one time or subscription, that gives them "no support" (like using an RHEL clone) with the benefit of having the actual RHEL software (and not a clone) ?
There is no need to be greedy and demand ONLY the BIG BUCKS license unless, like an insurance company, they expect to drive revenue from customers who do NOT use the support...!
Then, just to have even MORE fun, they could offer users of the clone versions an option to get support at the RHEL support price. Some might want that.
"Red Hat was, and is, leaving a lot of money on the table."
But is it? Consider some of the possible scanarios for different customers if the clones closed down.
1. RHEL user for both production and test, training etc - let's call these non-production uses ancillary - no additional money made.
2. RHEL user for production, clone for ancillaries but would be able to afford to convert the latter to RHEL licences - money on table here.
3. RHEL user for production, clone for ancillaries but converting latter to full licences would not be financially affordable , will manage to struggle along with late version clone/CENTOS-stream or some other distro - no additional money to be made.
4. As 3 but decides to use another distro for future projects - no short term gain, likely long term loss as current production purpose reaches EoL.
5. As 4 but decides to actively migrate existing production usage so as not to split work between two distros - complete medium to long term loss
6. As 5 but reluctantly needs to buy extra licence for ancillaries during migration - short term gain followed by loss in medium rather than longer term.
7. Running entirely on clones, could afford to buy licences - money left on table.
8. Running entirely on clones, licences would be unaffordable - no money left on table irrespective of what course of action they take.
9. Not currently RHEL users but were considering it until now - loss of future sales prospects
Those in scenarios 3 - 6 currently using some S/W or H/W product currently RHEL users are likely to be speaking to the vendors of those products in the near future if they aren't already and said vendors likely to be considering their positions already.
So in some scenarios there will have been money left on the table which they could pick up. In others there's no prospect of that happening and in others there's money to be lost in the longer term, especially as 3s slip into 4 or 5.
Whether this is a real money earner in the long term depends on the balance between the scenarios.
My guess would be that there are bonuses to be made in the next quarter or so as the immediate gains are made after which it will be time for the execs to emulate the rest of their customers and move on.
Your post is a longer, more complete version of what I was thinking about this move: that of the many possible scenarios, IBM has fixated on the one that would be a win for them. Given my (perceived) understanding of the FOSS mindset, that scenario does not seem like the one that's most likely to prevail. But after a century of selling stuff, it perhaps seemed obvious to IBM. Remember that "Nobody ever got fired for buying IBM": clearly [to the suits in Armonk] pushing folks in real businesses to use real RH with real RH support *must* be a win, right? /s
The deeper issue here is that unless you and I are both completely wrong about the results here, this shows how completely IBM misunderstands the end of the industry that they bought into with RH. I'm sure there were plenty of IBMers who could have explained it to them--a decade or so ago; they've all retired or been chased away (or RIFfed) by now, alas.
Quite a few people have been fired for buying IBM.
The situation isn't even as complex as your post might imply.
Even if a company actually does have the margin to afford buying more RHEL licences, they won't want to because that eats into their own profits.
So now every single RHEL customer is weighing up the ongoing increased future cost of more licences against the one-time cost of migration - whether now or at their contract renewal date.
And unlike IBM mainframes or Microsoft, there are several other entities very happy to take their money and help them migrate.
My employer has over 20,000 Linux installations. We have a lot of expertise inhouse so saw no point in paying RHEL for support, and ran everything on CentOS or Oracle Linux. We had a handful of RHEL licenses to allow us to access to bugzilla, which is probably the type of customer RedHat classes as a freeloader. However, we included full patches to fix the issues in all of our bug reports. We contribute multiple man-years and significant cash to upstream open-source efforts every year. We've never really had any success engaging with RedHat's support. Engaging with upstream authors such as kernel devs on LKML always got us better results.
We recently cancelled our RH contract and converted everything to a clone as upon renewal, as the new terms included a clause saying we needed to also purchase a license for any boxes using commercial RHEL clones, such as Oracle Linux. This was hidden in the small print but luckily our legal team spotted it. RH then quoted an 8-figure sum to license all of our boxes. For much less than amount of money we could just hire more people and do support internally. We've seen the direction that the wind is blowing and are already planning to migrate everything to a non-RHEL-based distro. The motivation is not just licensing - we've become increasingly disillusioned with RedHat's technical decisions. They make lots of breaking changes in minor upgrades lately, even things like the kernel source module API. The whole point of using an enterprise distro is for API/ABI stability and no breaking changes within a major version! They also seem to be behind a lot of the projects to make Linux needlessly complex and confusing - systemd et al which just make our jobs harder.
Once final thought: If RH/IBM are so against freeloading, why don't they stop distributing GPL software written outside of RH and rewrite GNU/Linux from scratch inhouse? It appears to me like they want to have their cake and eat it. They don't want anyone "freeloading" from them, but are happy to freeload themselves.
> I blame IBM
As do I, but everyone knew some kind of crap would come from from the sale And Red Hat still sold, and everyone working there (that wasn't laid off) still works there, etc.
Only so many excuses can be made on their behalf. The fact IBM now owns them doesn't absolve them of anything at all.
Red hat sold to IBM knowing the risks. Red Hat are now IBM. Ergo, I blame Red Hat and IBM.
This just isn't true. Red Hat is managed as a separate entity. It is controlled by IBM, sure, but it hasn't been absorbed the way other acquisitions (eg Rational) were.
You won't find an IBM logo anywhere on Red Hat's website except where IBM branded products are being integrated or distributed.
Red Hat was already doing bad things before IBM came along. Red Hat has arguably been on a downhill path since they took away the original free Red Hat Linux.
IBM makes things worse, of course -- Inevitable Bad Moves.
But to act like this latest mess is all IBM's fault is giving Red Hat management too much credit. They aren't victims here.
I blame IBM too. They should not have paid anywhere near $34bn for RedHat. Now as they realise they can't actually grow that revenue in any sensible way. RedHat's other products don't have the traction they made it seems like, 'OpenShift' I am looking at you! They now are trying to sure up the code base so they think they can make more money from previous users of Centos etc. When in fact those users will just move to another distribution.
RedHat's site has for a number of years made it look like you are paying for a license or seat for Linux when you are just paying for support. You can't license something that is ultimately Open Source or composed of open source. The other issue is that your banks and big businesses buy into the idea that there are paying for Linux licenses when in fact they are not. The number of organisations I see that pay full support for dev/test instances is insane.
https://discussion.fedoraproject.org/t/f40-change-request-privacy-preserving-telemetry-for-fedora-workstation-system-wide/85320
*Proposal* to use a telemetry system developed by the Endless OS people to gather data to improve yada yada.
I think the hatters are pretty good at being controversial all by themselves without IBM.
The directors rarely consider what engineering would consider objectionable. The directors will consider what shareholders want. And, as we've seen with the likes of Twitter, despite objections from hoi polloi in the cube farms, the management and board insisted on selling. Same here. If you're a shareholder and you get quite a nice premium on your shares by selling, you might vote for the sale.
Of course, individual shareholders tend to be inconsequential even if it comes to that. It's the institutional investors (pension funds, hedge funds, banks and the like) who chase after the quick buck, and when they can realise a decent profit on their shareholding, individual shareholders be damned. I've been there several times in my life as an individual shareholder.
"The directors will consider what shareholders want. And, as we've seen with the likes of Twitter, despite objections from hoi polloi in the cube farms, the management and board insisted on selling."
In Twitter's case, the directors and the board both attempted not to sell, adding extra blocks to the process. It was only after shareholders demanded that they change their mind and get them the Musk payout that they accepted it. Of course, when Musk allowed it to break down, they pursued the sale he had agreed to, but that was also something the shareholders would have made them do anyway. Sometimes, the management doesn't want to make a company crash, but if the people who own the company want the money now, they have the right to get it and to make the management act for that goal.
> There was never any real chance that IBM would get it.
> That must have been obvious before the takeover, so in that sense at least, it is Red Hat's fault, or at least their director's fault.
At that time I described it as a "drive by mugging by the Redhat C-Suite". They sold-out and b******d off with the money leaving their employees to face the consequences. I anticipated I would be gone within the year (being 50+), and I was right.
Not sure how so many people are so one-sided on the topic and paint RHEL clone builders as some great liberators
Competition in the open source Linux market is healthy - you have Red Hat, Suse, Canonical, Amazon, and recently Microsoft, with their own range of Linux offerings.
All of these companies share open source Linux code to build their offerings, and hire and pay lots of people to work on the engineering, test. documentation and ecosystem building of their products. This allows open source Linux offerings to be more widely used and to help the technology advance.
The Linux model enables Joe Blow Linux or Rocky or Alma, to start their own Linux version and do the same as the others above have done. But Rocky and Alma have chosen to just clone someone else's version, where that company invests in making the offering, and Linux overall, better. There is little to no value in what they do, and I think it is shady. But for some reason people paint them as some sort of hero's. Not getting it. Many Red Hat paying customers are likely angry when they see peer companies in their industry, not paying Red Hat for RHEL, but getting the majority of the value of RHEL for free.
Articles like this seem to take that same one sided view, and hark back to days of Bob Young at Red Hat, when Red Hat was losing money and had less than 1,000 employees as some sort of panacea...Don't get it.
> The Linux model enables Joe Blow Linux or Rocky or Alma, to start their own Linux version and do the same as the others above have done
The "linux model" is version 2 of the GPL, which permits everyone to do anything with the software except deny anybody else the same right to do anything. If you don't want people to take your stuff and wrap it up in a box to sell, don't give it away under a license which explicitly permits them to do exactly that. Being able to do that is the *very purpose* behind the GPL and if IBM don't like it they can go and play some other game.
> Many Red Hat paying customers are likely angry when they see peer companies in their industry, not paying Red Hat for RHEL, but getting the majority of the value of RHEL for free.
Nobody is stopping them from hiring people who can run a server without an American corporation to hold their hand.
It is relevant for a lot of their customers though - American companies will prefer to buy from other American companies(*)
Is it relevant in regards to if companies will pay for a sysadmin team that can manage it themselves rather than rely on support contracts?
harder to say...
Footnote: A few $JOBs ago, we had ecommerce clients from around the world.
Nearly all in the US were running IIS, most everywhere else were Apache (nginx was only starting to come in)
> If you don't want people to take your stuff and wrap it up in a box to sell, don't give it away under a license which explicitly permits them to do exactly that.
Or rather, if you don't want people to do that with "your" stuff, don't base "your" stuff- and your business- on the efforts of others who let you do so only on the condition that you were required to do that.
Red Hat was built on that principle. They could- in theory- have built their own OS from scratch and released it under a completely non-free license.
But in practice, they wouldn't and they couldn't, and Red Hat only exists in the first place as a result of freely-licensed code.
Now that its requirements work against them more, I'm sure that Red Hat's current owners- i.e. IBM- would like to pull up the ladder behind them *and* have their cake and eat it. Tough.
The fact that Red Hat contributes a lot of code to Linux doesn't entitle them to change the rules to suit themselves.
If they don't like that, no-one's stopping IBM from- again- writing their own completely proprietary, completely closed OS from scratch, unencumbered by the requirements of open source licenses.
In a nutshell.
It appears that business decisions aren't made on the basis of long term health and profitability of a company. But rather the demands of equity traders who make their money on the basis of shirt term share value. In effect liming the soil.*
*For non-gardeners, putting lime on the soil releases nutrients quickly. But it also depletes the soil by releasing the nutrients without replenishing them. In the short term you get a good crop, but long term you can end up with very poor harvests.
Where did you read they shutdown access to the source code ?
It says in their blog and in several articles that document the reality of what happened, that the source is now available as part of CentOS Stream or if you get a free Developer Subscription or buy a Subscription. You can look at the source all day long, and with two options where there is no payment required (CentOS Stream or a Developer Subscription).
Yes they stopped one of the options of how source code was made available, but they did not "shutdown" access as many people seem to be selectively sound byting.
> but they did not "shutdown" access as many people seem to be selectively sound byting.
You are being disingenuous. You have access to source to RHEL if you subscribe as long as you don't distribute, otherwise you will be unsubscribed and then you will lose access. CentOS Stream is *not* RHEL.
This is where the Red Hat defenders are getting it wrong IMO. The majority of the value of what RH provides exists in terms of professional support. You don't get that using a clone. The majority of the value of the OS was not created by RH, they didn't write the kernel, the user shells, the desktop ........ all these are open source and were taken by RH from the community, for free. And the condition of doing so was abiding by the GPL.
> The majority of the value of the OS was not created by RH, they didn't write the kernel, the user shells
They are in fact one of the biggest contributors to the kennel and other parts of Linux. I'm not giving them a pass on this, but don't under cut their actual contributions.
I have chosen other distributions so along as I don't like a lot of their business decisions and IBM absolutely should be avoided.. The same for Oracle.
"This is where the Red Hat defenders are getting it wrong IMO. The majority of the value of what RH provides exists in terms of professional support. You don't get that using a clone. "
It is clear that the Red Hat distribution has value by itself, otherwise people wouldn't bother cloning it, and you wouldn't have these howls of outrage when they try to stifle the clone builders.
Red hat support org is mostly useless. The value of rhel is the quality of the release engineering ( don't laugh too hard) and it's stability. With any other platform packaged software would change willy nilly. Eg Apache 2.2 would suddenly become v2.4 because they felt like it. With rhel it was still 2.2 with certain things backported. Which is its own problem, but not normally a showstopper.
Red hat got GREEDY many years ago!. $200 one time became $79/yr/host. Then it was 199/yr/host. Then 400 and even 1200+
There is no frickin way you can justify that kind of price gouging.
THAT is why CentOS etc became so big. Nobody could afford Rhel everywhere any more. Hell some people went back to friggin Windows because it was cheaper and by a lot.
Redhat killed the golden goose a good 10 years before ibm came sniffing around. Their own stupidity MADE Conical and Suse the players they are today.
I worked for Nielsen Media. We had ONE copy of Rhel for the purposes of scraping the patch repo, and a 1000 machines installed. That used to be legal. Rather than start paying redhat millions we pivoted to CentOS. Had redhat offered us $50/host/yr we would have taken it in a blink of any eye.
Red hat mgmt has fundamentally been just stupid. They wanted to get rich. They sold out to get their pay day, KNOWING the cancer that is IBM would destroy the company soon enough. Now the devil is here to collect.
Yes this is ALL about screwing Oracle as hard as they can. These 2 hate each other with white hot hatred. They really should just go get pistols at dawn and shoot each other dead and be done with it.
The trivial answer to this tightfisted money grubbing is incredibly simple. $50/host/yr subscription to the repos. 3yr qty discounts, Org level bulk discounts. And completely separate support contracts for actual support.
When running a rhel box is as cheap as bottled water, nobody is going to bother making clones.
Red hat is perfectly within their rights to prohibit further distribution of their binary and src rpms. If Oracle wants to get into that game, they can give redhat 100 million a year.
Because of CentOS RHEL was the default distribution for third party software. If a company needed third party software they choose RHEL because you knew it was either your only choice or was the primary platform for the software you needed. That's not true anymore, everything is available on Ubuntu and in many cases Ubuntu works better. I used Fedora and CentOS for over 20 years, all of my new systems run Ubuntu now. The older hardware with CentOS7 will continue to run it until they are decommissioned. The same is true for my clients.
I remember years ago being quite a Red Hat fan. I used Fedora personally and felt that it was a good choice because Red Hat were a great company who had given a *lot* to open source.
(also I did find Ubuntu's new releases to be crash-tastic most of the time, and Fedora to be quite stable, this outweighing the effort required to enable patent-encumbered things on Fedora).
I remember reading about how many projects/companies and how many millions of dollars Red Hat had spent acquiring companies and then open-sourcing the acquired technology. Seemed like a good idea to be a Red Hat / Fedora guy.
Feels like things have changed a bit.
No sane person will use Stream or Fedora for any kind of prod business app. Not even the simple Django+Vue tools that I have started to chuck out for free here and there. The OS support lifecycle is too short/non-existent.
... and discovered Slackware.
Imagine, 30 years (in 10 days) of politics-free, board-room free computing that just works. It could have been yours, too.
There is no reason it won't last another 30 years. Is there any reason you shouldn't start now?
slackware.com
Slackware is where I learned to not fear the tar.gz file and build from it.
But - that's where it ends. Given that many Makefiles don't have an uninstall option (this is where RPM and DEB are superior), Slackware sadly doesn't quite cut it in production for us.
My predecessors went with RPM-based stuff, so that's where we'll stay. :-/
"Given that many Makefiles don't have an uninstall option (this is where RPM and DEB are superior),"
Makefiles don't. But Slack Packages do. And have done for a long time. Since Slackware 9.1 back in 2003. Earlier if you were a Slack developer.
"Slackware sadly doesn't quite cut it in production for us."
Maybe not, over 20 years ago. Try it again, it might surprise you.
I like Slackware. As I like FreeBSD and OpenBSD. Minimal assumptions made, plenty of room to experiment and learn. But for the past few years I've been using the Mandrake branch - PCLinuxOS is its coolest branch IMHO - and then switching to a dual setup of Kubuntu and Fedora, because most of the people I met using Linux professionally seemed to be gravitating to those two distros.
Wouldn't mind running Slackware again.
> It could have been yours, too.
Indeed, it was (first Linux distro, I had played a bit with gnuwin and cygwin(*))
30 years... Nostalgia sure aint what it use to be
I didn't start at the start, but it wasn't much later - now I'm wondering which version I started with..?
Initially thought it might be 3.0, but I think I remember compiling to a.out files as well, so could have been 2.3...
Ahh, the days of running off of 2 floppies (because the hard drive had windows installed) - one to boot (with kernel) and one for userspace (including X86, I think - or maybe that was later and a 3rd floppy?)
I learnt a lot from that, but eventually got tired of having to figure out changes to config files (particularly X.conf) most updates (config files weren't managed - if you unpacked the one in the .tar.gz over the one in /etc you had to reconfigure it - and I hadn't yet figured out diffing and merging local changes with upstream)
Footnote: I thought I had played with cygwin for a while before slackware, but wikipedia tells me cygwin came out in 95, so maybe it was just gnuwin before and cygwin in parallel...)
Honestly... how clueless is IBM these days?
They were the biggest leaders of the hottest tech in the world and they flushed all that down the drain.
They couldn't arrange a piss up in a brewery. I certainly don't expect them to understand open source or the nuances of a third-party developer community around a product.
"All companies are ephemeral."
Well put, I'm going to use that. Much shorter than my usual description.
I've argued for decades that any solution implemented (hardware or software) has to have an exit plan -- how do you move to another platform, if your company outlasts your suppliers? How do you extract your data and reimplement on another platform?
Of course, most people (including managers) consider employment temporary and their current employer as a place to stuff things on their resume for the next job. If the supplier fails, you just go job jumping a little sooner than expected. Try to look at things long term, you are branded as "not a team player".
I am from the other side; that is, I worked on the IBM software side. I am retired now and no longer have to put up with IBM. Sometime in the 1990's IBM did a 180 in customer relationships. IBM charged for what used to be free. I would ask a question about hardware/software from my friendly local IBM person, and he would get back to me usually within 24 hours (longer if it entailed talking with someone else at IBM). All of a sudden, those IBM people disappeared. When I approached IBM for a question on their hardware/software, I was told to call a 1-800 number. When I did so, I was informed it would now cost a thousand dollars for what was once a free question. We were interested in buying a larger mainframe, and yes, they charged us to order a large mainframe. I am happy I retired and do not have to work with IBM.
In M&A situations, it is not unusual for the acquiring party to destroy much of the value they saw in what they acquired. Often it is because of the imposition of mindless bureaucracy on something that was successfully dynamic. Often it is because the acquiring party makes their problems your problems. Finally some mindless bean-counter or HR idiot will say "oh our costs just went up, we need to make savings and RA people" - blow the brains out of what we purchased. Having worked on some Italian projects in the 90s the general saying was "never underestimate their ability to disorganize you". Same principles apply here.
Red Hat was, and is, leaving a lot of money on the table.
That's purely imaginary money... You'll find when you try to FORCE free users to pay full price, they instead quickly cease to be users at all.
And more to the point... Red Hat is in the "free users" tier as well. They built their product upon the work of many thousands of others, which they did not have to pay for. They would not exist otherwise.
No, I don't think Big Blue's top brass gets it.
That's a difficult argument to sustain, as you've just listed several (but far from all) of the previous times Red Hat (before the IBM merger) did stupid things like that, too. Perhaps it's true, but we don't have any evidence to that effect.
While I agree with your analysis that the dictat came from above (IBM). I disagree that it was against clone in general. I think it was directed at one clone specifically, Oracle. IBM are in direct competition in cloud hosting with someone who is not only offering a comparable product, they are offering Their product. And undercutting their prices for it! That situation could only be allowed to continue for so long, and Rocky and co are just collateral.
Quite possibly. So instead of not careing about the OS part, since there are plenty of good non-redhat free competitors (being open source and all) and selling other applications on top of it, and creating a new cloud business to rival microsofts, oracle, and aws, they have really pushed the community away. Ubuntu has been making inroads to developer desktops for a long time in businesses that pay for redhat licenses, and those, like me, who ran rhel, or alma, or centos, or.. have been noticing that Ubuntu is becoming the defacto standard Linux and is, in general a better supported platform. This time, the redhat users didnt just scream, many have left. The writing is on the wall. This isn't just a bad decision, this is here to stay. Its time to drop fedora and redhat, and move to debian and ubuntu. My employer still pays for a redhat site license, and with it some redhat products which are not operating systems, but the mind share of wanting to be a redhat shop is out the window. Apps are moving in to containers on debian, workstations are going back to windows or to mac. Now that microsoft has improved and been more open source friendly (including a pretty good WSL2 implementation with, you guessed it, Ubuntu being the defacto OS) All redhat have done is removed one of the major reasons for staying with them if you have made a career of it. Redhat, or IBM, really shot themselves in the foot with this one. Microsoft is the new Redhat, and Redhat is the old Microsoft. And apple.. still apple for better and worse.
I've been wondering for a while if Oracle was the target here, but I'm not convinced.
Oracle are a Red Hat Certified Cloud Service Provider, which was only announced in January '23.
https://blogs.oracle.com/cloud-infrastructure/post/red-hat-enterprise-linux-supported-oci
https://catalog.redhat.com/cloud/detail/216977
If Red Hat/IBM had concerns about Oracle I can't see they'd agree to them signing up to be a CCSP.
CCSP is an interesting point. Obviously, it doesn't just apply to the OS, but the whole suite of systems and apps. Under the current/previous rules, I can't see how RH could refuse them... If you think of that scenario the other way round, Maybe the fact Redhat issued CCSP to Oracle was the catalyst to this change. CCSP means Oracle can sell Redhat, but not Oracle branded Redhat from the source the licence grants them. All hypothetical ofc. But it seems to chime for me.
Well, it's for "individual non-business" use..
"The Red Hat Developer Subscription for Individuals is still only available to individuals, not organizations or teams, and is designed for personal servers, home labs, and small open source communities."
https://developers.redhat.com/articles/faqs-no-cost-red-hat-enterprise-linux
I've been using Ubuntu for years, if not over a decade. I used to use RedHat; I switched because they broke my installation one too many times with their updates. Now mind, this was over a decade ago, but still - RedHat is not the panacea some think it is.
Are they only violating the "spirit" of the GPL? I don't think so. Not by a long shot. But it'll be up to the lawyers to argue it out. To my amateur reading, the GPL is quite clear that source code has to be made available when you release the product, and that you are NOT allowed to do "private builds" of modified code and keep it hidden.
The legalese about "paying customers" is just smoke and mirrors by IBM's expert legal team to try and screw over the open source community, and just TAKE instead of giving back. Even Amazon is learning their lesson in that regard; IBM is walking backwards at a run. A pox on all things Big Blue. They can't finish dying off soon enough.
Curious where you see that Red Hat is not "making the source code available" anymore ?
It says in their blog and in several articles that document the reality of what happened, that the source is now available as part of CentOS Stream or if you get a free Developer Subscription or buy a Subscription. You can look at the source all day long, and with two options where there is no payment required (CentOS Stream or a Developer Subscription).
Yes they stopped one of the options of how source code was made available, but they did not stop making the source code available as many people seem to be selectively sound byting.
no, it's not the same. But RedHat as a whole isn't GPLed (or at least, I think that's their argument), it's all the individual pieces that are GPLed. You can get the source code to any of the individual pieces all you want. You can get past versions, current versions, release, beta, development, whatever the authors (not RH) make available.
Kinda like making a dance floor greatest hits CD. You acquire rights to all the pieces, hopefully properly, you put it together, you sell it, and you now have the right to get grumpy if someone just duplicates it and doesn't give anything to you. You don't own the individual pieces, you claim copyright on the compilation.
RedHat isn't closing the sources on the pieces, they are making how they put it all together a "secret".
And that's what the GPL was about -- freedom to use the software as you wish, as long as any changes the publisher made are made available to the users. If you buy RH Linux, you get that source code. If you don't buy RH, you aren't a RH user, and what's your beef? All the advances RH is making to the code are out there, you can grab anything you want out of Fedora or Centos trees. You want to precisely replicate RH v9.3? Well, that's what they have an issue with.
This has turned into a "free as in beer" rage, and for decades, the GPL people have said it isn't about free beer. And I get it. I'm not at all happy about what IBM has done, but I wasn't a fan of RH before IBM either. But I see their arguments, and I think they fit the letter of the GPL law, even if it isn't what all the free labor that has made RH possible desired.
Technically, just to the person with the binaries. However, that person cannot be forbidden from sharing the source or the binary using that license, so they can then publish it. Additional agreements can be used to get someone to agree that they have the rights to publish but that they will not exercise them, but that is more complex and must be created when they get the binaries originally.
People are going to disagree but I always think of HP/HPE and IBM as modern-day "IP trolls": They buy a large company and s*ck out (or live-off) the finance (of the newly acquired company) and when the finances run dry, they discard the carcass and move on to the next victim.
(The only time this has "buck the trend" is when HP acquired Aruba but I think this was pure luck.)
> People are going to disagree but I always think of HP/HPE and IBM as modern-day "IP trolls"
Which is odd because the IBM of old did sterling work fending off SCO
That's what this reminds me of, way back in the 80s. IBM cut corners to get their PC out, and that included using common parts; they also published the specifications so that people could build add-ons that IBM wasn't going to waste time building themselves. And they also published the BIOS source, which a few people promptly clean-room cloned.
After suddenly realizing they'd created a marketplace out of a hobbyist niche, they tried to close things up. They soon discovered that they couldn't sue the likes of Phoenix, maker of a clone BIOS out of existence, because the law didn't agree with IBM's views. So they developed a brand new PC architecture, the MCA (microchannel architecture) and tried to establish that as the new motherboard bus. Which led to a lot of kerfuffle and the EISA (extended industry standard architecture), yadayadayada. All long since deceased, gone to the great big bit-bucket in the sky.
What we've had has been similar - a new software "architecture" has taken over, moving rapidly from mere hobbyist to serious business underpinning, and IBM has tried to corral it. And they're finding it's not what they think it is. It enables IBM to maintain its massive investment in mainframes, since it works just as well on mainframes as on PCs and smartphones. But it's not their product; it's owned by others who are willing to share, but not to let it be corralled off in any way.
Which is in large part, why I have every now and then, mentioned IBM and Microsoft letting go of the OS/2 code base (and prior such as the VAX VMS and later, such as the early MS Win 3x-9X and WinNT 3-5), so they can also let go of some of those assumptions. Because at the rate they're going, they're spending more time spinning in circles than actually getting anywhere.
I have several issues with what you've said. IBM did not create a system for the hobbyist market, they were looking at the business niche that Apple had created. You are right that they used commodity parts, but that is mainly because it was not a mainstream product development, but almost what you might describe as a 'skunk-works' product created in spite of, not because of IBM general policy. The specifications you talk about IBM publishing were really easily derived from the components, even if the physical design and pin layout of the ISA connector was defined by IBM. But the ISA bus itself is really just the 8 bit Intel bus plus a number of interrupt lines.
IBM actually had form publishing the specs for their systems. The 370 architecture was well documented in one of the first red books before the PC came along!
Secondly, I don't know that IBM actually published the BIOS. It was easily accessible, being a ROM present in every PC built, and could be dis-assembled and/or copied easily. The clean room reimplementation of the BIOS is well documented. It did not come from the source, at least not directly. They had a team that analyzed the BIOS, wrote a requirements document based on the function of the BIOS, and then handed this to a group of programmers that had never seen the original code for them to write a new, functionally compatible BIOS that relied nothing on the original code.
IBM was involved in the UNIX market for longer that Linux has been around. AIX has been a big part of IBM's business for 30+ years. And when Linux became available, they embraced it. There are many IBM products that rely on Linux, for example the Hardware Management consoles used for Z and Power systems (forked from the original Red Hat Linux), and the Power hypervisor, which is also Linux. I think that IBM (or at least part of IBM) did understand Linux.
IBM has both UNIX ports and Linux on the mainframe for a very long time. But I don't think that they make a huge amount of money from Linux on Z, although I'm sure people run it and it drives Mainframe sales.
I think that IBM has bought Red Hat, partly because of the customer base, but also for some of the technologies RH controlled, and also so they could start closely marrying RH Linux to Power, to extend the lifetime of the Power platform as AIX slowly withers away. I believe that they will try to introduce the Power RAS features into a mainstream Linux release to extend Power's life in the lucrative financial markets that still like the Power platform.
But I will concede that IBM is far from the company that it was in the '90s and '00s. It is no longer driven by technology, but more like a financial and consultancy company that happens to still have some legacy businesses.
"they're spending more time spinning in circles than actually getting anywhere."
This is DEFINITELY TRUE in the Micros~1 'Windows World'. I do not mess with RHEL much (other than light experimentation with the clones) and so I just have to assume that this is also true for IBM/RH.
...it's merely a division / brand of IBM.
The real question is, what's gone wrong in IBM? They're a shadow of their former selves. The company has gone from being one of the most enlightened of employers globally one that gets embroiled in law suits related to age discrimination. They have been significant contributors to the Linux kernel (largely to ensure that it runs on their mainframe hardware). They used to lead, and now they lag, in pretty much all areas. They used to be the company that everyone wanted to work for, and now they're not.
I think a lot of it stems from a far too conservative approach to business strategy - no get up and go out there and change the market, their management simply started to rest on the laurels of the past and collect profits without thinking of the future. I wasn't surprised that IBM were attracted to RedHat - making money out of the past (stable old modified kernels) is basically Red Hat's entire business model.
Ex Big Blue and it's been going downhill since Lou Gerstner left.
"IBM wants to squeeze more profits out of Red Hat by making it more like IBM, cutting costs, and increasing RHEL licensing volume. It won't work.
This move will hurt Red Hat's culture and IBM's bottom line."
What will it do to share prices in the next 12 months though? That's what IBM management care about exclusively.
It seems to me like Red Hat is taking a bunch of things built by others, packing them up with their logo and then suing anyone who tries to copy their strategy too closely. That feels wrong. RH is building on top of the work of many open source developers but doesn't want to return anything to the community.
The facts show that they provide source code available in 3 ways, including two free ways. People keep sound byting that they stopped providing source code or returning things to the community.
t says in their blog and in several articles that document the reality of what happened, that the source is now available as part of CentOS Stream or if you get a free Developer Subscription or buy a Subscription. You can look at the source all day long, and with two options where there is no payment required (CentOS Stream or a Developer Subscription).
Yes they stopped one of the options of how source code was made available, but they did not "shutdown" access as many people seem to be selectively sound byting.
The IBM bashing that's going on misses out on a few important details from the history.
20 years ago, we all cheered when IBM deployed its formidable legal department to stop a severely misguided but dangerous effort by SCO (with Microsoft skulking in the background) to essentially racketeer users of the Linux kernel, threatened to sue anyone using it without a license. IBM said "not on our watch" and prevailed. IBM had correctly calculated, before Sun, Microsoft and others did, that the Linux kernel would become the standard compute kernel in the enterprise and that it would be better to embrace it rather than try to compete with it. Along with others, they poured resources into it, among other things porting it to their mainframe line.
I don't say this to suggest that any supposed IBM transgressions should be set aside; instead I think IBM's historic support for open source is something that needs to be considered when evaluating their motivations.
The difficult reality as I see it here is that the relationship between commercial interests and volunteer contributors within open source is a symbiotic one. It doesn't always work as it should (hello OpenSSL) but I think it it's a fact that Linux and the suite of applications around it would have nothing like the stability and quality they have without heavy commercial input. At the same time, the commercial interests which benefit from open source would be worse off without the community. I think everyone involves should try to find a pragmatic compromise.
Red Hat's focus on trying to make rebuilding their distribution more difficult did not, as the article notes, start with the IBM acquisition. They bought out CentOS, and then changed the way they released their patches. CentOS is not really the target here - competing vendors are. Consider Oracle : a large, wealthy and highly successful and profitable corporation, taking a distribution built - entirely legally - on someone else's investment and selling it through their own channels, while refusing to support, maintain and open source their own enterprise-class operating system.
I think there is a case to be made here that Red Hat have perhaps mishandled the communication of this. But at the end of the day, the people who are complaining are people who want the stability of a tested and certified Linux distribution without contributing to the considerable costs of that testing and certifying. If you're running a mission critical workload without some sort of support, you're an idiot. If you're smart enough not to need support, then you're a Debian guy and you don't need Red Hat.
"If you're running a mission critical workload without some sort of support, you're an idiot. If you're smart enough not to need support, then you're a Debian guy and you don't need Red Hat."
These are the extremes. There's also a variety of other situations such as you have a mission critical workload for which RHEL is worth the money but some ancillary uses such as testing and training which production environment compatibility is needed, a clones is fine but RHEL prices can't be justified or it's used in production but the profit margins won't support buying RHEL. The first of these alternatives is one where the RHEL customer is going to review the market and the second one where they will never sell anything anyway unless they drastically cut prices.
"the SimilarTech research company reports that over 10 times as many businesses are using CentOS over RHEL" - As others have pointed this out already, there really isn't as much money on the table as IBM/RedHat seems to think. Eve though there really MIGHT be 10x more CentOS installations on the web than RHEL quite a few of them are run by SMEs who just don't have the budget for RedHat support. Pushing them hard will simply mean that they'll adopt something else, say Debian or Ubuntu.
The thinking is the same as the music industry during their massive crackdown on piracy. "Every downloaded copy is a lost sale". Except, it isn't, as those people using the free clones would never have afforded to use the paid version, just like pirates would never have bought the songs instead.
All it does, in this case, is reduce your potential market through bad press. Why would any new company decide to use RHEL when they look at how they treat the sector? They could use one of the other OS's instead, such as Ubuntu or SUSE.
In my vertical (research computing in higher ed) the cost of individual licenses isn't terrible. Our high performance computing nodes are in the order of $10k, so a few hundred dollars per node is not huge. I'd have happily written a check to Redhat for a few thousand dollars every year (if their support hadn't sucked) but only if provisioning were trivial. It wasn't (and the support does usually suck).
We remained on CentOS more because the provisioning challenges to automation were a constant irritant. These weren't insurmountable, of course, but with CentOS it was trivial. Rocky is our next choice, and similarly it is frictionless. And with CIQ, the support is actually great!
So, yeah, many thanks to IBM for screwing their pooch. But really Redhat had been there own sales prevention force in the first place.
this is just the so manieth attempt of redhat to make live difficult who want to use their source and build a distro from it.
this already had one incarnation, LOOOOOONG before IBM stepped in.
https://www.theregister.com/2011/03/04/red_hat_twarts_oracle_and_novell_with_change_to_source_code_packaging/
anyway.
who wants a linux with a backported kernel based on code from 5 years ago ? You ? really ? paid your enterprise subscription fee for your 5 year old software ?
think what ever you like :-)
this is just the so manieth attempt of redhat to make live difficult who want to use their source and build a distro from it.
this already had one incarnation, LOOOOOONG before IBM stepped in.
https://www.theregister.com/2011/03/04/red_hat_twarts_oracle_and_novell_with_change_to_source_code_packaging/
that was 12 years ago.
nothing changed much.
anyway.
who wants a linux with a backported kernel based on code from 5 years ago ? you ? really ? are you sure ? Paid your enterprise license then ?
think what ever you like :-)
Cory Doctorow and Mike Masnick have been writing recently abou the ensh-ttification lifecycle of products. The details differ, but I suggest that the path of Red Hat is following closely the steps of Twitter, Facebook, and others. Here is Jason Kottke's summary of the lifecycle from the writings of Cory and Mike.
https://kottke.org/23/07/seven-rules-for-internet-ceos-to-avoid-enshittification
Excerpt -- This is enshittification: Surpluses are first directed to users; then, once they're locked in, surpluses go to suppliers; then once they're locked in, the surplus is handed to shareholders and the platform becomes a useless pile of shit. From mobile app stores to Steam, from Facebook to Twitter, this is the enshittification lifecycle.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
