Ask Slashdot: How Long Should a Vendor Support a Distro? 137
Long-term Slashdot reader couchslug believes that "Howls of anguish from betrayed CentOS 8 users highlight the value of its long support cycles..." Earlier this month it was announced that at the end of 2021, the community-supported rebuild of Red Hat Enterprise Linux, CentOS 8, "will no longer be maintained," though CentOS 7 "will stick around in a supported maintenance state until 2024."
This leads Slashdot reader couchslug to an interesting question. "Should competitors like Ubuntu and SUSE offer truly long-term-support versions to seize that (obviously large and thus important to widespread adoption) user base?" As distros become more refined, how important are changes vs. stability for users running tens, thousands and hundreds of thousands of servers, or who just want stability and security over change for its own sake...? Why do you think distro leadership are so eager for distro life cycles? Boredom, progress or what mix of both?
What sayeth the hive mind and what distros do you use to achieve your goals?
The original submission argues that "Distro-hopping is fun but people with work to do and a fixed task set have different needs." But what do Slashdot's readers thinks? Leave your own thoughts in the comments.
And how long do you think a vendor should support a distro?
This leads Slashdot reader couchslug to an interesting question. "Should competitors like Ubuntu and SUSE offer truly long-term-support versions to seize that (obviously large and thus important to widespread adoption) user base?" As distros become more refined, how important are changes vs. stability for users running tens, thousands and hundreds of thousands of servers, or who just want stability and security over change for its own sake...? Why do you think distro leadership are so eager for distro life cycles? Boredom, progress or what mix of both?
What sayeth the hive mind and what distros do you use to achieve your goals?
The original submission argues that "Distro-hopping is fun but people with work to do and a fixed task set have different needs." But what do Slashdot's readers thinks? Leave your own thoughts in the comments.
And how long do you think a vendor should support a distro?
However long they decide. (Score:3)
It's not up to you. It's up to them. They're changing their model. You can switch if you want. There's plenty of distros out there.
Re:However long they decide. (Score:5, Insightful)
It's not up to you. It's up to them. They're changing their model. You can switch if you want. There's plenty of distros out there.
True. Would be nice if they stick to what they tell you up front though. Doing otherwise just makes them untrustworthy.
Re:However long they decide. (Score:5, Insightful)
It's not up to you. It's up to them. They're changing their model. You can switch if you want. There's plenty of distros out there.
True. Would be nice if they stick to what they tell you up front though. Doing otherwise just makes them untrustworthy.
This 100%. RedHat made a promise to keep CentOS [zdnet.com] as it was when they took over the distribution. On the back of this many people have used CentOS and then moved to RedHat when they wanted support. Now there is Rocky Linux [rockylinux.org] to take the place of CentOS and little apparent reason to recommend RedHat over Amazon or Oracle for the compatible, RPM based distro you move to when you start to need support.
I've always recommended people who want something solid and stable to install CentOS and then move to RedHat if their install becomes critical to their business. Sure it's put some millions of pounds of business in the direction of RH. From now on I'm not going to give any specifics about which commercial distribution you could move to, just mention each of the three being available if needed.
Re:However long they decide. (Score:4, Insightful)
Re: (Score:2)
Only if you trusted them. And anyone who trusts non-binding statements, and not make adequate contingency plans are idiots.
That is also true. It's a wonder anyone uses unsupported software.
Reputation is still worth something to those who do though. Sometimes anyway.
"However Long?" ain't it (Score:2)
True. Would be nice if they stick to what they tell you up front though. Doing otherwise just makes them untrustworthy.
The problem isn't "how long should they support X". That's secondary. The problem is "why in the hell should I trust an open source product when it becomes wholly owned by a corporation?"
The logical thing to do would be a model where your open source OS is run by a non-profit foundation that offers paid support for corporate users. You pay the bills, expand the operation, but the OS won't be pulled from the public because of the next quarterly profit projection.
Debian would normally be an ideal candidate fo
Re: (Score:2)
IceWeasel wasn't political/idealogical IIRC - it was due to trademarks on names and copyright on images etc, not the code. Also seem to remember there were similar naming issues with that "code family" and the mail client (thunderbird was gonna be firebird but there was a small DB system named firesomethingorother that it would've trampled on).
Re: (Score:2)
More to the point, virtualization is probably the correct response. I still have a version of Red Hat 4.5 (I think) that I run in virtualization so that I can run a few things like Civilization: Call To Power and Alpha Centauri. I haven't run across any games for a more modern Linux distro that are as appealing.
Re:However long they decide. (Score:4, Interesting)
Linux is remarkably backwards compatible. Just the other day I followed some instructions I found for getting WordPerfect 8 up and running again on Fedora 32 for fun. It involved installing a bunch of old libraries and configuring the ld.so to handle loading these old libraries, based on libc.so.5.
I bet you could get those games running natively on a modern distro. For example, here's a little shim someone used to run Alpha Centauri on modern distros: https://github.com/ZeroPointEn... [github.com]
Re: (Score:2)
Probably, but why bother. Virtualization works perfectly well, and handles adjusting to modern screen sizes and pixel depths.
Re: However long they decide. (Score:2)
Uum, it's just a matter of dependencies.
1. Add the package index of the old OS as an overlay to the index of the new OS.
2. Add the packages from your old OS as a source to your new one.
3. Install your old software like normal.
If your package manager is worth a damn, it can install that old dependency side-by-side with the new one.
Lacking abilities to do the above intuitively is what being a point-and-click user gets you.
Re: (Score:2)
How long will that "package index" remain valid? The virtual disk image has remained valid for over a decade. You look at your dependencies, and I'll look at mine.
OTOH, I have disabled the ability of the virtual image to contact the internet. But this would be desirable in any case. The old software hasn't been maintained in decades.
Re: (Score:2)
Linux is remarkably backwards compatible. Just the other day I followed some instructions I found for getting WordPerfect 8 up and running again on Fedora 32 for fun. It involved installing a bunch of old libraries and configuring the ld.so to handle loading these old libraries, based on libc.so.5.
You don't happen to have a link, do you? I'm curious because I did a similar thing years ago for an old MATLAB copy as it happens, but when I got curious about it recently I coudn't find the same instructions.
Anot
Re: (Score:3)
Here it is. It may be applicable: http://www.xwp8users.com/ [xwp8users.com]
Up to me to decide the support of the distro I cho (Score:2)
What you said is true in a way. It's also true if you look at it in reverse.
It's up to *me* to choose how long I want my servers to remain stable. It's up to each distro to decide if they are going to provide a distro that works for such servers, or if the industry moves away to a different distro.
I think most companies don't run servers on Fedora, because they want stability more than they want the new shiny thing of the week. They need long term stability, and they'll get ling term stability. Red Hat ha
Go over as well as Slashdot audience (Score:2)
Ugh, I accidentally hit Submit before I was done writing.
It seems to be going over as well as "Slashdot Tries Something New; Audience Responds!"
https://meta.slashdot.org/stor... [slashdot.org]
Less than two years later, Dice was no longer in the Slashdot business.
Re: (Score:3)
More relevantly, I do not think the model for RHEL is changing. If you want long-term support, you can still pay for it. That kind of work does cost labor and equipment time, therefore money. (Labors of love do exist, but usually not for long-term maintenance of someone else's software.) If people need an OS baseline that is stable for years, they should be willing to pay for it. RHEL uses that model, and it is not changing.
Re: (Score:3)
Seriously, why else would RedHat have bought CentOS other than to stop the hurt to its bottom line. And who can blame them?
My question is why would the CentOS folks, fully aware that that must've been the case, sell? My guess is that they made a bunch of money off of it. They could've remained independent and continued to maintain their distro - which, face it, was just a recompile of RedHat sources, so about as easy a distro to maintain as is possible. But what was in it for them?
There are Linux contri
Re:However long they decide. (Score:5, Informative)
Seriously, why else would RedHat have bought CentOS other than to stop the hurt to its bottom line. And who can blame them?
The explanation they gave at the time was to benefit from the community. They were leaking people to Ubuntu and so unifying with CentOS allowed them to protect that.
My question is why would the CentOS folks, fully aware that that must've been the case, sell? My guess is that they made a bunch of money off of it. They could've remained independent and continued to maintain their distro - which, face it, was just a recompile of RedHat sources, so about as easy a distro to maintain as is possible. But what was in it for them?
The CentOS guys were promised that RedHat would continue supporting their users and give them better service. Some of them went to work for RedHat. It was supposed to give stability and long term guarantees of the resources needed to maintain a distro.
CentOS basically was in it to rip off RedHat. Probably because they wanted a free RH distro for themselves, but whatever. At some point they were running an extortion scheme on RH.
Bullshit. CentOS was what made RedHat acceptable. Just as CentOS takes the free software content from RedHat, RedHat takes from everyone upstream of them. We (I'm included here) gave our code on the understanding that it would be possible to copy it, alter it, experiment with it and sell it but in the most important cases with the condition of giving the same freedom to the people you give the code to. RedHat uses trademarks and proprietary branding which makes that impossible for their whole distribution. However that freedom was regained with CentOS and Rocky Linux [github.com] should now restore it. That's not extortion, that's just following the spirit of the copyleft licenses that they get their code under.
So the question is, do we want RedHat to have a profitable business model? If not, why not?
Up until this point for almost all of the time I have wanted RedHat to be profitable. I was an early user. I have been a very happy customer. There was one point, where they switched from Red Hat Linux to RHEL, where, briefly that changed when they abandoned the desktop I was quite invested in. Right now I am not sure and the reason is that I no longer trust them. Amazon also contributes to free software to a small extent, but they drain it much more. If RedHat is going to be like that then I would rather they were dead.
We need to find license models which force cloud providers to provide full source access and allow people to experiment with and change the cloud services they rely on. This is the key part of the future. RedHat could be part of that. As it is they seem to be aiming to be just another Google or Amazon exploiting the community.
As long as possible OR pass the baton (Score:2)
Re: (Score:2)
I think that the answer for an Enterprise grade software product is "At least a long as originally promised". For operating systems in particular, that time is usually 10 years or more. Hell, Windows XP was supported for about 12 1/2 years.
If you say that something is supposed to be supported until 2029 and then cut it off 8 years early, you're going to piss off its user base.
Re: (Score:2)
But it SHOULD be up to "your" in the sense of "as long as 'you' are willing to pay for the distro to be supported, then it should be supported". It really should come back to "show me the money" because support is basically kind of tedious and boring for most people, but it's still necessary and the "normal solution" is to pay people to do such work.
Continuously searching for a fresh distro is not a constructive solution approach. More of a portable and ongoing disaster approach as each new distro provides
Re: (Score:2)
Re: However long they decide. (Score:2)
Sorry, no can do. He is changing his model.
CentOS 8 a good idea? (Score:3)
Seems like CentOS got too close to Red Hat... and that's why they're pulling out. Way to get hated by the community, and there's plenty of other distros out there. If you want close to Red Hat, get Fedora.
Re: (Score:3)
If you want close to Red Hat, get Fedora.
Fedora and CentOS have fundamentally different purposes. CentOS is intended to be fully compatible with the corresponding RHEL release. Fedora, on the other hand, is the upstream distribution that RHEL is based on, so if you use Fedora you are using software that hasn't gone into RHEL yet.
While Fedora is useful, it's not fully compatible with existing RHEL releases, and so if people chose CentOS over Fedora in the first place, then they likely don't want to move to Fedora. For the same reasons, they also
That's what it used to be (Score:5, Informative)
To be clear, CentOS *was* perfectly compatible with RHEL, it was Red Hat debranded. It is now becoming pretty much what Fedora was. Which sucks, because we need what CentOS has been
Re: (Score:2)
To be clear, CentOS *was* perfectly compatible with RHEL, it was Red Hat debranded. It is now becoming pretty much what Fedora was. Which sucks, because we need what CentOS has been
Absolutely. What needs to happen now, is some group comes along and just builds RedHat from the required to be available source with all the branding removed. Like CentOS did, before RedHat took them over and killed them (how was that not predictable anyway?).
The best thing that could happen is a fresh new start. Without RedHat's permission or involvement. Like it used to be.
Rocky, Springdale, minor releases (Score:5, Informative)
I understand one of the co-creators of CentOS plans to resurrect the concept as Rocky Linux. There is also Springdale Linux.
According to ArsTechnica, CentOS might not be changing as much as I thought. They say the new system is:
RHEL version x.0 forks from Fedora
CentOS Stream version x forks from RHEL version x.0
Development work for RHEL x.1 is done in CentOS Stream version x repos
RHEL x.1 forks from CentOS Stream version x
So it's only point releases that bring CentOS ahead of RHEL.
Re: (Score:2)
I'll be watching Rocky, I hope they succeed. Will be interesting to see RedHat's reaction, if any.
Re: (Score:2)
Re: (Score:3)
You'll be watching, but I'm not seeing anything about if you'll be paying. Because unless you pay for it, this will all happen again.
Probably. I've used many Linux distros over the decades. RedHat has lasted longer than most.
Honestly I prefer *BSD myself, but in the end it is all some form of open source software.
It's like that Liberty Mutual insurance commercial. Only pay for what you need.
Re: (Score:2)
That still means CentOS is ahead of RHEL and you're now an RC tester instead of a beta tester.
Re: (Score:2)
Re: (Score:2)
Any distribution may participate in the bugfixing while the bug is embargoed. There are several subscription-restricted security mailing lists where work is coordinated. Why doesn't CentOS fix with zero delay? No idea, they have chosen to do that. E. g. openSUSE does fix with zero delay.
Re: (Score:2)
Re: (Score:2)
Rocky Linux is a community enterprise operating system designed to be 100% bug-for-bug compatible with America's top enterprise Linux distribution now that its downstream partner has shifted direction. It is under intensive development by the community. Rocky Linux is led by Gregory Kurtzer, founder of the CentOS project. There is no ETA for a release. Contributors are asked to reach out using the communication options offered on this site.
They have since posted a Q2 2021 ETA for their first release [zdnet.com].
I've been playing with Stream for several months, but gotten more serious about trying it lately. I think it could work for some use cases I have, but not all.
Re: (Score:2)
But I am watching Project Lenix as well and if that one seems to have more momentum say Q3 next year, when I need to start getting serious about CentOS migrations, then that will be the one I u
Re: (Score:2)
Seems like CentOS got too close to Red Hat...
I figured it was IBM's ownership of Red Hat and pressure to increase (short-term) profit.
Re: (Score:2)
Not likely. IBM bought RedHat for their cloud technologies. If they were interested in being a Linux distributor, they could have bought RedHat a lot cheaper 20 years ago.
It's not like RedHat wasn't already heading down that path before getting involved with IBM. Think about them tying systemd to Gnome and PulseAudio. At this point, Linux is all but a proprietary RedHat OS.
Re: CentOS 8 a good idea? (Score:2)
My server needs stability. Fedora is fine for a desktop OS if you do not mind occasionally fixing something that an update breaks, but it is not stable or secure enough for a sever. Most CentOS users run servers.
Not one right answer (Score:2)
There are several factors in determining this. What is the OS used for? Why can't one upgrade more frequently?
If it's an embedded product, or an appliance, LTS releases make a lot of sense and really should support at least 4-5 years for many types of products to be viable. This is certainly true for IoT uses.
For servers, it's another story because we're in the era of spinning up new ec2 instances or using docker/kubernetees. If you do it right, it shouldn't be such a big deal to migrate to a newer OS beyo
Re: (Score:2)
Desktop users can migrate more often for personal use, but in business environments you start getting into things like VPN and other security software. It's hard to control that and new versions are often not supported right away.
In the "many tens of thousands of users" business environment, it seems the "sweet spot" for most IT departments is to wait on upgrading the OS until it's a least a year old, and then keep on that OS for ~4 years. Likely upgrading the OS with hardware refreshes. Which means a distro would have to support it for 5 years.
Now, they are constantly applying patches to those OSes, so it is quite debatable whether or not they're actually running the "old" OS anymore. And because of that they probably should upg
As long as the support policy is clearly stated (Score:5, Insightful)
The makers of a distro can support it for as long as they want, but if they say they're gonna support in for x years, they really ought to support it for x years. There are distros with different philosophies of stability vs. shiny, etc., you should be able to find one that is comfortable.
If they want to change policy, that's fine, but maintain the most recent release according to the policy in effect at the time. People get invested in a distro, and to have the rug pulled out from under them is not good. I'm thinking primarily of commercial/enterprise users.
Re: (Score:2)
Well, I switched away from Red Hat when they discontinued the "Professional Edition". It's not quite the same, but I suspect the reaction should be the same. (Whether it will be is another question.)
Re: (Score:2)
Re: (Score:2)
I base it on how much I paid for it. If I got it for free download, I don't expect much support. Of course, it would be nice if they made it stable so it isn't a continuously moving target - after nothing, nothing's worse than installing Windows and spending the next 2 days rebooting to get eve
Timer starts when a new alternative is viable (Score:2)
Given that integrators need a new target to migrate to, the answer really should be based on when a upgraded version is available and generally viable. Even with this consideration there's always someone who will be burned by a termination too soon, and much potential left on the table to maintaining support too long.
Of course, "end of support" as a security end-all is something people need to unlearn. There's plenty of legacy devices that could have been properly segmented and operated without updates fo
I'm fine with centos/8 dying as a parrelell to RH (Score:2)
...I've ran into way more problems they I would like to admit the last 3 years dealing with a stuborn linux admin that has insists on using centos in lue of RH like our production environments because he doesn't want to bother to register for the "free" developers edition.
He's messed up just about every package install that has been requested of him for our RH environments..its gotten to the point I don't even bother asking, I put the request in for myself to do it, and I am the software admin.
Its amazing..
Yum (Score:2)
I hadn't used RH in years. Our cloud team was setting up an Apache proxy for us, and their RHEL distros are *dirt* old. The Apache build was way out of maintenance. It took me, maybe, ten minutes to figure out how to set up the backports repository and install the latest Apache build on the system. This stuff isn't hard.
Re: (Score:2)
Unix/Linux get a life....
18 years (Score:2)
Then it's out of the house and get a job.
The answer is sort of in the summary (Score:2)
For those who have work to do, Redhat have RHEL, the fully commercially supported version. You want free, you get what you pay for.
Wrong Question! Use the distro that has a (Score:2)
Personally I use stable and LTS distros.
That's the wrong question (Score:4, Insightful)
There's no right answer to the question you asked. The right question to ask is: "Once a distro commits to a support lifetime, should it go back on that commitment without much notice?"
Re: (Score:2)
Re: (Score:2)
No one is asking for a contract. They are asking for commitment. Contracts are for those people who are unwilling to demonstrate commitment.
Re: (Score:2)
Contracts are legally binding to set the terms that are agreed to so that both sides have clear understanding of their obligations and what is not an obligation. Nothing about willingness. The need for that has been understood for centuri
Re:That's the wrong question (Score:4, Insightful)
Nobody says Red Hat has an obligation to stick to the originally-stated lifespan for CentOS 8. Clearly, they're within their rights to do whatever they want.
However, what they did is arguably a crappy thing to do, even if it's perfectly legal.
Re: (Score:2)
Let's say you decided to bake cookies and have a stall on a street where you give them away for free. It becomes so popular that it just becomes an
Re: (Score:2)
Disclaimer: I have no skin in this game, since I use neither RHEL nor CentOS. However, your analogy is not quite right. The correct analogy would be opening a stall in the street giving away cookies, and saying "I intend to continue giving away cookies until 2027, so rest assured your cookie supply will be stable." and then changing your mind with very little notice. Again, perfectly legal... but crappy.
Red Hat's press release from 2014 [redhat.com] even uses the word committed: "CentOS owes its success not just t
Re: (Score:2)
"I intend to continue giving away cookies until 2027, so rest assured your cookie supply will be stable." and then changing your mind with very little notice. Again, perfectly legal... but crappy.
No, it's not crappy. Because that is only an intention. This is the height of entitlement culture. If you depend on people giving away things for free, you have no right to complain when it goes away, even with no notice, because there is no moral obligation to keep providing. It's not crappy - things change. You and the other guy are worse than those entitled bridezilla stories that pop up on social media.
CentOS community while remaining committed to our current and new users.
This means nothing. Like I said to the other guy, it doesn't matter one bit if they use the word "com
Re: (Score:2)
The phrase "no right to complain" is not about l
Re: (Score:2)
As I pointed out above, the words "committed" and "commitment" appear in Red Hat's own press release [redhat.com].
Red Hat has since removed most pages referring to the original CentOS lifecycle, and the Wayback Machine is unfortunately giving me 504 Gateway Timeout errors, but I'm sure I can find plenty of evidence that Red Hat did state a support timeline for CentOS 8 that was much longer than what it is now.
Re: (Score:2)
If "commitment" means nothing, then I stand by my assertion that it was crappy of Red Hat.
Why are you so angry? I don't use Red Hat. I don't use CentOS. I never completely trusted RH and my mistrust appears to have been well-placed.
Re: (Score:2)
So committed they couldn't even bother to announce it in the press release...
The "commitment" is worth as much as the paper it's written on - it doesn't exist, no matter how much you want to believe it. Anyone can say they commit to something. I don't go around planning my future finances based on a commitment by some Nigerian prince saying he would send me lots of money.
They didn't pull the rug out from CentOS users, because CentOS
Re: (Score:2)
There is no commitment because there is no contract.
You don't seem to understand the difference between those two words. Go look them up in the dictionary and realise that just because someone doesn't write something down and sign it in exchange for some monetary consideration doesn't mean they can't be trusted to hold their word.
You sound like you have serious trust issues. When you invite people over for dinner at 6pm do you send them a written contract to sign or do you rely on the fact that they are committed to showing up and keeping their word?
Re: (Score:2)
The Evil Atheist writes: "...take your ad hominem "trust issues"..."
The Evil Atheist writes: "Dickhead."
Oh my. The Evil Atheist pot calls the kettle... :)
For zero days unless you pay them (Score:2)
How Long Should a Vendor Support a Distro?
They need not support it at all. They gave you the source code. Support is on you. ***Unless you pay them to do this work for you***. That is the open source way. Any support they do offer is purely a courtesy.
Re: (Score:2)
(oh, and the answer is Debian Stable of course. The answer is always Debian Stable...)
As long as it makes financial sense? (Score:2)
Supporting any distro takes resources and costs money, there's clearly a point the margin becomes small enough that it just no longer makes any sense to support it commercially. I find it hard to believe that the main distro vendors haven't already factored this in when determining the length of their respective LTS windows and are just leaving money on the table.
Re: (Score:2)
From my longer comment on this story I realized that "financial model" is a key term, but this is the only other mention if "financial" in all of the visible comments. No moderation and no replies. Such is Slashdot 2020.
Maybe because you didn't actually offer anything like a solution approach? I doubt it would avail you. I keep pushing for the idea of a CSB (Charity Share Brokerage) to fund OSS (and perhaps journalism), but never detect much comprehension or interest. I would actually welcome constructive c
You get what you paid for (Score:2)
Not more, not less.
Compatibility offers MORE value (Score:2)
Personally I prefer an OS that will run my shit with minimal hassle. Computers are after all are supposed to be tools to get shit done. The value proposition of computers is reduced when you constantly have to waste time dealing with churn that doesn't help you in any way.
Something you can forget about for years or decades at a time would be ideal.
Re: (Score:2)
You might want to try OpenBSD. I have had 2 year uptimes.
They don't (Score:4, Informative)
The dirty secret of CentOS/RHEL is they don't really provide what it says on the tin.
I was there when they broke all the chrooted named's and wouldn't fix it.
I was there when they wouldn't backport some apache 2.4 security fixes to apache 2.2 because the task was too enormous. ("Software Collections" came out soon after)
I was there when the spectre/meltdown patches made Xen DomU's unbootable and "switch to KVM then."
My "main machine" was in-place upgraded, stepwise, from Redhat 9 to Fedora 28, all with yum, then dnf, plenty of hell, and occasionally with a boot DVD (like the i686 to x86_64 arch upgrade).
Now I'm on Debian everywhere, mostly Buster, and the upgrade from Stretch was a breeze. They're not perfect but they like people to upgrade and try their best. Spinning up a new VM isn't that big a deal anymore because the upgrades are straightforward.
95% of everything is better on Debian (and I can say for sure that wasn't true 15 years ago).
One of the main advantages of getting 5+ years with IBM (nee Redhat) was that upgrading was neigh impossible. Side-step the LTS problem with Debian (or even LTS Ubuntu if you favor it).
Re: (Score:3)
Sounds about right... back when Ubuntu started acting erattic I decided it was time to go back to Debian on my own machines. The more recent behavior of RedHat has just confirmed my impression that just like you should avoid relying on proprietary software, you should avoid relying on commercial linux distributions. The incentive structure of the commercial world invariably puts them at odds with the people using
Just wait for self driveling cars to stop at 3-5 y (Score:5, Insightful)
Just wait for self driveling cars to stop at 3-5 years where the stop getting updates or maps updates.
Unless you pay like $299 year for map updates and maybe an $500 OS upgrade.
Also maybe even very over priced storage upgrades like $299 + dealer install costs for an 1TB flash drive.
Re: (Score:2)
Unless you pay like $299 year
Nobrainer: Absolutely worth it. Hell I'd pay double that to have the car drive itself.
Re: (Score:2)
In the US, there probably won't be legislation because of lobbying by the car industry, and then the support of all the free market nerds, because the freedom to kill people to make money is an god/market given right.
Re: (Score:2)
Just wait for self driveling cars to stop at 3-5 years where the stop getting updates or maps updates. Unless you pay like $299 year for map updates and maybe an $500 OS upgrade. Also maybe even very over priced storage upgrades like $299 + dealer install costs for an 1TB flash drive.
None of that may actually matter. If its self driving do I want to own it or just ride share it? If the latter would I care? Those costs should be spread over a large number of rides and perhaps not really be noticeable.
Centos needed an upgrade from 7 to 8 in place (Score:2)
Centos needed an upgrade from 7 to 8 in place but they did not port that.
ubuntu does and it's free.
The impact.... (Score:2, Troll)
I'm the Director of my shop. We use a lot of Linux in some pretty mission critical places.
We're done with CentOS and Redhat/IBM over this. We will not transition to stream. We will no longer pay for the support.
We had just finished our transition to RHE8/CentOS8 in order to set ourselves up for running on it until 2027. The bomb was dropped, a bunch of things broken in most recent release, and in a no warning situation.
Instead I'll be transitioning all the machine in Q1 to something else. There's a number o
Re:The impact.... (Score:4)
You seem to be confused, CentOS was $0 and so you have no right to complain. You want support, pay for it. Red Hat supports their distros for 10 years.
You don't have the resources to properly support a distro.
Re: (Score:3)
What Redhat should have done was keep their promise for 10 years of support, and move Centos 9 to the new model. There would be complaints, but they would have still remained a company that could be trusted.
Re: (Score:2)
Nonsense, "promises" about zero dollar gifts in the future aren't the same thing as a business deal. Most of Red Hat and IBM's customers couldn't care less about Centos. You want guaranteed 10 year support, cough up the loot, quit being whining welfare queen.
Re: (Score:2)
Should have added it's only "trust" from people paying $0... why would Red Hat give a shit?
Re: (Score:2)
"Commitment" about future gifts and free rides? that's laughable, be thankful for your years of free ride, now it's time to pony up. 10 years support takes immense resources, bad business to give it away.
Your alternative is to get on free Linux or BSD distro with several years support but be agile enough to migrate every five years or so. That's what most the whiners should have been doing anyway.
In house distro - a security plus (Score:2)
We're a breath away from rolling our own in house distro tailored to our industry.
A security plus. A minimal distribution can help reduce opportunities for attack.
It's not so much that CentOS isn't being... (Score:2)
... supported until the date that was originally set. It's the shift of CentOS being a unbranded clone of RHEL to a test bed for RHEL. (It'd be interesting to see the internal memos behind this decision.)
Competitors DO offer long term versions. (Score:2)
Should competitors like Ubuntu and SUSE offer truly long-term-support versions
Well, Ubuntu LTS releases are supported for five years. Maybe do research before asking? All I had to do was gewgl search "how long is ubuntu supported" Here for the lazy:
Ubuntu Release cycle page [ubuntu.com] (LTS five years)
SUSE lifecycle page [suse.com] (Enterprise up to 13, thirteen years)
Well, updates *are* support. (Score:2)
The reason people are staying with old versions is twofold:
First, new stuff might break their system. But of course if the system wasn't broken, you would not need any support.
Second, it takes work to transition, adapt to new storage format and settings. But there the problem is laziness. Laziness of the user to do that work, and laziness of the developer to make sure that however you are using the thing, it is converted automatically and losslessly and that that is always possible. That means no feature re
slackware (Score:3)
Slackware has no set schedule, and having a small team it can be hard to support a release for a long time.
But looking at Wikipedia, one release (14) reached 8 years of support, and there are 3 active releases in addition to current are being supported now.
So, if a small team can manage that, larger and richer distros should be at least able to match that. So I would say no excuses.
simply: forever (Score:2)
Anything else I find to be lack of professionalism and lack of work ethics.
I don't mean software should be supported on new hardware, or new features added but bugs should be fixed forever. "Oh, but who has the time for that?" you say... well, you had the time put the bugs there to start with, right? Then, find the time to correct them. Once nobody uses your software or no more bugs can be found, supporting your software would mean nothing anyway, so your burden will only be as heavy as you want to: shitt
Re: (Score:2)
Anything else I find to be lack of professionalism and lack of work ethics.
I don't mean software should be supported on new hardware, or new features added but bugs should be fixed forever. "Oh, but who has the time for that?" you say... well, you had the time put the bugs there to start with, right? Then, find the time to correct them. Once nobody uses your software or no more bugs can be found, supporting your software would mean nothing anyway, so your burden will only be as heavy as you want to: shitty software? heavy burden. Great quality software? you can make the promise and it won't take you that much.
*ahem*
"This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE."
In other words, I, as the author of my software, owe you precisely nothing, and have made you no promises. If you want me to promise to fix bugs that affect you or provide any other sort of support, then you're going to have to promise me something of value in exchange. You know, a couple of chickens, weekly blowjobs, or w
Forever and a day (Score:2)
The obvious answer to this dilemma is of course a 'rolling release' distro. They are supported forever, as long as the user updates regularly.
I've been using rolling release distros (Gentoo, Arch) for the last 13-14 years, and I'm never going back!
Forever with an asterisk (Score:2)
Yes, in principle, a distro should be supported forever. That can have an asterisk if upgrades are 1) truly upgrades (offer new function, features, and real bug fixes); and 2) is still usable on the same hardware as a previous distro. At some point, one presumes it will become inconvenient to keep supporting anything on old hardware, but that should not mean dropping it entirely; the last distro to work on the old hardware should continue to get security updates and other patches that apply to it forever, o
12 years for essential apps (Score:2)
Just leave a legacy app downloadable for phones that have downloaded it before. Delete after 12 years. Few economies outside the USA support new phones every year. It recycles 1st world iPhones into 3rd worlds. Saves planet resources and enhances Apple sustainability not producing a throwaway platform.
Americans (USA) forget the rest of the America’s. iPhone 4S works faster in Central/South America than it evr did on T-mobile. It even would lose mike and drop out of conversations. Down here none of tho
WIth so many distros this has to happen (Score:2)
RHEL/Centos mix (Score:2)
My organization has RHEL licenses for at least 500 machines. It also has a large HPC farm which runs Centos because the RHEL licensing costs for that number of cores would be absurd. HPC transitioned to Centos 8 earlier this year. Now it will probably go back to 7 while the IT division decides what to do about everything.