I've run 200+ server environments on centos4, a 40+ server cluster on centos5, and now run a 50+ server cluster on rhel6. I'm pretty much dead on their target audience here.
This seems almost mean. Its a billion dollar company calling out a volunteer group of a couple of guys for being slower than them.
Part of the reason I switched to rhel was exactly this graph, I didn't have a lot of faith in centos as a going concern given the delays on 6. But also, its clearly not easy for a reason. There is a lot of work to be done, and the farther you trace your way up the rhel/epel/fedora tree the more you realize this is a community providing an insane amount of value and deserves to be funded.
So if Oracle is going to sell me on using and/or paying for their distro, being faster than a handful of volunteers isn't gonna do it. Going toe-to-toe with redhat on funding, contributing, producing and supporting open source software such that I want to fund you is what'll do it.
That's not quite the intent. Look, here's the deal: I know we all like to think that corporations are evil and the presence of corporate interests actually messes stuff up.
But this really is one case where you want the oomph of bilion-dollar company making sure that you get your security updates in time.
No one else is providing that. Oracle happens to be doing so, and we're making it available to you for free, so why shouldn't you benefit from that? (And seriously, I get that you're not paying for CentOS, and I'm not interested in getting you to pay for Oracle Linux either.)
The guy you're replying to has implied he works for Oracle and helped bring this to fruition. Instead of asking what the advantage they see is or diving into it, you've accused him of lying and being a part of an evil organization.
Oracle's a business--they make good products that many businesses rely on. While they've done some stuff that this crowd greatly disproves of (see: Java copyright/patents lawsuits), their motivation for doing this could be as simple as keeping the brand's security image in place.
I just want you to know that you had a good opportunity here, and you wasted it.
The "guy you're replying to" is Waseem Daher, who was COO and co-founder of KSplice, a startup acquired by Oracle. KSplice provides a mechanism for streaming updates to Linux Kernel and used to support CentOS, Ubuntu and Fedora. After Oracle acquired KSplice, they discontinued support for CentOS while continuing to support the other distros. Oracle also no longer makes available the formerly GPLed code of KSplice.
We didn't discontinue support for CentOS; our legacy customers can continue to get access to Ksplice for all the distros we used to support. We did stop accepting new customers for anything other than Oracle Linux, though.
The bottom line is that CentOS is now a second-class citizen with regard to support from KSplice, compared to Ubuntu and Fedora. See the KSplice download page at http://www.ksplice.com/uptrack/download in which CentOS is missing.
Regarding the source code, yes the original code was GPLed is still available. But are you saying that Oracle/KSplice will continue to make available ongoing versions of KSplice?
The link you provided is to a source tarball that was mostly last updated in August 2009, except for a small change in July 2011 (removal of zlib detection). Therefore, either the KSplice code is not being maintained at all, or there are ongoing updates to it that will no longer be released in source form.
Sounds like the classic bait and switch. Down the road there surely will be more interesting features (virtualization comes to mind) that can unfortunately not be supported in the free version of Oracle linux. But hey, special deal, just $999/yr and you won't have to migrate to a different distro for that one feature!
Plus included with the price you get our awesome support-plan which you never call, but when you do it will only take weeks before they return a canned response designed to make you buy more products instead of actually solving your problem!
Personally I don't regret sticking with deb-based distros. They have problems of their own, but at least none of this extortion-bullshit.
As long as we're noticing how you guys sold out, I'd like to bring up the Ksplice blog.
It used to have some really interesting stuff on there, but after the move to Oracle, the interesting posts disappeared.
I'm talking specifically about 'Coffee shop internet access'. Man that was a cool post and it taught my brother and I a LOT. Then it disappeared, along with most mirrors of it. I was able to find one left, here:
http://www.getoto.net/noise/2011/01/18/coffee-shop-internet-...
after around a full 60 minutes of searching-- and it doesn't have the images from the original.
I emailed you guys, and J was kind enough to send me the text of the post, but again no images. She said she hoped the coffee shop post would make it to the oracle blog, but this was about 6 months ago so I don't think that'll happen. Anyway, I hope you're able to do good at Oracle, and I hope that you're still trying to, and I hope that you don't get taken advantage of.
Yeah, getting these posts back up has largely been a casualty of our lack of resources, not any sort of coordinated malicious plot against them. But we're glad you like 'em! (and I think your comment was the little bit of kick we needed to just go and bring it back.)
Maybe so, but if I was Oracle I would take this as great feedback. My guess is that the majority of people running CentOS have a very similar attitude, so they should understand that this will be the reaction, whether it is fair or not.
It is up to Oracle to figure out how to build the required trust to market this, not up to consumers to accept it given Oracle's history. Millions of people are missing billions of opportunities every day (and you're probably one of them!) but you don't blame them for it, you figure out how to market a product more effectively instead.
It's quite possible, that, being a huge company, they do a lot of bad crap (they are not high on my list), but there are some good people there trying to do good things.
I've met a lot of smart and friendly developers who work for Oracle (or companies acquired by Oracle) so I know for a fact there are many people in that organization which are genuinely good and really smart.
However, it seems to be a culture which promotes aggressive "shark" type personalities into leadership positions - people who, in my experience, have less focus on ethical business practices and more focus on increasing profits, crushing competition, and advancing up the ladder.
It obviously makes for a successful company but it's a not one which I would trust to have my business' best interests at heart.
Given that Red Hat still can't get some people to move off RHEL 3, I'm amused that you think this could actually kill CentOS :)
And in the unlikely event that it were to happen, you could just switch to Scientific Linux, or whatever new RHEL clone would crop up in the wake of the news.
or we could stick with the guys who have given us centos.
i am not sure why your compassion is so selective. people who work for fucking oracle are misunderstood saints, while the poor sods that put centos together - who have never done anyone any harm - should be thrown to the wolves.
I don't think I've ever espoused that opinion! I like CentOS a lot. I had a beer with Karanbir in London in 2009, and I'm totally impressed that they're able to put out an incredibly popular Linux distribution with next to nothing, resource-wise.
But if I were a sysadmin, would I run CentOS in production? Is CentOS perfect? I'd say no on both counts, and I think there's nothing wrong with saying "There's room for improvement, and we think we've done better."
I'm having problems with seeing why CentOS may cease as a result of this move from Oracle and Mr Daher.
CentOS is free, so as long as the sponsors continue to sponsor the work on CentOS and as long as whoever pays for the servers &c continues to do so, the project can continue.
Scientific Linux and PUIAS exist for other reasons than commercial server provisioning and those reasons will probably stay around.
The CentOS and Scientific Linux forums contain useful information, the mailing lists contain more technical information. Oracle won't to my knowledge be providing self-help forums or other community features as they wish to package support as a product. I can imagine Oracle Linux users who do not require support dropping into, posting to and contributing to the CentOS and SciLi forums.
How will CentOS (or another clone) be impacted to the point of extinction by Oracle Linux? Am I being dense here?
Most users of either one will likely not contribute anything at all back besides simply being part of the "network effects". I'm not a fan of Oracle either, but your rhetoric is in the red zone...
you're right i'm angry. i'm sorry if it offends you. but still, i don't see what you're argument is. are you saying that because i am angry and "rhetorical" it's ok to switch to oracle?
is noticing that i am angry somehow going to make everything
work out? when centos folds and oracle decide to "raise margins" me being angry will make centos come back?
because if not, i would suggest worrying a little less about my emotional state and a little more about the future of your operating system.
I don't use CentOS actually, I'm fairly happy with Ubuntu, and if I didn't use that, I'd go back to Debian.
For many years I was a Debian package maintainer, and Debian got by ok without much corporate support, despite the existence of Redhat.
In other words, your thinking is way too zero-sum. I think there's room for both. Perhaps the competition will make Redhat think about their model some too, improving things for everyone.
An example of software-industry practices by Oracle: After Oracle bought ATG (e-commerce software vendor), they assigned an engineer to remove the code that detected and limited the number of concurrent users. ATG software is licensed by the number of concurrent users, and had for a long-time a mechanism to limit usage to the actual number of licenses for that specific customer. Oracle removed the limit code so that some customers would unwittingly exceed their licensed number. The intent was that these customers could be targeted by Oracle's license police, who'd extract high financial penalties.
A hard-nosed business decision. Potentially lucrative, profitable. Not illegal. But demonstrating utter contempt and manipulation of their customers. If this is how Oracle treats its own customers, how do you think they treat competitors like CentOS and the OSS community?
I am quite familiar with ATG, having worked with the product since 2001. I don't believe that Oracle removed the license limitation on purpose, but rather to make ATG consistent with their other products which you can download, install and develop with, but as soon as you use them for production you have to pay.
You also have to factor in that ATG's old licensing department was chronically incapable of delivering the correct license files: they used to send out production licences for the wrong IP addresses, session limitations or bad expiration dates, and it wasn't uncommon to have 3 or 4 runarounds with them in order to have what you paid for.
That said, I'm not looking forward to use Oracle Linux in any of our production servers, unless my clients specifically asks for it and pays for the support. And this support has to be of very high quality, it will take me just one canned reply to put CentOS back in.
Businesses can be evil. Crushing your competition via patents is evil. Damaging others way out of proportion to your own gain is evil. Buying elections with a flood of money is evil. The Libor scandal is evil because it screwed millions of contracts. Stomping a kitten is pretty tame by comparison.
Businesses are sociopathic - especially since corporations claim the rights of a "person" under law.
Current North American society appears to deem sociopathic behavoir as being evil - so it's not much of a stretch to claim businesses are evil.
Not saying they can't act good or do good in society though as many obviously do. It's also worth noting that good/evil are not objective binary things - so your definition of evil is likely different then the GP's.
True, and one that exists to maximise shareholder value. Alas, that aim may not contribute positively to the larger picture, hence various forms of regulation that exist in many countries.
This is an issue too large for this particular page.
And the GP was giving his honest opinion. An opinion that is held by many people I've interacted with in the open source community regarding Oracle. Maybe instead of browbeating him over it you could consider encouraging Oracle to pay attention to things like this and do something about recovering their image. BTW, it isn't illegal in any jurisdiction I am aware of to harbor personal feelings towards brands and companies. That's what marketing is all about anyway.
For a large segment of us (presumably the target market) there's no point in Oracle trying to recover their image. Based on direct personal experience I will fight tooth and nail against giving Oracle one red cent or one idle second. I can't think of anything they could do in the short or medium term that would change my mind so why waste time trying? Buying purchasing managers steak dinners and lapdances has a much better return on investment.
I was giving my honest opinion of his honest opinion.
An opinion that is held by many people
These people have a weak grasp of the concept of 'evil'. In my opinion.
Maybe instead of browbeating
I wasn't doing that. My opinion.
You could consider encouraging Oracle
I'm nobody Oracle is going to listen to. I say this as a guy who has been dealing with Oracle for about a decade, supported database and application servers.
Oracle pays attention to guys that write checks. And not just any checks but big checks.
If the CIO of (say) AT&T brings this up with his sales rep, things will happen. Me .. not so much.
BTW, it isn't illegal
I believe you are reading things that I did not write.
Anyone decrying Oracle as "evil" is falling into a trap that I have warned about[1]: they are making the mistake of anthropomorphizing Larry Ellison -- they are infusing Oracle with humanity that it does not in fact have. Oracle is, as another commenter put it, a corporate sociopath. The inability to empathize is fine for a monopolist (and indeed, it's a terrific asset -- and after five months of working there, I concluded that it's the only thing that I think Oracle has any real talent for), but it's terrible for a technology company. In particular, the inability for Oracle to see anyone other than themselves leads to decisions that don't just erode trust, they destroy it beyond all repair. What you are seeing here are the costs of Oracle's decisions: it's what Oracle did to Hudson, to MySQL, to Java and to OpenSolaris. If you honestly don't understand that -- if you truly believe in your heart that the dismissal of Oracle is somehow merely Occupy-esque anti-corporate sentiment -- then you yourself probably have an inability to empathize that should bode very well for your future career at Oracle.
You're awesome. I watched the full vid of that talk a month ago or so and it really hammered home what matters.
What we do as developers matters. Our time on this world is short enough already. What reason is there to spend years of our lives to further the aims of a company that doesn't share our goals and hopes, and in the end may be acting completely opposite to what we believe is fair and honest competition? What a waste...
Looking forward to seeing more and more of the OpenSolaris/SmartOS/Lumiere innovation spread to the rest of the community now that Oracle has caused everyone to jump ship. In some really cool technical ways SmartOS is really advanced, but the usability definitely needs work... things should improve as more people start hammering on it. Would be good to see more cooperation/coordination between SmartOS & Linux distros.
It depends on what you mean by "significantly diverge." If you mean "innovate", then yes -- absolutely. We already have tons of evidence of differentiated innovation in illumos[1][2][3] -- none of which can be brought back into Solaris.[4]
You are saying that Oracle's goal is to make sure that CentOS customers can access the resources of a billion dollar company in order to get security patches on time.
If that were all that was going on, then Oracle would just provide funding and engineering support to CentOS.
Since you aren't doing that, one of two things must be happening:
1. This is part of a long term strategy, which means it will somehow make a play for my wallet on some future date.
2. This is not part of a long term strategy, which means it is a low-level initiative. Most likely it will chug along until the responsible middle manager retires, transfers or gets bored with it. If by some happenstance it becomes successful enough to attract the attention of top management, then goto #1.
Bottom line - Oracle is a business and wants my money. I know it, and you know it. Using free products from Oracle is like accepting a favor from Tony Soprano. Sooner or later, the bill comes due.
"We do want big companies behind the Linux ecosystem. There are already many out there (Red Hat, Canonical, among them) doing just that."
Canonical isn't that big on a scale of Red Hat, and is rounding error on a scale of Oracle. They punch well above their financial weight and good luck to them
I don't think corporations are evil. In fact, I bet most people on HN don't think that.
We do want big companies behind the Linux ecosystem. There are already many out there (Red Hat, Canonical, among them) doing just that.
So if I have a budget for paying for a Linux distro, where do I want my money to go? Red Hat, Oracle, or Canonical? In my mind I'm ranking them
1. Canonical
2. Red Hat
3. everyone else in the world
4. Oracle
You need to fix that perception before Oracle Linux will go anywhere. Freatures won't do it. You could offer free support for a year and I wouldn't do it.
At the end of the day, most of us know that the value out of a 'support' contract isn't really the support. It's knowing that the product will continue forward. So who do I want my budget to fund?
>I don't think corporations are evil. In fact, I bet most people on HN don't think that.
I don't think that anything is evil per se, but I do think that corporations and modern businesses are defective by design. Why are they defective? Several reasons.
A) They are frequently more populous than can be effectively managed (also see: Dunbar's number).
B) They are unethical by design since money can be made by simply throwing money in a particular direction. This make no fucking sense whatsoever. Imagine if money didn't exist and we only had physical goods. It would be the equivalent of having 100 logs, putting them on a particular patch of ground then ending up with 200 logs a week later. This just makes no sense.
C) The are unethical by design since the wage gap is huge and the scenario is essentially that many hundreds or thousands of people are working to make a tiny minority extremely rich. It ain't slavery, Bob, but it sure as fuck smells like it.
C part deux) They are unethical by design because when a worker creates something, it is by default not his. Common sense dictates that what comes from your hands belongs to you. The rules of business are such that what comes from thousands of peoples' hands belongs to one person. If there was ever a scam...!
D) The purpose of the business is to ensure it's own survival. A more enlightened approach would be a business which ensures the parent society's survival. This makes more sense because when an entire society flourishes, so do all the businesses contained in it. When a single business flourishes, other businesses and, by extension, parts of society will suffer for it. In other words, the world ain't zero sum and modern business practices do not appear to take this into account.
I don't see how this story doesn't end with a whole bunch of people having to convert back to CentOS when someone at Oracle questions the value in offering this distribution for free.
I could imagine Oracle's long term goal would be to replace Red Hat's role as The Linux Distribution for commercial stuff. One way to do this would be to start this way. Without giving their stuff for free, they would never gain significant share of the total market.
I've tried to analyze this a dozen ways and so far the only one that makes sense is that Oracle has decided to kill RedHat. They want a billion dollar OS business and RedHat has that, but they can't just buy RedHat because with Linux the guys just quit and set up service as 'BlueHat' and continue on their merry way and the customers follow them.
So. The only way to kill RedHat is to take all of their customers. Since RedHat has a GPL'd product Oracle has just as much right to give it away as RedHat does. They try to convert 'free' customers to 'support' customers when a problem comes up.
If they are successful, people migrate to their version of Linux, RedHat dies, Oracle gets the support business because they are not dependent on Linux support revenues to survive. "Poof" Oracle 'owns' a billion dollar OS business. Their investment in getting it is probably less than a 'normal' buyout price if RedHat was a commercial software company.
Of course if they don't convert customers fast enough I'm sure Larry will grow tired of this ploy and flush it down the toilet. At which point people on Oracle Linux will be either sudden new RHEL customers or back to CentOS/SL.
Normally centos users would "grow into" payed support from RH. Now Oracle claims to give them that support without cost.
Unbreakable Linux was never an attempt to have their own distribution, it was to undercut RH, and now they're found a new way to do that through Centos.
It may cost Oracle money, but it costs RH more money, because, after all, they actually build a distribution.
Ideally, kill/buy RH, phase out Unbreakable, and cram in proprietary Solaris. But for now just destabilizing the RH ecosystem has to do.
It's obvious why Oracle is doing this. It's the same reason Red Hat does Fedora. It doesn't cost anything and eventually, you might reach a size or a problem where you need support. If you do, and you are running the free version of a company's Linux, that is the company that you will inevitably turn to (with your cash) for support.
Not to mention Red Hat is their competition in the Java space. Capturing people from Red Hat's ecosystem and moving them towards Oracle's is a pretty aggressive move. It's war on Red Hat. And also on Microsoft, which is using CentOS as a weapon against Red Hat on its Azure platform.
Because if you're running Oracle Linux for free, when you need support you're more likely to start paying Oracle than you would be if you were running CentOS or Scientific Linux.
If you switch to Oracle Linux and if you have some major problem you can't fix, then the next logical step is to pay for support. It would be a simple step to take since you're already running the official version.
Particularly because this is a weak point for Red Hat: they've even orphaned paying customers in the past and there's currently no supported migration path for CentOS. If you're using CentOS or Scientific and decide you want support, there's significant value to being able to pay for support without being told to do a bare-metal reinstall of everything first.
I'm not an Oracle hater (I just dislike them a little). But if this is true, then they have an economic incentive to _not_ issue free fixes in a timely manner. Or, (not to sound to tinfoil-hattish) to release _slightly_ broken things that don't impact an individual installation, but impact clusters.
Again, I'm not saying that they _will_ do the above; just that there is economic incentive for them to bring users into paying for support, and as long as they give it away for free, no one will bite.
There's an old wives saying (used in a different context), that says "why would a guy buy a cow, if he can get the milk for free?"
wdaher, I don't question for a moment your intent and your altruism for this project, and please don't take my criticism of Oracle as anything personal.
I agree with you that corporations are not inherently evil. However, I am of the opinion that Oracle is inherently evil, and it's their business practices, pricing practices, and licensing practices that are ruining their reputation in the industry.
I form this opinion from personal experience dealing with Oracle over hardware, database, middleware, support, and operating system issues with my current employer. Oracle here is not well loved and the concept of converting CentOS instances to Oracle Linux is greeted with peals of laughter at the very suggestion.
I wish you luck in your endeavors, but in your enthusiasm please also keep in mind that Oracle is not known as the most trustworthy company on the block and people are righly justified in being wary of Oracle's intent in providing a "free" product or service.
Just an FYI but I'm fairly certain wdahher is Waseem Daher who is the cofounder of Ksplice. So if you're wanting to hear it directly from the "horse's mouth", this is probably a good place to ask some questions.
This is a really smart move. I started typing a huge write up why, but there are too many reasons why this free product/paid support model are good for both Oracle and their future clients.
That matches Scientific Linux in terms of full time paid packagers/QA &c
I'm not sure how much paid development time PUIAS has but there is some as it is the system used at the host institutions.
I imagine that the Oracle Linux team has plenty of extra people providing the paid support.
What I find interesting is that you can't simply download an .iso, you have to create an account or run a dodgy script. Strange combination of free and corporate approaches.
An excellent question. Honestly, because the page was intended at people already running CentOS -- I thought people would be much more inclined to switch a system they already have than to reinstall a brand new one. Clearly I was mistaken :)
Would you expect those who tend servers to run a script that changes the update repositories and installs new packages as root on a production server without testing the operating system first?
I take the point that the script and the reversion of the changes introduced by the script could be tested on a development box running CentOS.
Disclaimer: I am not part of the target market for your product. Just a desktop RHEL clone user.
> I take the point that the script and the reversion of the changes introduced by the script could be tested on a development box running CentOS.
Yes, that was my thinking here. That you'd spin up a CentOS VM or run it on a test system, see if you were happy, and then you'd run it for a while. Once you were happy, you'd then run the script everywhere.
I've dealt with both RHEL and OEL support contracts over long periods of time, in large scale environments (a few orders of magnitude larger than your example). OEL was free as we had other Oracle licensing in place and the thought was "Why not switch and save money?"
We eventually switched back to RHEL. It's well worth the money for any serious player. The "support" from OEL seemed fairly amateurish -- for example broken hotfix packages that appeared to be one-off builds.
This seems almost mean. Its a billion dollar company calling out a volunteer group of a couple of guys for being slower than them.
Part of the reason I switched to rhel was exactly this graph, I didn't have a lot of faith in centos as a going concern given the delays on 6. But also, its clearly not easy for a reason. There is a lot of work to be done, and the farther you trace your way up the rhel/epel/fedora tree the more you realize this is a community providing an insane amount of value and deserves to be funded.
So if Oracle is going to sell me on using and/or paying for their distro, being faster than a handful of volunteers isn't gonna do it. Going toe-to-toe with redhat on funding, contributing, producing and supporting open source software such that I want to fund you is what'll do it.