They are already disliked in the open source community. This is not helping their case. I am not sure what they are going for. This sets them up again as a giant corporation attack the little guy. Geeks and nerds usually hate when little open source projects get attacked.
Why didn't they just offer paid support for CentOS? Why not dedicate full time engineers and donate?
Can't tell you how many time government and military refused to pay less for our product running on CentOS and instead wanted to pay lots more for a Redhat license so they can have support. Nevermind that once setup they only have to do yum update to get the latest security patches and not much else.
There are some companies that offer CentOS support but they are not as prominent perhaps. I think Oracle should have done that, and in the process win some brownie points with the OS community.
From their point of view pissing off some idealistic OS nerds wearing T-shirts and flip-flops is nothing to worry about. However, the problem is a lot of these nerds end up making technology choices, directly or indirectly by advising their managers what is better, etc.
The government (at least in the US) prefers RHEL more for consistency and accountability than anything. RHEL also has Red Hat standing behind it, which, for security, is important (as a read through the entire CentOS codebase would be cost-prohibitive).
Ironically, as a t-shirt/flip-flop nerd myself, I was hoping that this would be something that would appeal to nerds, rather than pissing them off: "Hey, Oracle is giving me this thing that sounds useful for free".
The product itself is appealing. The issue is the company behind it.
I don't disagree with anyone that is wary of Oracles motivation with this product. I also don't disagree that the product itself is nice. But given Oracles known interactions and history with open source an attitude of "fool me once, shame on you, fool me twice, shame on me" isn't unwarranted.
Not sure how to measure "better" in this case. Better for RHEL clones usually means "less changes" than other clones and close to upstream. I guess "better" is support and speed of updates. So in case of a software project you can usually just look at the features and say "oh, look, can see how they added a,b,c and now it is exactly what I want" This is a cloning project, so the best feature is how closely it tracks upstream. But one can't know that unless one switches to it first, and I am guessing, there is enough antagonism in the community that many will not make that first step.
Not only due to the litigious bent. Many past events left a bitter taste - the management of the MySQL products, the end of OpenSolaris (which was a very good Unix distro - and one that lives on) and so on.
This can be interpreted as an attempt to suck the air out of CentOS and Red Hat at the same time. Considering Oracle's past, the hypothesis cannot be ruled out easily.
That's interesting. The company that I work for (which is known for a poor reputation in some tech circles) just hired developer to work on CentOS fulltime as an effort to contribute back to the project and the ecosystem.
We use CentOS on basically all of our Linux servers, so it seems appropriate to support the project financially.
Paid support may tie your hands too. Nokia used Red Hat Enterprise Linux and found a bug in its network stack. Patching it themselves would have invalidated their RHEL support contract. Instead, they had to listen customer complains for another 6 months before official RHEL patch arrived.
Last time(2008) I dealt with Oracle support in regards to linux, they did provide support for CentOS, but funneled customers to oracle linux. IMO they have a good model for helping ensure oracle linux doesn't become a fork of redhat, by creating emergency patches on their side, then sending code/reasoning to redhat to get included CentOS/RHEL. They are helping the quality of redhat's product as long as they get good margins on support contracts.
> Why not dedicate full time engineers and donate?
Do you know ANYTHING about CentOS ?
They aren't exactly the most open and accommodating OSS project around. And with the very worrying communication issues around that developer who went AWOL I am surprised you think Oracle would give them any money at all.
Yes, they had that issue. And some people switched to Scientific Linux. That is not different than relying on Oracle and find out overnight that management changed and now the project is discontinued. In large companies whole department get laid off and focus changes.
Has Oracle tried to help. How about a donation first.
> I am surprised you think Oracle would give them any money at all.
I would be surprised but not for the same reason, but just because it is Oracle. It doesn't exactly have a good standing with the open source community. If it donated, I would be shocked. Instead if chooses to trample.
CentOS is well known, used by many, it had its issues, and it seems to me Oracle instead of trying to help is trying to strangle is replace it. I don't see that as a good move.
The main problem is that Oracle's open source track record speaks for itself.
Now they suddenly want to play "nice", and so they do it by forking a well respected, volunteer driven distro rather than simply contributing to it and helping streamline the security update process. And then they top it all off with a negative ad campaign.
I'm sorry, but you're still not getting it, Oracle.
Security updates are available in patch form pretty quickly, and can be compiled pretty quickly if needed. Has everyone in the tech media forgotten that people can actually compile RHEL/CentOS from Source RPMs?
To be honest i can only find marketing blabla and a very poorly written pdf. The only thing that stands out is ksplice (which in itself is a prime example for Oracle being "evil"!) and a 75% performance boost claim without proof (i read those claims often on the internet and most of the times they are vastly exagerated).
They're probably referring to this, from http://en.wikipedia.org/wiki/Ksplice. I've read the creator of Ksplice explaining that the dropping of Red Hat support wasn't as abrupt as it sounds here:
Ksplice is an open source extension of the Linux kernel ... It was developed by Ksplice, Inc. until 21 July 2011, when Oracle acquired Ksplice and started offering support for Oracle Linux. Support for Red Hat Enterprise Linux was dropped and turned into a free 30-day trial for RHEL customers as an incentive to migrate to Oracle Linux Premier Support.
The "real" engineering was done by MIT students. Before Oracle bought ksplice.
You probably didn't notice but the whole thing about linux and this "open source" everyone is talking about, is about contributing back and community. Not about buying and shutting down stuff. Guess what Oracle is known for and what RedHat is known for. Both make money.
Non sequitur my friend. Just because someone contributes to Linux projects, doesn't mean they are a. Not making money for their services, while contributing back to the community he uses to gain his income, or b. he has a job (related or otherwise) that provides him with an independent income.
You seem to be saying that the Ksplice guys don't deserve to get paid for their hard work, because "community, maaann". Which is an odd sentiment for a site based around startup companies, but ho hum. If they had been acquired by Google would that be any better?
One ironic thing is that, at present, apparently the only way to run Oracle Linux without registering an account on oracle.com (and agreeing to an awful lot of legalese) is to install CentOS first and then use Oracle's little upgrade script. Unless that somehow requires registration later on.
Interesting, thank you. I wonder why you can't get there from linux.oracle.com (or why the mirrors don't all mirror the same stuff). It'll be interesting to see what the state of Oracle Linux and those wiki pages is a year from now.
Why doesn't anybody EVEN TRY to make a comparison with Debian Stable or Ubuntu Server LTS in any of the discussions around this topic I've seen so far?! Maybe my company wants to switch from CentOS and will go with Ubuntu or Debian. Shouldn't Oracle guys try and convince me to go their way instead? (This "over-targeted" approach from the Oracle marketing guys seems more like an anti-CentOS campaign than a pro OL one in this context... and they seem to compare with RHEL only on price - 0 vs something - because they know they are otherwise out of their league).
It's not like these alternatives are not in the same league now, with the latest 2 Ubuntu Server LTS seen as rock solid by most and good commercial support available...
Debian/Ubuntu are not competing in the same space. The only reason you would've been running CentOS in the first place would be if you needed redhat compatibility, most likely so you could run a commercial application that was only supported on redhat, most likely an application from oracle themselves.
I know what you mean (though lots of people run CentOS/RHEL simply because they got used to it and Debian/Ubuntu Server was not perceived as stable enough some time ago so they first started deployment), but there's no reason for a company with Oracle's resources not to try and make a true general purpose server Linux distro... otherwise they'll just wake up to a future when their customers will ask them for debian binaries and support for them and they will have to provide this in order to keep their customer base growing, and when they'll go this way, their whole Linux business will roll faster and faster on the downward slope... I never understood why big corporations have such short sighted strategies. I understand that focusing on short them profit would make sense for a startup, but a company like Oracle should really be aiming for long term growth (after they clean their "evil" image, if they ever manage to do this)...
There's no benefit to Oracle in providing a general purpose server linux to compete with Debian, because there's no profit in it. They're providing this linux primarily so they can sell a vertically integrated stack with their applications on the top, and secondarily to get some revenue from the kind of companies that insist on big name support for anything they use. Debian does not compete in that market, and they're unlikely to have customers demand debian binaries because the kind of companies that run debian don't run oracle applications.
so oracle says during 2011, centos was slow with updates.
Centos responds: no we weren't.. look at this chart of 2012
CentOS was terrible with updates during 2011.. and there's no guarantee that it won't happen again (after all, it isn't paid support). 2011 is the reason why I deployed several RHEL webservers instead of centos recently.
Q. After RedHat releases a security errata, how long until it shows up in Scientific Linux's errata?
A. Within a couple days.
Q. That seems like a long time for errata, why so long?
A. RedHat is not perfect, and sometimes their errata completely break programs.
Q. What happens when the people recompiling the errata go on vacation?
A. Because these security errata are part of Fermilab's security procedures, the entire Scientific Linux development team is not allowed to go on vacation at the same time. So there will always be at least one main developer able to do recompiles.
AFAIK, it's different enough from RHEL that some binary-distributed software breaks. I never had much trouble with it, but I try to keep enough distance from distro details as not having to worry too much whether it is a Red Hat or a Debian I am using.
Redhat really is the only good risk if you need RHEL and timely updates on a production system. CentOS is a great idea but after last year, I don't trust them any more than Oracle. Different kinds of mistrust, but still.
I completely agree. Forget about Oracle's treatment of the OSS community.. At the end of the day, Oracle AND centos are 3rd parties, and will always be playing catchup to RH's patch releases. So for important or internet-facing servers, if you want to use something RHEL-based.. RH is really the only option.