Linux.com

CentOS did not ADD anything to the operating system aside from removing RH branding so please refrase. There are people who spend thousands of hours programming for Red Hat and it is unfortunate to see someone who did an rpmbuild --rebuild of all RH's software say that THEY added virtualization.

Give credit where credit is due..

#

I have to agree here.

The article makes it sound as if the CentOS guys have done everything..
I don't know if CentOS use older versions of RPM's - I wouldn't expect the versions to be too far different time-wise.

I would like to see more covered on the stuff they actually DO contribute more to though, and do appreciate they contribute greatly to RH.

--
Tr0n

#

CentOS did mainstream yum before RH did (though this is now included in RHEL5). In general, CentOS tries to hew very closely to RHEL while adding a few pieces which have made system management very much easier (RHN is a PITA, and raw 'rpm' should have died or been subsumed by more powerful tools ten years ago), and allowed for integration of third-party packages within the CentOS yum repos. To this extent, CentOS is somewhat more responsive to end-user needs than RH has shown themselves to be.

What CentOS also accomplishes, however, is to commoditize Red Hat's Linux offerings. In much the way RH sought to commoditize Sun's Unix offerings. Oddly enough, Sun is among the sponsors of CentOS. Go figure.

In my own recommendations to management, I suggest that if you're going to go with a RH-like system, you may as well go with RH, because you're ultimately shooting the value proposition in the foot (to mangle metaphors) by cutting off its air supply. Which isn't to say I generally recommend RH: while it offers VAR and third-party ISP endorsements (the principle value), actual support has been<nobr> <wbr></nobr>... wanting (a fact RH has shown painfully increased awareness of lately). The growth of ABRH (anything but RH) deployments many locations, whether Fedora, CentOS, Suse, or increasingly Debian/Ubuntu, has been significant over the past several years. While RH still appears to have legs, the company has played very close to the line of providing useful service and annoying the hell out of its customer base and in particular technical evangelists.

For those wedded to a commercial RPM-based distro, Novell's Suse exists. I vastly prefer Debian or Ubuntu in both server and desktop roles, and generally recommend either over RH/Suse for a host of reasons which largely culminate in "more power, less pain".

User-responsiveness and policy will get you a long way.

Karsten M. Self

#

"CentOS did mainstream yum before RH did (though this is now included in RHEL5)"

Of course the difference is the yum-rhn plugin.

"RHN is a PITA"

That's BS. RHN does more than delivering packaging. It has provising, monitoring etc which CentOS LACKS.

" (the principle value), actual support has been "

Incorrect. Red Hat has been number one on support for n number of years.

"For those wedded to a commercial RPM-based distro, Novell's Suse exists"

That would be getting wedded to Microsoft all over again.

" I vastly prefer Debian or Ubuntu in both server and desktop roles, and generally recommend either over RH/Suse for a host of reasons which largely culminate in "more power, less pain"."

That would really be "unproven. Pretty much no ISV's"

<a href="http://www.redhat.com/promo/vendor/" title="redhat.com">http://www.redhat.com/promo/vendor/</a redhat.com>

#

That's BS. RHN does more than delivering packaging. It has provising, monitoring etc which CentOS LACKS.

The problem with RHN is the per-system licensing requirements. While you could argue that this is a "we want something for nothing" situation, the truth is that a key GNU/Linux feature has been deployment flexibility and freedom from licensing constraints. RHN singlehandedly destroys a large part of the user's value proposition. There's the additional annoying problem of vast amounts of otherwise useful information (bugtracking, discussions, and of course binary RPMs) disappearing behing RHN registered-user-only firewalls. These are complaints coming directly from experience at various corporate settings, ranging from tech startups to Fortune 500 firms. It's a wonderful way to alienate your technical evangelists and make supporting your product more difficult. RH have succeeded beyond all expectations on this last front.

As for monitoring and provisioning, our heterogenous (10,000+ systems comrising RH, Suse, commercial Unixes) environment requires use of cross-platform tools for provisioning (Opsware) and monitoring (Ganglia, Nagios, in-house tools). RHN is not an appropriate global solution.

[WRT RH support] Incorrect. Red Hat has been number one on support for n number of years.

You and I apparently inhabit different universes. Can you tell me more about yours? My physicist friends would be fascinated.

In mine, three tiers of RH customer reps came on site (F-500 tech customer) recently to make Alberto Gonzales's current Congressional hearings look like a love-in. I've never such a contrite performance from a vendor in my life. RH are well aware that their support execution has been sorely lacking (and if this still puts it at "number one" on unspecified criteria this damns the entire field) and may actually be turning things around. I'm just reporting my own experience.

[WRT Novell/Suse] That would be getting wedded to Microsoft all over again.

Which was pretty much exactly my point.

[WRT Debian/Ubuntu] That would really be "unproven. Pretty much no ISV's.

Yes and no.

Most third-party apps I've encountered run just fine on Debian, whether or not they're officially supported. Truth is, they're likely not supported on anything other than RH or Suse, which puts all alternatives on largely equal footing (and as I pointed out originally, this is among the last remaining true advantages of either of these two distros).

A great many Linux deployments don't run any significant third-party, proprietary apps, but are used instead for deploying in-house development, as build systems (basic GCC toolchain), or can be fully provisioned from within the ~20,000+ packages available within the Debian archives. That's over three times the offerings of RH (roughly 6-7k packages last I checked), and among the reasons my own deployments tend to be Debian (or Ubuntu): I've very rarely got any call to go outside the available packaged software (current exception list: Google Earth. MM Flash. That last I'll lose when Gnash is sufficiently developed).

Finally: Ubuntu, being a well-capitalized (Mark Shuttleworth took in over $1b from sale of Thawte), clueful company is well positioned to establish ISV partnerships and is in the process of doing so. This is an advantage of corporate-based distros over community ones, in that they've got the infrastructure to do this sort of thing. Ubuntu has to date played pretty closely to Debian in other aspects which gives it a win in my book over other Debian-based distros (several of which have also inked ISV partnerships) including Linspire, Xandros, and others. RH are on shifting sands here as well.

It's interesting to note that the one notable ISV which is most often mentioned as a reason for deploying RH (often with the phrase "we might want to run Oracle so we must run RH") is, oh, I gave myself away, Oracle. Who last I heard wouldn't even think, no sir! of, say, rolling out their own distro to free themselves of some pesky little complaints about RH such as, oh, just stabbing in the dark, insufficient support levels.

I'm not saying RH are dead. I'm saying they should be gravely concerned about their position and execution.

Karsten M. Self

#

"There are people who spend thousands of hours programming for Red Hat"

So what? There are people who spend thousands of hours writing Open Source that RH release with their product. Gee, it's not like they haven't benefited from the hard work of others

#

I've been working with different distros of linux for the last 7 years and have come to the conclusion that if Red Hat decided to shut down tomorrow, Linux at the Enterprise level would suffer a devastating blow.

The value proposition of Red Hat to the enterprise customers (those willing to pay dollars) is way under estimated. Red Hat brings together bug fixes, documentation, validation and most importantly enterprise quality global support.

Just think that Novell released SLES in 2006 with Xen when XEN was clearly not ready or stable.

An an example, if it were not for Red Hat, the enterprise customers would have been deluded into implementing Xen when it was not stable.

Oracle's Linux, CENTOS, Fedora, etc would not survive or progress without the contributions of Red Hat. There is a place and market for these recompiled versions of the Red Hat source, but it is not at the enterprise.

Linux for the hobbyists and academia would survive without Red Hat just as it started without Red Hat.

Ubuntu is the only alternative to Red Hat and it is still gearing up an organization (has been for the last 3 years). Ubuntu is a wait and see as far as I'm concerned.

The press needs to do an analysis on :
1. What would happen to Linux, at the Enterprise level, if Red Hat decided to shut down tomorrow ?

2. What would happen to mass/cheap web hosting, if Red Hat decided to shut down tomorrow ?

I'm willing to bet the results of such will be sobering.

#

This is kind of a "no duh." If Microsoft shut down tomorrow, Windows would suffer a devastating blow. If Apple shut down tomorrow, Mac OS would suffer a devastating blow.

I don't think anyone is willing to waste their time on an obvious situation. I think that most people are bright enough to understand that GNU/Linux is not a zero sum game.

#

> What would happen to mass/cheap web hosting, if Red Hat decided to shut down tomorrow ?

Someone would fill the void. Ubuntu, FreeBSD, whatever, would show up on newly deployed machines where possible. The world would not end.

#

"Someone would fill the void. Ubuntu, FreeBSD, whatever, would show up on newly deployed machines where possible. The world would not end. "

No but Linux development and deployment will show down drastically.

<a href="http://fedoraproject.org/wiki/RedHatContributions" title="fedoraproject.org">http://fedoraproject.org/wiki/RedHatContributions</a fedoraproject.org>

#

Doubt it. The people working for RH would migrate to other Linux companies and do pretty much the same thing they do now.

#

No biggie. Linux would keep on keeping on, apps would get ported to other systems... and commercial users would look to supported alternatives. Solaris is free now, and well supported.

#

This points up the Achilles heel of free software. Red Hat has put a huge amount of investment into making Linux enterprise-ready. They have contributed most of their work to the community.

Then along comes CentOS, copies Red Hat's code, takes off the "Red Hat" label and substitutes "CentOS" - and proceeds to try to destroy Red Hat's business.

If CentOS succeeds, Red Hat dies and so does Linux in the enterprise, because Red Hat is the company that has the know-how, the resources, and the incentive to keep Linux up with evolving requirements in the enterprise.

BTW CentOS did not "add virtualization". Red Hat did. As far as I can tell from the article, CentOS has added nothing, apart from a few apps whose developers think aren't ready for prime time yet anyway. They're just parasites on Red Hat.

#

You are not very familiar with the GPL license. CentOS is recompiling the available sources to make freely available binaries. When you buy RedHat you are not actually buying the OS (basically identical to CentOS), but the service and support. CentOS simply exists because RH does not offer an unsupported but stable version of RHEL. This is not illegal or wrong, in fact it is the very good part of being open source. Calling them parasites is an offense to the open FOSS community, which you don't seem to belong to. By the way, lots of the "amazing" technology that RH develops is "stolen" (according to your reasoning) from Fedora, which is community based and uses collection of applications developed outside of RH itself. The simple fact of the matter is that it's not stolen, it's for everybody to use it and be modified. Which part of the "open source" don't you understand? Maybe you think in Microsoft terms, so everything non MS is in violation of their licenses, so it's "stolen"?

P.S.The fact that you can try RH technology for free is thanks to CentOS.

#

May I suggest you actually read a comment before replying to it?

"stolen" (according to your reasoning)

I have no idea where you got this from. Certainly not from anything I wrote.

The raw material for a distro is code that the community has developed, but there's a lot more to putting together a distro than downloading a bunch of code. Other distros (for example Debian) use the code that others (including Red Hat) have developed, but they're certainly not parasites - among other things, (1) they have developed their own vision and aims, (2) they contribute significantly to the community. There are distros based on Debian, but again they have their own, different, vision and goals and they contribute to the community.

Sure, the GPL gives you the legal right to copy a bunch of code, or a whole distro, slap it onto a DVD or 6 CDs, and stick your logo on the result. But ethical distros do more than that. It is perfectly possible to stay 100% within the law, and still be a parasite.

#

Amen!

Alot of posts here seemed unhappy with the way the article was written. No hesitation to bash the good Red Hat clone.

Here is what I see in the last few years I have been reading comments from various media outlets, forums, blogs, etc.:

Whenever a distro gains in popularity, some people will find anything to discredit it. Look at the popularity of Ubuntu, PCLinuxOS, LinuxMint, for examples. I saw endless bashing of these distros: they are a rip-off of Debian/Mandriva, they are just Debian/Mandriva with dressed up desktop, they don't contribute anything back, blah blah blah, etc.

These people totally ignore the GPL license which grants anyone the right to rebuild these distros anyway they see fit.

Hypocrisy if you ask me.

#

I'll only say it is open source.
The people making claim that it is stolen software are probably MicroSnot users using something in the range of at least 20% copied software.

I would like to point out that this poster sounds very young.
I say that as the young who have not bothered to learn the English language use things like "alot."
Sorry folk, there is no such word in the English language.
I believe you meant to say " A lot" which means many or much.

Sorry about the Anonymous post.
I'll sign up later when I have more time.
I'll be back!
Pete G

#

Red Hat's licensing is forcing the issue here.

For example, if I have 500 servers, all EXACTLY the same running exactly the same configuration and software (think massive web cluster,) RH's pricing model is 500 X $Retail - 20% (quantity discount.)

Since you are not "licensing the software" really, (since it's all GPL) you are paying for support. Support for 500 clones does not cost 500 times as much as support for one machine. This is EXACTLY why I don't run RH and run CentOS instead. RH really needs an alternative model where you get a site license and pay $X for Y hours or Z incidents of support.

#

centos's success will have no negative outcome on redhat's profit. redhat is the crem de la crem in linux in the enterprise.

#

Hi,

In many, many situations I've installed CentOS at
the evaluation stage of a project where the
budget is tight. Once the system is proven and the
client is happy to roll-out the system into production,
it is often switched to a RedHat installation
as the customer wants the big-company support
behind it - the company that is approved by
the software vendor.

CentOS for me is a great sales/evaluation tool.
It vastly reduces the time/complexity in getting
a single-server install up, running and patched.
Ultimately, it has resulted in sales for RedHat
down the line.

Tony

#

Interesting...
How easy is it to migrate an "evaluation" CentOS installation to RedHat?

#

I was the original poster of the comment on "Enterprise Linux sorely needs Red Hat".

I accidentally posted my comment to the wrong article.

My comment was intended for the article on Linux.com (maybe an external link ?) where the reporter was highlighting a business that was proclaiming how they were saving money using CentOS and that Red Hat did not have a value proposition for their business.

For comments posted on how everything is free and all that Red Hat is doing is slapping it together on a CD, maybe they should try to get the stuff themselves from the CVS repositories.

I have been in the software business long enough to recognize that testing, integration, validation, upgrading, quality documentation, knowledgebase maintenance, separating the software that is hyped (e.g., ReiserFS) from software that works (e.g., Ext3) and a global support organization take a lot of time, dollars engineers and facilities.

#

How many months have you been running CentOS to lay claim to its reliability?

By what criteria do you define reliability?

#

I have been running "a version" of CentOS since it's debut 2003. I can say that all of the releases to date have been stable.

#

So, then ubuntu doesn't add anything on top of debian?



Red Hat itself does not seem to share in your hatred of the CentOS Product:



<a href="http://www.internetnews.com/dev-news/article.php/3671841" title="internetnews.com">http://www.internetnews.com/dev-news/article.php/<nobr>3<wbr></nobr> 671841</a internetnews.com>



In the above article, Red Hat, Inc discusses the advantages that CentOS provides for customers of Red Hat, and also discusses how CentOS helps them engineer their product.



Here are some Red Hat quotes from that article:



"From the technology side, CentOS broadens the customer base for Red Hat Enterprise Linux technology," Nick Carr, a product marketing director for RHEL, told internetnews.com. "They are active in the mailing list, and from an engineering viewpoint they certainly assist us in finding problems in the product."



"We certainly get CentOS customers who will see the value of the Red Hat model when what they are deploying becomes more and more critical to their infrastructure," Carr noted.



"That's not to say that what CentOS offers to customers isn't of great value to customers," Carr said. "It's just not the same as what Red Hat is offering."



CentOS and Fedora recently shared a Dev Room at <a href="http://www.fosdem.org/2007/schedule/devroom/centosfedora" title="fosdem.org">FOSDEM 2007</a fosdem.org>.


Max Spevack (Fedora Board Chairman and Red Hat employee) says in his <a href="http://spevack.livejournal.com/" title="livejournal.com">blog</a livejournal.com>:


I have always thought that the work the CentOS project does is very well aligned with both Fedora and Red Hat, and that Red Hat should look at the CentOS project as one of our biggest community allies. Talking with the leaders of CentOS only reaffirmed that in my mind. They're a good group of guys, and all of their users are folks whose operating system is part of the Red Hat family (who otherwise might not be using a Red Hat-based distro).





Please do a search for "centos" in bugzilla.redhat.com.



CentOS users and the CentOS developers find and fix many bugs in the Red Hat codebase.



Surely you would not suggest that Red Hat would be better off if a user (who was going to use a Free {as in beer} OS) picked debian or ubuntu over CentOS. If they picked one of those OSes, anything they did would not easily be moved to RHEL in the future. If they pick CentOS, thier work can easily be moved to Red Hat's premire paid product, if / when official paid support is needed.



If Red Hat is OK with CentOS and if Fedora is OK with CentOS, then who are all these CentOS haters who are commenting that CentOS Developers are parasites.

#

Centos adds freedom (no license cost) and extra (newer) packages / technologies (such as XFS) that Red Hat deems to be unworthy. The article is misleading / wrong about what Centos adds, but this does not mean that Centos doesn't add anything.

#

CentOS does not add anything. It just remove RED HAT branding from the latest RHEL released. The author is either ignorant of this fact or deliberately mis-representing this. With 10 years in IT the author should know better.

I acknowledge the fact that CentOS is a good effort at proving a free, supported RHEL and I appreciate their efforts - but one must give due credit to RED HAT as well.

#

The power management is broken, udev sometime or everytime (depend on NB type) hanks , shared irg is not properly handled - USB
look to http://bugs.centos.org/view_all_bug_page.php before instalation to notebook

#

Use Slackware cloobs....

#

I think it's great to be able to take advantage of all the nice RHEL features without having to buy a licence. Hey! I'm a student, but I still want RHEL. As for sales, I belive what has allready been written. If CentOS users/developers can help out in bugfixing, it is a deed to RHEL, not a copycat operation. And, CentOS is not making money on doing this favor, so shut up, and enjoy RHEL, with peace in mind and money still in your pocket.
Go donate some $'s to the project instead of whining about it.

#

RHE & CentOS are good working together. In an enterprise enviroment I use large cluster built on RHE + certified HW + certified SW (DB and Apps) and some file servers, proxys, dnss, running over CentOS with great relsults in performance and costs.

#

Red Hat adds a great deal of value with their packaging, bug fixes and institutionalized choices. This makes for a stable environment, unlike Fedora where every time you type 'yum update' something new pops up.

What Red Hat stopped providing is a 'home server' type of product, where I want to run Linux on my 1 home machine. I don't want to assemble all of the possible packages and pick through the inter-relationships (this is more involved than simple dependency checking done by yum). Yet they stepped up the cost to something an enterprise feels is worth it, but a home user may want to switch back and forth between distos and a subscription up front doesn't make that easy. CentOS allows me to run a stable environment without a long-term commitment.

One could argue that if I'm storing my ever-growing email archive on a machine I should be paying for the OS, so I've contributed to CentOS. But not as much as if I had to shell out $$ for MS Exchange.

I personally feel a business should invest in RHEL and home users (and possibly public libraries, etc) are the proper target for CENTOS.

#