The comment may be a bit snotty, but it serves as a good lesson for other blogs or media outlets that are trying to review Linux distros and post their results into the ether of the Internet. The review in question in this case is one for CentOS 5. Or, I should say, it is a review for the CentOS 5 LiveCD that proclaims to be a review for the enterprise release of CentOS. It’s an important distinction to make and, if you’re trying to catch a break as a Linux review site, you should probably know the difference before your fingers hit the keys to type out a headline.

First, the juicy bit, or “controversy” regarding CentOS 5 Live CD, for reference:

Conclusion: So is CentOS 5.0 worth every penny? Not really. A live CD of its size should have been a little more complete. Even the smaller-sized Zenwalk can do much better than CentOS live. If its main purpose is merely for testing or for rescuing a broken system, lots of Mini distributions can do the job just the same. CentOS 5.0 live CD edition shouldn’t have been released in the first place because it is not ready yet. For now, I wouldn’t recommend downloading it because it’s just a waste of time. But to those who are really eager to try CentOS, perhaps the DVD installer version will do just fine based on the good things I’ve heard about it in some reviews. Maybe I will download that version also, that is after I have fully recovered from my disappointment with the Live CD.

Now, back to what I said about knowing the difference. The CentOS 5 DVD is something we’ve touched upon here at SearchEnterpriseLinux.com a number of times in the past, more recently for a series of Linux support articles where CentOS was featured thanks to its update sans subscription support model (which we compared to Red Hat Enterprise Linux 5). We spoke with enterprise level customers doing enterprise level things in their enterprise level data centers, and they provided us with real world examples of how this OS was helping keep their businesses running. The LiveCD, on the other hand, is a recovery tool usually found right at home on a workstation. If comparing it to the DVD seems like a venture in comparing spherically shaped orange and red types of fruit, you are not alone in your thoughts.

A Live CD definition via Wikipedia:

LiveDistro or Live CD is a generic term for an operating system distribution that is executed upon boot, without installation on a hard drive. Typically, it is stored on a bootable medium, such as a CD-ROM (Live CD), DVD (Live DVD), Floppy (Live floppy), USB flash drive (Live USB), among others. The term “live” derives from the fact that these distributions are a complete, runnable—i.e., “live”—instance of the operating system residing on the distribution medium, rather than the typical case of a collection of packages that must first be permanently installed to a hard drive on the target machine before using the OS. A LiveDistro does not alter the current operating system or files unless the user specifically requests it.

Used incorrectly, I would assume any piece of technology, no matter how masterfully designed, would yield less than stellar results (well, except maybe Mac OS x ;-P). That said, if we’re talking about CentOS as an enterprise distribution, the facts we’ve assembled here at SEL would appear to show this OS has the chops to run a business as long as the user is willing and able to independently handle support. There are handful of other OS’s out there too that we’ve covered with similar results (Debian and Ubuntu come to mind).

But as far as CentOS goes, it is “the” clone right now, says Beranger.org (who laced into Bohol’s post on CentOS, naturally):

Being the most popular of the rebuilds of RHEL, you have a very good chance to benefit enormously from their mailing lists. Also, if some blunder occurs upstream, you might find a workaround in the CentOS community even before the upstream comes with a fix!

Installing CentOS 5.0 right now has a slightly annoying downside: you’ll have to pull tons of updates right away. Note that CentOS 5.0 issued “already obsoleted” media, for they understood the binary compatibility ad litteram: they have included the exact same versions of the packages as per the upstream RHEL 5.0 install media. It’s just they released several months later, and updates were quite a lot…

CentOS also has some extra repos you might benefit of. However, they were much more consistent with version 4 than with version 5. Last but not least, Karanbir Singh does a great job with his extra packages too.

CentOS is simply a different approach to the whole Linux support puzzle. Some people pay a subscription because it fits their needs, and others use IRC and mailing lists to fill in the holes. My highly unscientific opinion says that there are more people happy with CentOS in their environments than not,so any budding reviewer should probably get in contact with a few of them before writing any articles. I know we did at SEL.com.

Of course, we mustn’t be completely negative with this post, and as I write it I realized there could be an issue out there worth looking into that stems directly from this little parlay into the world of OS reviews. Do a majority of users understand the difference? Instead of attacking the review, which gets its CDs and DVDs crossed up, could we instead ask if there’s an issue in how the two are defined? I find that, unfortunately, in cases like this, the Linux community attacks the messenger, instead of addressing the overall problem: Not everyone on the planet Earth understands, uses or even likes Linux. Instead of labeling them an idiot, or a noob or whatever, maybe some hand holding is in order. Who knows, you could ultimately be adding another member to your ranks.

Regardless, this whole exercise is a lesson in headline writing; when you post a headline that promises a review of an operating system, make sure it’s the right one!

The query isn’t out of the blue. To close out 2006, SearchEnterpriseLinux.com conducted a broad survey of its readership about a host of present day Linux and open source software issues. Support, surprisingly or unsurprisingly depending on where your experience with the matter lies, made the list as a “barrier to adoption.”

This is not to say that the support structure for Linux and other open source software is viewed as inadequate. Just three years ago SearchEnterpriseLinux.com reported that only the most diligent of IT managers were considering Linux or open source in their data centers. At the time, Linux deployments in mission critical areas just didn’t happen as they do today because they lacked the standard commercial support that businesses were accustomed to receiving from IBM, Microsoft and others. This perception, some of it earned and some not, was one of the main reasons SUSE Linux couldn’t compete and its founders chose to sell to Novell.

Since then, however, Red Hat dropped support of older Red Hat Enterprise Linux (RHEL) versions to streamline the cost of support. Ubuntu, based on Debian, had commercial support in two years–an accomplishment Ubuntu corporate sponsor Canonical Ltd. said indicates the community got the message that support is of extreme importance. HP made a bundle supporting Debian in Europe. The list goes on.

But what also goes on is this residual effect wherein IT managers not accustomed to Linux continue to view it as less supported and more difficult to administer than Windows, for example. Fair or unfair, it’s out there, and as of the end of 2006 IT managers — and these are somewhat pro-Linux guys who read SearchEnterpriseLinux.com, mind you — are still listing support as a barrier. This is even in light of the cost savings and success associated with Linux and open source software. IDC, as you may know, says open source will go gangbusters in 2011,with revenues through the roof.

So some of this is a skewed perception, the result of Old Guard admins who were raised in the late 1980′s and early 90′s on a diet of Unix and Windows, and who are not as familiar with Linux as they could be. These guys are aware of the forums and mailing lists and the fact that Red Hat sells a very robust (albeit premium-priced) support structure, sure, but are they *really* aware?

Green, in our discussion last week and subsequent podcast this morning, doesn’t seem to think so. He thinks some due diligence and re-education is in order; some hard work on the part of IT staffs to demonstrate the savings of open source AND the richness of the support offered by the major distros and leading commercial open source companies to their higher ups and CIOs.

But that’s not all there is to it. There’s also a documentation problem. Documentation exists but it’s not uniform across the board. Green suggests commercial vendors put some serious effort into making their documentation hum. And this is all while they re-educate users on the importance of community forums, chat channels, and mailing lists. In other words, the support outlets that some open source advocates take for granted.

There’s so much more to be written about this topic. It’s as broad as it is deep, and it’s definitely on the radar of more than a few IT pros out there in the trenches of today’s data center. Information and education are key, and it’s something we here at the Log and SearchEnterpriseLinux.com take seriously enough to dedicate a substantial amount of time and effort into addressing.

All that said… What’s your criteria for evaluating an open source company? Linux vendors? Linux communities (Ubuntu, Debian, etc)? How far have things come, in your opinion, from even just 4-5 years ago? Are users today still doing a lot of their own support? As for the criteria for evaluating support question, here are some examples: time of response, quality of first contact support person, how quickly your question moves up the help ladder, exclusions from contract or things not covered, flexibility, cost, etc. Choose your own adventure, but with support.

It’s been a long time coming (perhaps too long, eh Jan?), but SearchEnterpriseLinux.com will soon be running a regularly updated series on the broad topic of Linux support in the data center. Having a user-generated list of support criteria and some testimonials in place will not only help you get your job done better (and save a buck or two), it will also help your fellow open source community leaders do the same. You guys swap code all the time. How about we try the same with support war stories?

The Debian Project is pleased to announce the official release of Debian GNU/Linux version 4.0, codenamed etch, after 21 months of constant development. Debian GNU/Linux is a free operating system which supports a total of eleven processor architectures and includes the KDE, GNOME and Xfce desktop environments. It also features cryptographic software and compatibility with the FHS v2.3 and software developed for version 3.1 of the LSB.

It’s finally out, and not a moment too soon. With Ian Murlock calling the project a “process run amok” and Ubuntu gaining momentum everywhere but in the kitchen sink, things were getting a bit hectic.

UPDATE@10:35: Seems Debian also has a new leader as well. iTWire is reporting that Sam Hocevar, a French developer, who has been with the project since 2000, was elected as leader for 2007-08 on Sunday.