Review: CentOS 6.0

Main Screen

Red Hat Enterprise Linux (RHEL) 6.0 was released a little less than 9 months ago, while RHEL derivative Scientific Linux 6.0 "Carbon" was released a little less than 5 months ago. Yet, it took 8 months after the release of RHEL 6.0 for CentOS 6.0 to be released. Two weeks after that, the CentOS 6.0 live medium was released, so I'm trying it out and reviewing it now.

So what is CentOS? Well, before I say that, it's important to know that RHEL is the flagship product of Red Hat; it's an OS meant for business and server environments, and its price is almost entirely for technical support, which can be purchased at different levels. It isn't available for home users for free; that's where Fedora and CentOS come in. Fedora is basically like RHEL's upstream, in that every few Fedora releases, a new version of RHEL based on that Fedora version is released; for example, Fedora Core 6 formed the basis of RHEL 5, while a combination of Fedora 12 and 13 formed the basis of RHEL 6. CentOS, on the other hand, is RHEL's clone; it is 100% identical to RHEL, except that all Red Hat branding is replaced by CentOS branding, and all references to Red Hat and RHEL are replaced by references to a more generic "upstream" or "Enterprise Linux", due to Red Hat's policies regarding its logo and name. Replacing the branding and rebuilding the packages is not trivial, and as far as I can tell, this release was particularly problematic, which is why it took a full 8 months (as opposed to the typical 1-3 months) after the corresponding RHEL release for the new CentOS version to be released.
Unfortunately, this unprecedented delay irritated some of the more vocal members of the Linux community, and as Scientific Linux 6.0 "Carbon" was released many months before CentOS 6.0 (while typically the newest Scientific Linux version is released after the corresponding CentOS version), former users of CentOS started switching to Scientific Linux. Well, CentOS 6.0 is finally here, so hopefully those disgruntled users are happy.

I tested CentOS on a live USB made with MultiSystem. I did not test the installation because I did that with Scientific Linux, and I don't see any reason why the installation should be any different; this is the same reason why I don't test the installation for every Ubuntu derivative I try out. Follow the jump to see what it's like. Also note that I originally wanted to do a direct comparison between Scientific Linux and CentOS, but I got impatient in waiting for the CentOS 6.0 live CD to be released, so I went ahead and tested Scientific Linux 6.0 "Carbon" by itself. Therefore, this article will make frequent references and comparisons to Scientific Linux. Finally, as was the case with the Scientific Linux review, though CentOS is really targeted at enterprise users and servers, I'm going to be reviewing it from the perspective of a home desktop user. Why? Well, Microsoft Windows 2000, the first home version of Microsoft Windows to be based on Microsoft Windows NT, which was previously just for servers and enterprise users, was very well-received among home users despite it targeting enterprise and server use. I'm reviewing CentOS from that same perspective.

Read more »

Read More

FOLLOW-UP: Linux and Breakfast Cereals

I wrote a post a couple months ago regarding Mr. Graham Morrison's assertion that the overabundance of choice in the open-source software community is its failing and Ms. Caitlyn Martin's counter-assertion using the example of breakfast cereals. In it, I mostly agree with Ms. Martin's statements, though I do question the use of breakfast cereals as an analogy because that industry has not ever been so thoroughly monopolized like the PC OS industry.
One of Mr. Morrison's gripes was the confusion in package management, with so many different front- and back-ends. While I still believe that the multitude of front-ends can only be good for users (as anyway most distributions' preferences in that regard are pretty clear), after having talked to a friend in college who is extremely familiar with Red Hat and Fedora, I'm rethinking my stance on the multitude of back-ends.
DEBs are fairly standard (they're just glorified compressed files), so as long as all the required dependencies are present, a DEB from KNOPPIX should work on Linux Mint as well. For RPMs, the situation is a bit more complicated, because RPMs themselves are a good deal more complicated than DEBs. As it turns out, there's a good deal of variation even within the different implementations of RPMs. For example, RHEL (and CentOS) and Fedora use the original kind of RPMs. Mandriva and SUSE have modified the RPM format to fit their needs, as has Scientific Linux (otherwise based on RHEL). Thus, an RPM on Mandriva won't work on SUSE or Fedora; the differences can be so large that it would be like trying to install an alien file format (e.g. DEB).
So my question is, why have all these differences sprung up? For example, the Skype site shows different RPMs for Fedora and openSUSE. (Then again, it shows different DEBs for Debian and Ubuntu as well.) Why can't the maintainers of these distributions pare away the differences as much as possible to maintain inter-distribution compatibility? Wouldn't this just make everyone's life easier?

Read More

Stuxnet, Microsoft, and the Media

There have been a slew of articles about a new piece of malware called Stuxnet which has infected tens of thousands of computers in Iran without the computers' users' knowledge. There's an article by Ellen Nakashima in the Washington Post about how Stuxnet could be used against the US, considering the target of the original attack was probably one of the nuclear power plants in Iran. I wondered what sort of havoc it could wreak on our country's computers. Then I clicked on page 2, and my suspicions — not about Stuxnet's fearsome capabilities, but about its modus operandi and how the mainstream media would report it — were confirmed.
Of course, reading the article again, I should have been suspicious on page 1 itself, considering that "[t]he antivirus security firm Symantec analyzed the worm this summer." Does anyone seriously expect Symantec to be a disinterested party in this? It's a question of computer security, so of course they're going to inflate numbers a little (though whether they've actually done so this time or not is another question) to scare the public into buying their products.
But the second page holds the real "goodies" of this article. Let's go through the major ones.

But "not even two days later," he said, a hacker Web site posted the code so that others could use it to exploit the vulnerabilities in Microsoft.

I should have figured as much. It only affects Microsoft software. Why must the mainstream media equate Microsoft software with all software, considering that in higher levels of the government (e.g. the Department of Defense) Linux is in widespread use for its security benefits? For goodness sake, the military uses RHEL/CentOS!

* It exploited four Microsoft "zero-day" vulnerabilities, allowing Stuxnet to spread automatically without computers users' knowledge.
* One vulnerability allowed the worm to spread via the use of a thumb drive or other removable device. That flaw and one other have since been patched.
* It is autonomous - it requires no hidden hand at the control stick to direct its moves. [...]
* Once it found its target, it was designed to inject code into the controller to change a process. What that process is, is not yet known.

All of these have to do with the fact that Microsoft Windows automatically elevates users to administrator privileges and grants executables administrative privileges as well, so of course this virus will spread without the user's knowledge, spread via removable media, spread autonomously, and inject code autonomously. With Linux, the concept of user privileges (as well as the way Linux handles executables, which is very different from Microsoft Windows) means that this sort of thing would require a lot more effort to execute. And don't counter with Apple's Mac OS X; a recent Secunia report has showed that Apple software has experienced more security vulnerabilities this year than Microsoft software.
So please, Washington Post: don't conflate Microsoft software with all software, and please do some more of the investigative reporting that made you famous in the 1970s with regard to a certain president; is that too much to ask, in this day and age?

Read More

Red Hat, Fedora, and Small Towns

I was visiting $relative's house in a small town, where access to the cutting edge in technology is not the easiest to come by. First, I was surprised to find a digital photo frame (and a big one, at that) hanging on the wall. Then, $relative asked me to turn on the computer so that I could show some recent family pictures to other relatives. I turned on the computer, expecting to find nothing more than the Microsoft Windows XP loading screen. To my shock (which quickly morphed into delight), I was greeted instead by this (note: not verbatim):

GNU GRUB Version 0.97

Please choose which operating system to start:

Fedora Core 6

Microsoft Windows XP

I wasn't sure if $relative used Fedora or Windows, so I left it at Fedora. It stopped at the login screen; by then, $relative came in and restarted the computer (to start with Windows) as $relative didn't know the login information for Fedora, so the question of how their Fedora installation was became moot.
(For those who don't know, up till version 7, the official Fedora releases were termed Fedora Core because the base installation only had the official Red Hat code, while the extra downloadables were community-supported extras.)
This leads me into my next point: RHEL 6 is coming out soon! The great site Dedoimedo has a nice review of a beta release of it (though this beta is fairly representative of the upcoming official release, as far as I know). The review also expresses excitement over the near-simultaneous release of CentOS 6, CentOS being the free-of-charge community version of RHEL, as it brings a stable yet modern operating system to the desktop once more. (Fedora was supposed to be the community version of RHEL when it started, but it decided to go for cutting-edge technologies while RHEL stayed with stability above all else.)
But, Dedoimedo, may I remind you of Scientific Linux. It has all of the stable goodness of CentOS with the added goodies of proprietary codecs preinstalled as well as some useful scientific tools preinstalled.
So let me say that while I am certainly excited about RHEL 6 and CentOS 6, I am really excited about Scientific Linux 6.

Read More