“Scientific Linux is an unknown gem, one of the best Red Hat Enterprise Linux clones. The name works against it because it’s not for scientists; rather it’s maintained by science organizations. Let’s kick the tires on the latest release and see what makes it special.”
I switched all my servers over to SL this past year. The CentOS project, which I used for several years, is really broken. That became apparent during the EL 6 release. The CentOS team refuses to open the project up to the community and also doesn’t have anyone doing this full time. The biggest issue was the lack of updates while they were working on a point release. Sorry, but you can’t stop putting out patches and call it an enterprise linux. There is also no community in the creation of it.
As for SL, they may not always beat CentOS, but they have people working on it for a living. There are also a great number of organizations that rely on it, so support seems pretty good.
CentOS did have a “blip” before releasing 6 (there was also an issue with 5.X releases coinciding with 6). However, if you hung on with CentOS like I did, they’ve fixed the update issues in fine style now.
Firstly, they introduced a Continuous Repo for both CentOS 5 and 6, which provides the important updates from the next point release very quickly, allowing the CentOS devs to spend more time working on that next point release.
Secondly, they sped up the whole build process for the next point release, so much so that CentOS 6.2 came out only 2 weeks after RHEL 6.2 and well ahead (2 months!) of Scientific Linux 6.2. The thing I like about CentOS is that it strives for 100% binary compatibility with RHEL, which Scientific Linux doesn’t. Hence anything certified for RHEL is also going to work out of the box with CentOS, again not true for SL.
CentOS is now, IMHO, the preferred RHEL clone and it’s probably time that Scientific Linux merged with CentOS because we really don’t need more than one RHEL clone (and we certainly don’t need Oracle’s clone either, that’s for sure!).
+10000000000000000… except I’m glad SL is still around, competition is good. But we can definitely do without Oracle, this one can fsck off and die.
Edited 2012-03-24 11:21 UTC
I’m really sort of curious how much the 100% binary compatibility really matters. Does anyone have any examples of a piece of software that has ever worked on CentOS but not Scientific?
I know many people that have RHEL in production, but use CentOS in development to save a bit on licenses. For them, it makes sense to use the closest possible OS to RHEL, which is CentOS. Those kinds of things are good to be paranoid about.
My current employer does not use a RHEL clone, but another distro. We do routinely get bits of source code ( drivers/utils for some exotic hardware) that are RHEL certified. That doen’t work ( mainly due to kernel versions, gnulibc versions ect) and have some interesting security holes. Im almost glad we run into the obvious versioning problems so it forces us to look at the whole code base to find the security implications, rather than just installing the RHEL certified software and walking away after a successful install.
I have nothing personal against the people making or using CentOS. Its Karanbir’s project, and he is free to run it as he sees fit. But the reason I left is that the dev group is dysfunctional and really doesn’t care to change. The name, Community Enterprise OS, is completely wrong. It should be Karanbir’s Hobby OS. The main problem with the release of EL 6 has still not been addressed as far as I know. And that is to have some transparency in the build process. There is no good reason to keep everything secret. As for the binary differences, I have yet to find an example of something that worked on RHEL or CentOS but not on SL.
/metoo And I thought I was bitching about CentOS.
Scientific Linux is a really nice distro.
There are a few misunderstandings in the article. CentOS and Scientific Linux have to totally separate goals. CentOS is a bug for bug RHEL clone, and SL is to keep the licensing costs down for the various participating laboratories which need a RHEL compatible distro and who just happen to make it available for public download.
Well, talking with my friends that work on Scientific, their goal is really just to have a high performance Linux distro that they have intimate knowledge of and some level of control over while leveraging the best code base around.
There’s that too.
Which is still different then CentOS’s reason for existing.
As someone that used to work at CERN, the article seems very bad written to me.
If memory serves me well, SL was only intended to be used at CERN and Fermilab, among other research institutes that collaborated with them. As a means to have an uniform distribution across all institutes.
It was never intended to be used in the outside world.
Well, yes and no. It was never intended to not be used by the outside world as well. I have been evangelized by some core devs in the past. I would absolutely use it, if it weren’t so gosh darn stable. I like my desktops to break every now and then from unstable packages that I can learn important lessons by fixing them. Like having multiple offsite versioned backups of important documents at all times
You can break it. You just have to use the third party repos indiscriminately and replace pieces with stuff you’ve compiled yourself.
Like they say, where there’s a will there’s a way.
Indeed.
Same goes for PU-IAS, another RHEL clone targeted towards scientific computing.
http://puias.math.ias.edu/
“all those non-paying customers who vowed to never be non-paying Red Hat customers again.”
Heh. That’s kinda funny.
Well, I was going to go to bed, but I guess I gotta download and install first.
I don’t understand why reviewers keep harping on about the name Scientific Linux. I’ve never read a review which complained Fedora wasn’t designed for people who wear hats or that Mandrake Linux wasn’t for people into plants or that Pear wasn’t for people who grow fruit. Why is it that so many reviewers complain about Scientific Linux but none of the others? Are they really that simple-minded?
Because like the reviewer said most people (which honestly included me just a while ago) think that it’s made for scientists. This is a perfectly reasonable assumption given all the distributions built for certain jobs or professions that use computers.
We have a problem with names within open source community. It’s crystal clear as if no one is really interested in the foundations of commercial marketing. Projects come with the most ridiculous or inexpressive names. I know that’s subjective to taste, but I think name has to have a good sounding tune, a catchy feeling. Scientific is no proper name here. As XFCE is no proper name for that GTK desktop. (Although I still think the most ridiculous name ever made is Debian = Deborah + Ian; now you can mod me down…)
Ok, back to topic… well…
I have used CentOS and found it like the article said: bug by bug clone. I was interest in knowing if SL has packages like latest software: Firefox, LibreOffice, Audacity, etc on top of that old GNOME and so on.
Oh, I know what you mean, Windows has been called windows for 25 years, but it didn’t even glass until 2006! There is no fruit, snakes, or woman dressed in fig leaves in the manufacture of an Apple computer, christ, they aren’t even scottish!
Just if you didn’t notice… Free Updates and Errata for Oracle Linux –> https://blogs.oracle.com/linux/entry/free_updates_and_errata_for
ok its a _NO_FREE_ support, (you are on your own), but its good to see a “free” third RH alternative there…
Edited 2012-03-26 09:05 UTC
People are stupid. The name is fine. It is linux for science.