FedoraThis forum is for the discussion of the Fedora Project.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
Hello to all
I am a sysadmin (well... two weeks of experience is not enough to be called "sysadmin"....) and in my hands (dangerous ;-)) I have a cluster on which people heavy computations. This means that usually they open a thread that works for a while and then get a result. I do not expect a high rate of thread creation but big threads running for a long time.
The cluster is for instance running RH9 and I wanted to know if it was worth upgrading to Fedora (which core ?). I was told one of the big improvements was that in Fedora the threads are much more lightweight. Is there a site which more technical information about the differences between RH9 and Fedora in order to let me take the right decision upon the release my cluster should run.
Although I don't really know what your setup is like, I strongly reccomend upgrading to Fedora Core 3. The new major feature is SELinux which is enabled by default. This is Security Enhanced Linux, which has a lot more in terms of access control etc.
Also, RH9 runs on a 2.4 kernel, doesn't it? FC3 runs the latest 2.6.9 kernel, and www.kernel.org has a LOT of detail about the changes and improvements.
Also, there are nowhere near as many updates etc. for RH9 as FC2/3, especially with yum.
If this server needs very very high uptime, however, it is a different story. Fedora releases are very quick (less than 9 months between versions) and this means upgrades!
If you need a server expected to be online 24/7, maybe try Slackware (http://www.slackware.com) or Gentoo (http://www.gentoo.org - this could take a while to compile - took me approx 12 hours to install (then again, I left it running 'emerge system' overnight) but it is optimized for YOUR exact system)
Not only does the Fedora team release versions rather quickly, the used technology is quite new, like 'bleeding edge'.
If you need to run it for a business that can't have down time, you'd probably want rock solid solutions that use proven technology, like RedHat Server editions, but they cost money.
Indeed, Slackware and Gentoo are an option as well, which is made for people who love to tune the OS to be completely custom to their system (read: optimized), for best performance.
I don't know about Gentoo or Slackware using 'bleeding edge' or 'proven technology' though.
If you are running a server in a professional setting, I would maybe go with whitebox (if you want to stay free). I know people run FC1,2 and 3 as servers but I am not sure of the criticial nature of the usage. I have never had any problems with stability but I do not run a server (yet). I want goggle around and find some options, they are out there.
well...
thanks to all for your answers. About the question "does my clustering software recommend any distribution", the answer is ... no. My cluster is supposed to run heavy physical simulations. It uses mosix, which (I need to verify) is dedicated to RedHat, but we might throw mosix out. Info about mosix for the interested: www.mosix.org. We also run condor, and to make it fun, the nodes (read PCs) are booted on the LAN using CLiP (tool which is a research and not perfect yet from the Hebrew University of Jerusalem) which manages the kernel images and the /etc/ partitions of each PC.
Explainations: No cluster is symetric, there is always new PCs and old ones. This means different kernels and different options in /etc/ (think about XF86Config for different video cards for example), so when it boots the PC gets the loader (Syslinux/PXE) through TFTP, then sends a DHCP request (again) and gets the kernel (LAN), its personal kernel. The tool that knows which kernel should be sent to a specific PC is CLiP.
Wel.. If you are still reading this messed up explaination, I will not try Gentoo (do not have enough aspirin sorry) but will visit the sites you sent me about the difference in the kernels.
Another question: I was told that the NFS server of RedHat is somewhat... fragile, meaning that it fails when the load begin to be problematic, that is why I saw that the critical NFS server runs FreeBSD, do you know if there were and modifications on it (from RH9 to FC3) ?
The question you have to ask yourself also: Do I need a new distro, doesn't the old one run fine?
If this is a productive professional system, you really can't afford downtime (as mentioned before) and upgrading to another system (especially non redhat) will likely cause downtime, especially when things go wrong.
Example? I upgraded RH9 to Fedora 2 and my networkcard stopped working. After a few hours troubleshooting (downtime) I found out that KUDZU, the hardware autodetector of RH and Fedora broke my network drive for the 3COM netcard.
I had to disable it. This with a prior case in this forum, so if the problem is brand new, expect no help. In this particular case, the new software, more efficiency gained from it (which will not be THAT much), is wiped out by all the new problems that kill your productivity and cause massive downtime.
If it runs, don't touch it, if you don't have a good reason for it. This redhat bucket will run great, even if it is not the newest and coolest around.
A good thing to do is to setup a testsystem, where you can find out the problems before it stops your productive system working.
This said, I am always assuming that you are talking about a productive system, which stops people from doing work when it goes down. I am always amazed when I did such stunts how much things run on that particular machine that you absolutely depend on.
(Upgrading your server to the newest distro, run into a problem, google for it and realize that one of the roles your server was acting in was NAT and ROUTER for internetaccess, WHICH does not work anymore now... ;-)
It is always a pull between new features and stability: New systems have new features and promise more efficiency, but are not as stable and reliable and old systems are stable, but boring and lack features.
BTW: Mosix is not at all only for RedHat, you can make it run on almost everything.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.