On Sat, 2007-11-24 at 11:07 -0600, Les Mikesell wrote: > Thomas Cameron wrote: > > >> So, to explain in more detail: Fedora was meant to help the development > >> of Red Hat’s codebase with the help of the community. Red Hat uses > >> Fedora (good as it may be) purely as a test-bed, where they can try out > >> new technologies that could prove to be too unstable for RHEL without > >> any risk. > > > > Bullshit. Red Hat sponsors Fedora for the good of the community. Now > > don't think for a second that Red Hat doesn't also do it because it > > benefits, but again, this myth that Fedora is some RHEL beta is bunk. > > A better way to find what Red Hat considers fedora to be suitable for > might be to ask where they use it themselves. Is there a single public > facing server managed by Red Hat that runs fedora? Highly unlikely, but I do not know for sure (I'm not in Red Hat IT, they would know). The reason that this is highly unlikely is not because of quality of product. There are two real reasons why it is not likely: 1) Red Hat's IT group is just like the IT group at any corporation. They don't want to have to deal with daily updates and semi-annual upgrades. That's a *lot* of work. 2) Red Hat does not sell Fedora subscriptions. Red Hat sells RHEL subscriptions. There are very legitimate marketing reasons for "eating your own dog food." Before anyone gets too pissy about this, remember that that marketing is what allows you to get Fedora free of charge. > > To summarize, Fedora is an independent distribution from RHEL. They > > share many bits to be sure, and the Red Hat developers work on both, but > > they really serve different purposes. Fedora is community based and > > community driven, with a very rapid rate of change. It's not > > appropriate for many large enterprise customers who don't want to > > constantly chase the latest bits. But anyone who says that it is not > > ready for prime time is just smoking crack. > > Does Red Hat have any critical data that is stored solely on fedora boxes? Well, that depends on what you think "critical" data is. I suspect that for varying definitions of "critical," the answer is undeniably "yes." But again, I am not sure. Thomas
