On Tue, Nov 27, 2007 at 04:25:35PM +0100, Adam Tkac wrote: > I've put this software to F8 because it has nice new features. Some > bugs is tax for them. Additionally I don't think that users use newest > Fedora on important servers and 9.5 will come into beta stage very > soon. I use the newest Fedora on important servers. They were running FC6 before, and the EOL is approaching soon. I'd rather upgrade once to F8 now rather than F7 now and then again to F8 when F9 is released. I guess that was a bad decision on my part, but I've never had major problems on servers with the latest Fedora before. I've been testing F8 as Rawhide for a while now, so I thought it was ready for my servers. Unfortunately, I didn't test BIND--slap my wrist for that one. Of course, who knows if I would have encountered this problem in a test server--it may be related to the load one puts on the server that would never have been seen in a test environment. I don't mind beta software and release candidates of software in a stable Fedora release--heck lots of software stays in that phase for a long long time (ISC dhcpd for example). But alpha software I think is pushing it a bit too far. This is just my opinion, and I will work around whatever problems I cause for myself by using the latest Fedora on my important servers, but the lack of a policy on this makes it hard for sysadmins to choose correctly. Now it seems that the choice should be "always run the previous Fedora release because the newest one might introduce new software that is considered alpha quality by upstream". It is a fine line to walk on stability vs. new features. Fedora is about being on the leading edge, sure. But bleeding edge should be reserved for Rawhide and Test releases. Especially for software as important as BIND and DHCP. -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
