"Horst H. von Brand" <vonbrand@xxxxxxxxxxxx> writes: > Currently you have *6 months* to move over to the next Fedora... Sure, and in the vast majority of cases I make it on time. It's just for a few not so important services. >> It wouldn't be good for anything too serious, > I.e., completely useless. Useless for you, not for me. >> but if noone steps up >> for a necessary security patch, I could do it myself > > Sorry, I very much doubt that. You doubt that I could? I'd generally just rebuild the SRPM from a newer release or from CentOS. It's not particularly hard, even for me. >> and then get to share the result with everyone else on that >> obsolete release. I would have to test all updates locally, but for >> services on the way out that is generally easier than testing the >> full upgrade. > It is even easier to migrate to CentOS and keep going with that one... the > difference is that the bumps in the road are much bigger, but you have a > couple of years to migrate your stuff. It is in fact quite difficult to migrate to CentOS, unless your timing is very lucky. CentOS and RHEL live in the past most of the time. If you need a newer kernel feature, you end up running RHEL+custom kernel, and even running your very own mini-distro is better than that. >> If it breaks something for someone, well, they should be more careful >> which updates they apply, especially once official support is gone. It >> might even encourage an upgrade and perhaps keep one box from being >> compromised. > > How is that less work in the longer run than just keep upgrading or moving > to some LTS distribution? In the longer run the server gets shut down. If the service is to be kept alive for more than perhaps a year, upgrading or moving to some LTS distribution are indeed the only options. /Benny -- fedora-devel-list mailing list fedora-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/fedora-devel-list
