Quoting Michael Simpson <mikie.simpson at gmail.com>: see my remarks below > On 4 April 2011 23:11, David Brian Chait <dchait at invenda.com> wrote: >> I have to provide a reliable and scalable infrastructure, and that >> requires a reliable provider / updates. While I do not need Centos >> 6 today, this development cycle has certainly raised questions as >> to whether the development process can be relied upon. The whole >> "when it's ready" mantra works well for academic/individual users, >> but you can't plan business processes based on it. >> > > Yet you can. > The only 5.6 update that has been rated as critical has been firefox. > The previous critical update was exim which was for 5.5 which we had. > I would place this firefox update at low priority as i would guess > that close to 100% of the millions of installations will be running > CentOS on servers rather than on workstations. > Whilst i use CentOS for my desktops, and appreciate the complete > stability that i have enjoyed since deploying 5.0 on these platforms i > really care about my internet facing production servers and these are > not impacted at all by waiting for 5.6 (or 6). > > I am looking forward to 6 coming out but just so that i can play with > it and install it on some boxen that i have waiting in their packaging > but i am in no rush. In the same way i would rather have 5.6 when it > is done. Therefore the business process for remaining on 5 doesn't > change especially with php53 and bind97 in testing so already > available > > Based on previous experience, if there was a critical update for a > core server service (or if there was an issue which was going to be > critical to systems within a certain time zone &c) then it would be > pushed sooner. > > If your business process demands some feature of 6 (kvm / tpm / power > savings / storage drivers) then you have enough money to buy some > licences for rhel 6 to enable your testing and the beauty of CentOS is > knowing that you can then replicate and upscale your testing > environment to production on CentOS 6 without worrying about having to > go though another full testing cycle due to the promise of full binary > compatibility, not sure that you can do that with SL as they have a > different raison d'etre > > With regards to communication to the community IMHO you can assume > that the lack of it indicates the effort required to get 4.9, 5.6 and > 6 out the door and underlines the devs determination to get it right > first time. As evidence of this, follow CentOS mailing list and look > at how many "help" threads are from problems with the core product. > > It must be quite a burden to know that releasing CentOS that isn't bug > for bug compatible with RHEL or is flawed in some way could cause > many, many production servers to fall over. > > I would like to thank the devs for all their time and effort > > mike sure, me too. I run CentOS servers and do all the patches every day and get great value for the money. But that doesn't make me deaf dumb and blind, the project management badly needs work. The firefox issue is a bit misleading for reasons mike points out, but for instance I can't deploy Drupal 7, for which there is a lot of demand, unless I open up non-CentOS repos. Not the end of the world, but one more avenue into the system and one more thing to watch out for. I mean I'd be happy to make a cash donation to get more bodies on the problem (when taken with all the other well-wishers and would-be supporters) but it doesn't seem as if there is a way. Dave > _______________________________________________ > CentOS mailing list > CentOS at centos.org > http://lists.centos.org/mailman/listinfo/centos > -- "It is no measure of health to be well adjusted to a profoundly sick society." Krishnamurti