> Eric Rostetter > > Quoting "Pettit, Paul" <ismanager@xxxxxxxxxxx>: > > > The overhead (work hours) is high if said downloads are on multiple > > servers and updates must be run on each (as well as dependancy > > checking). My dad always said "work smarter". > > But is it "working smarter" if doing so results in your machines or > services going down? I'd say not. > Yes it is if you know that updates will come out a random and un-uniform times. Then you can plan ahead or program something that will correct for the lack of forsight by those in charge of updates. > > It's sort of just one of those times when multiple normally > unrelated > > occurances happen at the same time (i.e. mysql update, > holiday, broken > > restart function in said update). Many ways to fix it but I > think one > > way to help would be to maybe schedule pending updates. > That might be a > > problem but hey, this is a discussion right? > > Yes, it is a discussion. Your solution won't work. But maybe > we'll get > a solution that does work by having this discussion. > > -- > Eric Rostetter > Actually I possed no solution except re-writing the cron script (and that from a idea by Tom) that runs the auto update. Yum it's self is not the problem here. If you feel somehow that is not fesable then please explain why. Now as to my SUGGESTION that updates be put out in a more conscientious fasion, if you find that unworkable then so be it. Other's have no problem structuring their updates in a manner that helps admins. That is not to say you should or can, only you can determine that and from your statements it's not possible for FL to do that. C'est la vie. Been interesting guys. Paul Pettit -- fedora-legacy-list@xxxxxxxxxx http://www.redhat.com/mailman/listinfo/fedora-legacy-list
