On Fri, 21 Nov 2003, Tom Diehl wrote: > Would someone please explain why it would be a bad thing(tm) to do > yum-update --with-obsoletes all of the time?? Side effects. In many cases obsoletes require reconfiguration of utilities that might be core/critical in your organization. It would really suck to go to bed one night and discover the next morning that every system in a cluster or on a campus was "broken" because e.g. a core library or program was replaced by one that used a different API with a different configuration. It is the same thing that always makes upgrading a dicey business. When a revision number changes, it is a signal that there exist SOME packages that are obsoleted or deprecated relative to the existing stable revision, and that a full installation is recommended to auto and/or manually configure the replacements. For years reinstallation was the only sane way to upgrade. Yum is rather remarkable in that it CAN upgrade a running system and sometimes/often not break anything, but that really does depend on what changes between revisions that can be broken. Even within a revision (when almost all updates won't/shouldn't obsolete anything without somehow managing the configuration port) well, this depends on whether EVERYBODY has been playing nice in RPM space, doesn't it...;-) For most people running off of most repositories, this isn't something that you want to have happening automatically and nightly. If you manage the repository and ensure that all obsoleting updates work fully automagically and robustly, w'hell, go for it:-) Yum remains very conservative in its basic design, according to a "do no harm" principle. Installs and updates are pretty safe. Upgrades are less safe. Removes are downright unsafe (in the hands of an inexperienced user who doesn't realize that removing certain things effectively forces removal of half the software on their system). So you have to do some extra work to select an update with obsoletes, just enough to form a barrier to those unready to cope with what can happen. rgb -- Robert G. Brown http://www.phy.duke.edu/~rgb/ Duke University Dept. of Physics, Box 90305 Durham, N.C. 27708-0305 Phone: 1-919-660-2567 Fax: 919-660-2525 email:rgb@xxxxxxxxxxxx
