> Every run of yum? I dont see that happening. My understanding that > the logic behind the "MD Read" step was added specifically to make > repeated runs of yum faster, if a repository has not changed. If a > repository has changed since the last run of yum the new xml metadata > has to be parsed, there is no way around that. > So if you run yum only once a day against the development tree... you > won't see a benefit here because the development tree changes every > day. If you run yum in a cronjob in a clever way, so all its doing is > pulling metadata somewhat perioudicly and building the cached binary > pickle object(s), any subsequent run of yum will use the cached pickle > if it can to save you time. > > Then again... using the newest version of yum im not seeing the "MD > Read:" step anylonger so i have no idea if my understanding is now > out-dated again. No, you're metadata is just cached. if you want to see it again run: yum clean cache and then it will reimport it all again. OR if you want generate all the caches of the metadata run: yum makecache <go get coffee> -sv