On 12/1/05, Jeff Spaleta <jspaleta@xxxxxxxxx> wrote: > On 12/1/05, Panu Matilainen <pmatilai@xxxxxxxxxxxxxxx> wrote: > > It's better than that Actually this leads to all sorts of bad situations depending on the yum command. Because yum doesn't sync its full repodata set on 'typical' operations. 0)start with a bare cache 1)yum check-update > sets the cachecookie and you cache against one of the staler mirrors a little bit later before the cookie timesout 2)yum provides /usr/bin/emacs > this needs filelists.xml.gz but now most if not all the mirrors have synced up the updates so filelists.xml.gz off any mirror fails the checksum against the cached stale repomd. So all you do is keep trying to download this file from each mirror in the mirror lists...making the provides operation take vastly more time than the syncing to the contacted mirror use to take. Can we really rely on this caching without having available functionality suffer, before we do something about mirrors being out of sync? This landed in rawhide too. I don't see how I'm going to be able to continue to do a lot of the troubleshooting I do with yum unless i forcible turn off this cachecookie stuff or force it to update. -jef -- fedora-test-list mailing list fedora-test-list@xxxxxxxxxx To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-test-list