On Wed, Dec 24, 2008 at 11:48 AM, Akemi Yagi <amyagi at gmail.com> wrote: > On Wed, Dec 24, 2008 at 11:42 AM, Les Mikesell <lesmikesell at gmail.com> wrote: >> Somewhere in the process of doing a yum update from a terminal window >> open in an NX/freenx desktop session, the whole session blew up - >> perhaps when ssh restarted or when freenx was updated. The yum update >> did not complete beyond that point and the session was gone when I >> reconnected. Is that supposed to happen? (Centos 5.x) > > Yes, it does happen. For more details, please see (warning, very long thread): > > http://www.centos.org/modules/newbb/viewtopic.php?viewmode=flat&topic_id=17166&forum=37 > > If you now have a broken rpm database because of the interrupted yum > session, the thread has a remedy for that. OK, that long thread caused a headache for someone (can't tell you who), so I'm going to give you a shortcut. Run the command: rpm -qa --qf "%{NAME} %{ARCH}\n" | sort | uniq -c | awk '$1 != 1 { print $2 }' | sort If the output shows anything other than gpg-pubkey or kernel*, then your rpm database is most likely broken. Go to #45 of that thread. The OP nicely summarized the procedure to fix it. Akemi / toracat