At 10:41 PM -0400 6/18/07, Eric wrote: >At 09:55 PM 6/18/2007, Tony Nelson wrote: > ><TN>>>>>In many cases a reboot, or just making sure that nothing is using >RPM and`rm /var/lib/rpm/__db00?` will fix strange RPM issues.<<<<< > >Hmmm ... there are four /var/lib/rpm/__db00x files (000 through 003). > >I said "ps aux | grep rpm" and got back nothing, so can I assume nothing is >using rpm? Yum and yum-updatesd and I expect pirut also use RPM. >Should I just blow away those four __db00x files? Maybe do `lsof | grep __db` first? Note that the files are supposed to exist. ><TN>>>>>Sometimes it helps to rebuild the RPM database, with `rpm >--rebuilddb`.This functionality may be going away; if that happens, you >will need to re-install the OS instead.<<<<< > >Oh, for Pete's sake ... What goes around, comes around, right? I've had to >put up with "re-install Windows" as a cure for what ails ye for years, now >you're telling me that Linux is tumbling down that same path? :-( Anything can break. For a while a kernel bug (along with some other issues) made it happen to a fair number of people, but that's fixed now. A Linux system will still work with a damaged RPM database, but package updates won't work. The RPM database is important enough that a single failure per year in a thousand users is noticed. When RPM seems to stop working, the locks are the most likely issue, but if not, it's possible that the RPM database is damaged. -- ____________________________________________________________________ TonyN.:' <mailto:tonynelson@xxxxxxxxxxxxxxxxx> ' <http://www.georgeanelson.com/>
