Hmm...I can't guarantee this wouldn't have undesirable effects but deleting 
the transient /var/lib/rpm/__db* files may do the trick. 

When my cron.daily RPM job occasionally hung, deleting these transient files 
was the only way I could access the RPM database without rebooting.

Although this might not be the exact problem, this URL might explain a bit 
more:
http://www.geocrawler.com/archives/3/87/2001/8/50/6439387/

Good luck!

Brandon

On Friday 30 November 2001 12:08 pm, you wrote:
> I was removing some RPMs using rpm -e [package] and the process
> appeared to hang.  After many minutes of waiting, I attempted to kill it,
> with no success.  And I cannot do any other rpm commands until it's
> done/killed. What can I do?
>
> [root at walleye atanner]# ps -ef | grep rpm
> root      5863  5836  0 11:47 ?        00:00:00 rpm -e apmd autofs
> xisdnload isdroot      5971  5906  0 12:03 ttyp2    00:00:00 grep rpm
> [root at walleye atanner]# kill -9 5863
> [root at walleye atanner]# ps -ef | grep rpm
> root      5863  5836  0 11:47 ?        00:00:00 rpm -e apmd autofs
> xisdnload isdroot      5974  5906  0 12:08 ttyp2    00:00:00 grep rpm
> [root at walleye atanner]# rpm -q apmd
> error: cannot get shared lock on /var/lib/rpm/Packages
> error: cannot open Packages index using db3 - Operation not permitted (1)
>
> thanks.