2002-Dec-06

You can get a copy of named's current database and cache by doing a "ndc dumpdb". Under Debian, this defaults to save to a plain-text file in zone file format at /var/cache/bind/named_dump.db. Usually it is at /var/tmp/named_dump.db. (You can also get it by sending SIGINT to the named process.)

rpm was hanging under Red Hat 8.0. Strace showed it just continuing with:

select(0, NULL, NULL, NULL, {1, 0})     = 0 (Timeout)

A posting said to delete /var/lib/rpm/__db* (but I moved instead) and run "rpm --rebuilddb". Also this may be bug 73097 on www.rpm.org where the version 4.0.4 needs to be downgraded. More info at http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=74726 and https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=68056 and http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=73097

Once I had killed all rpm processes and moved thosse __db* lock files and rebuilt the rpm database, then it worked.

This system had over 1600 packages installed. I slowly removed kde et cetera. By trying to removed, saving error output, cutting out the list of RPMs that they were need by and removing them too. Started with:

grep kde rpm.list | xargs rpm -ev

Later did:

cut -d")" -f 2 J J2 | fgrep -- -  | xargs rpm -ev