Lots of Lame Servers

Traynham.Ken at epamail.epa.gov Traynham.Ken at epamail.epa.gov
Thu Feb 8 14:01:29 UTC 2001


We are converting all our BIND 4.9.7 servers to BIND 8.2.3.  Two test
installations went fine, but on the first production installation we are
getting lots of "Lame Server" messages.  I understand the meaning of the
Lame Server message, but am unable to understand why my new BIND 8
secondary is acting so differently than its companion BIND 4 secondary or
the test BIND 8 secondary.  Not only is the new BIND 8 machine logging tons
of Lame Server messages, it returns a non-authoritative response to an
nslookup query for one of the external "lame" domains, but its BIND 4
companion and the BIND 8 test machine return an authoritative response for
the same query.

One possible answer is that the new BIND8 machine was running a caching
nameserver before installation of BIND 8.  (It is not now.)   Could it
still be consulting the cache and thereby returning non-authoritative
answers?  It is over 12 hours since the installation and the Lame Server
messages are still rolling in.  Can this cache be cleaned?  We restarted
named as part of the installation, or course, but have not restarted it
since.  We have not bounced the box and we have not cleaned out the zone
files and resloaded from the primary (because of the time), but that is
next on my list.

Here are the options from the named.conf file

options
{
     directory "/var/dss/namedb";
     query source address * port 53;
     notify no;
     pid-file "named.pid"
};

The new server is running as a non-root user (named) but not in chroot
jail.

The new server and test machine are Red Hat Linux 5.2, BIND 8.2.3.

The companion server is Red Hat Linux 5.2, BIND 4.9.7-1.

Thanks,
Ken Traynham



More information about the bind-users mailing list