What is a "lame" server ??

[Barry Margolin]

I have a Bind 8.x master server holding master records.
I have a slave server running bind 9.x
When I start the bind 9 server It all looks ok but then comaplins of a lame
server (see log) what does this mean ?
berger is the bind 8 server alesi is the bind 9

A lame server is a server that's expected to be authoritative for a zone
(because the domain is delegated to it) but isn't answering
authoritatively when queried for names in that zone.

It looks to me like you don't have alesi configured as a slave for the
projecthugo.co.uk zone.

Dec 26 23:27:42 jordan named[10337]: transfer of 'bathgirls.co.uk' from
217.34.1
94.217#53: end of transfer
Dec 26 23:27:42 jordan named[10337]: transfer of 'bathjobs.com' from
217.34.194.
217#53: end of transfer
Dec 26 23:27:42 jordan named[10337]: transfer of 'matthewdarcy.com' from
217.34.
194.217#53: end of transfer
Dec 26 23:27:48 jordan named[10337]: lame server on
'alesi.projecthugo.co.uk' (i
n 'projecthugo.co.uk'?): 217.34.194.218#53
Dec 26 23:27:48 jordan named[10337]: lame server on
'berger.projecthugo.co.uk' (
in 'projecthugo.co.uk'?): 217.34.194.218#53

[Simon Waters]

Dec 26 23:27:48 jordan named[10337]: lame server on
'alesi.projecthugo.co.uk' (i
n 'projecthugo.co.uk'?): 217.34.194.218#53
Dec 26 23:27:48 jordan named[10337]: lame server on
'berger.projecthugo.co.uk' (
in 'projecthugo.co.uk'?): 217.34.194.218#53

A lame server is one that doesn't answer correctly for a zone it
is expected to.

Typically it is delegated from above, but does not set the
authoritive flag on answers (AA in the flags section of dig's
output). BIND 9 also complains that servers are lame under
slightly different conditions, but the basic principal is
something bad happen such that I didn't get a usable answer from
a server that should have been able to answer the query.

My guess is that you have some suspect data in the zone file
that isn't acceptable to alesi. Check the logs from the last
time BIND restarted on "alesi", and the last zone transfer logs
on "berger". Failing that post the current contents of the zone
file from alesi.

[Simon Waters]

Post by Barry Margolin
It looks to me like you don't have alesi configured as a slave for the
projecthugo.co.uk zone.

And the other answer - by some sleepy chap called Simon, got
your servers around the wrong way, and so didn't realise Alesi
is slave and not master. Doh.

[Simon Waters]

can you expalin the difference in configuring a slave and a master, I
thought this was done using the zone configs ?

It is configured in the config file. Rather than "type master"
you set "type slave".

If you haven't set it up correctly, then you get the current
behaviour where the slave answers queries okay, but only because
it is asking the master as needed like any other DNS server, not
because it has made a local copy of the data.

Thus a DNS zone with a lame server will be less robust than one
without to failure of DNS servers.

If you've tried to set it up and made a mistake you should be
getting error messages, read the logs on both machines.

[Barry Margolin]

that is exactly how the config is,
berger has type master
alesi has type slave
with masters [ $IP address ]
I'll post the config files so you can see what I mean, I am %90 sure I have
config this correct, I think it maybe something to do with a bind9 server
taking zones from a bind 8 server. I have configured master and slave before
on to bind 8 servers on solaris without problem

The domain that it's complaining about is "projecthugo.co.uk". Where is
that in the except below?

config files for alesi (slave)
};
zone "bathjobs.com" {
type slave;
file "named.bathjobs.com";
masters { 217.34.194.217; };
};
zone "matthewdarcy.com" {
type slave;
file "named.matthewdarcy.com";
masters { 217.34.194.217; };
};
zone "matthewdarcy.co.uk" {
type slave;
file "named.matthewdarcy.co.uk";
masters { 217.34.194.216; };
};