Dear Sirs,
I face a very annoying problem with client dns queries.
I have manually configured a nameserver (retec.linux.bogus -
192.168.10.1)
for the domain linux.bogus on RedHat 7.2.
The corresponding critical sections in named.conf, linux.bogus,
192.168.10,
nsswitch.conf, resolv.conf, /etc/hosts, /etc/host.conf files, on the
nameserver are the following:
named.conf:
-----------------
zone "linux.bogus" IN {
type master;
file "pz/linux.bogus";
};
zone "10.168.192.in-addr.arpa" IN {
type master;
file "pz/192.168.10";
};
==============================
linux.bogus:
---------------
$TTL 2D
@ IN SOA retec.linux.bogus. root.retec.linux.bogus. (
2002072701
2D
3H
6W
1D
)
IN NS retec.linux.bogus.
retec IN A 192.168.10.1
retec-2.linux.bogus. IN A 192.168.10.2
www IN CNAME retec
======================================
192.168.10:
-----------------
$TTL 2D
@ IN SOA retec.linux.bogus. root.retec.linux.bogus. (
2002072701
2D
3H
6W
1D
)
IN NS retec.linux.bogus.
1 IN PTR retec.linux.bogus.
2 IN PTR retec-2.linux.bogus.
===========================================
/etc/nsswitch.conf:
------------------------
hosts files dns
=====================================
/etc/hosts:
---------------
127.0.0.1 localhost.localdomain localhost
192.168.10.1 retec.linux.bogus retec
192.168.10.2 retec-2.linux.bogus retec-2
======================================
/etc/resolv.conf:
----------------------
search linux.bogus
nameserver 192.168.10.1
domain linux.bogus
=========================================
/etc/host.conf:
-------------------
order hosts,bind
===========================================
The dns name resolution on the server, works fine: Whether I ask for a
hostname (retec, retec-2, www), FQDN (retec.linux.bogus,
retec-2.linux.bogus, www.linux.bogus) or ip address via nslookup, dig or
if
telnet or ftp using any of the previous "names, IPs", sitting on the
nameserver, the response is correct and IMMEDIATE (no delay
experienced).
The problem arises when I try some of the previous operations on the
client
machine! More specifically, in case I try to telnet or ftp the client
itself
or the nameserver, sitting on the client this time, and use the short
names
(retec-2, retec, www) there is a delay in the response. However this
does
not happen if I use the FQDNs. Moreover, I have no delays if use the
nslookup utility on the client with either any of the short names or
FQDNs.
In that latter case the respone is immediate. The strangest of all is
that
if I do ftp/telnet using the short name alias "www" of the nameserver,
the
response is again immediate.The problem is for sure not on the topology
or
the cabling of the LAN (three machines on a hub.) The client and the
nameserver both have two NICs but the second one is not UP. I tried to
use
the name service on the third machine, retec-3, which is running Windows
NT
4.0 in order to verify that the problem lies on the client and not on
the
linux nameserver, due to the ease of the configuration of the name
service,
and I had the same problem exactly on the client. That is, the problem
MUST
BE on the client resolution.
The files I used on the client are the following:
/etc/hosts:
------------
192.168.10.1 retec.linux.bogus retec
192.168.10.2 retec-2.linux.bogus retec-2
======================================
/etc/nsswitch.conf:
-----------------------
hosts files dns
====================
/etc/resolv.conf:
---------------------
search linux.bogus
nameserver 192.168.10.1
domain linux.bogus
=========================
/etc/host.conf:
------------------
order hosts,bind
==========================
/etc/sysconfig/network:
--------------------------------
NETWORKING=yes
HOSTNAME=retec-2.linux.bogus
===============================
If I erase the contents of /etc/resolv.conf or if remove "dns" from
/etc/nsswitch.conf, then only the /etc/hosts is used and the responses
are
ok.
I have to make clear that the problem is the delay when using short
names on
the client. After this "short" but irritating delay,which appears only
for A
records - not CNAME records, the login prompt does appear. The thing is
that
it is not a normal resolution delay. It lasts more than that.
I have read that the problem could be on the reverse zone, but it looks
to
me fine.
I sould also mention that this problem arose when I made an "empty"
custom
install ( 300 MB ) on the client and configured it. Prior to this, I had
already installed the name server (custom install - full) and a client
(custom - full ) which I had configured as I backup (slave) name server
to
the first one, and as name service client to the primary name server:
/etc/resolv.conf:
--------------------
search linux.bogus
nameserver 192.168.10.1
nameserver 192.168.10.2
domain linux.bogus
======================
At that time everything worked fine. ( After both of them were
configured as
name servers - I had not tried the client resolution before I had
configured
the secondary name server )
However, in each of the cases mentioned in this posting everything is
manually configured.
Please help me with this because it has driven me mad. It is so strange,
to
the length of my knowledge and experience of course.
I am looking forward to reading from you, as it is very urgent. I am on
my
way to RHCE!
Thank you for your time,
George Raptis
_______________________________________________
Redhat-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/redhat-list
