Internal DSN Resolution Issue?

[Wendel]

I am having a weird problem with my internal network's name resolution. I
have a name server set up to specifically deal with resolving all of our
internal clients, printers, servers, etc... (I am using RedHat 7.2 and Win
98 for the majority of these machines by the way)

Let just say that on a given machine if there isn't an entry in the host
file that corresponds to machineX and I ping machineX, I get this as output
back from ping.

# ping machineX
PING machineX.mydomain.com (192.168.2.246) from 192.168.2.90 : 56(84) bytes
of data.
Warning: time of day goes back, taking countermeasures.
64 bytes from 192.168.2.246: icmp_seq=0 ttl=128 time=429 usec
64 bytes from 192.168.2.246: icmp_seq=1 ttl=128 time=9.034 sec
64 bytes from 192.168.2.246: icmp_seq=2 ttl=128 time=18.060 sec
64 bytes from 192.168.2.246: icmp_seq=3 ttl=128 time=27.083 sec
64 bytes from 192.168.2.246: icmp_seq=4 ttl=128 time=36.108 sec
64 bytes from 192.168.2.246: icmp_seq=5 ttl=128 time=45.132 sec
64 bytes from 192.168.2.246: icmp_seq=6 ttl=128 time=54.157 sec
64 bytes from 192.168.2.246: icmp_seq=7 ttl=128 time=63.181 sec

(notice the pattern?)

Now when I have an entry that corresponds to machineX in my host file I
receive this output:

# ping machineX
PING machineX (192.168.2.246) from 192.168.2.90 : 56(84) bytes of data.
Warning: time of day goes back, taking countermeasures.
64 bytes from machineX (192.168.2.246): icmp_seq=0 ttl=128 time=474 usec
64 bytes from machineX (192.168.2.246): icmp_seq=1 ttl=128 time=247 usec
64 bytes from machineX (192.168.2.246): icmp_seq=2 ttl=128 time=250 usec
64 bytes from machineX (192.168.2.246): icmp_seq=3 ttl=128 time=237 usec
64 bytes from machineX (192.168.2.246): icmp_seq=4 ttl=128 time=245 usec
64 bytes from machineX (192.168.2.246): icmp_seq=5 ttl=128 time=245 usec
64 bytes from machineX (192.168.2.246): icmp_seq=6 ttl=128 time=253 usec

(notice the beautiful pings?)

My resolv.conf file looks like this:

search mydomain.com
nameserver 192.168.2.254

The weird thing is, is that the first line of output from the ping statement
(when the given machine is w/o the machineX host file entry):

PING machineX.mydomain.com (192.168.2.246) from 192.168.2.90 : 56(84) bytes
of data.

comes back very fast...

Also, when I do something like try to ssh FROM machineX TO this given
machine (and I don't have an entry for machineX in this machines host file),
it takes forever with respect the time it takes when this given machine has
an entry for machineX in it's host file.

I am not certain what to make of this and would love your help in resolving
this problem...

Thanks is advance!

[Wendel]

dar... DNS

jeez... where is the edit button? =D

[Ron Kinner]

First, I would upgrade to the 2.4.9-13 kernel.

see:
http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=37485

That would get rid of your time of day error message.

Your other problem is caused by a bug in ping which is Bug 68212. That interacts with a DNS bug 57998

http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=57998

Ron

[Wendel]

ok... this problem was indeed caused by a problem with reverse zone info...

The person who configured the named.conf file before me had this in the file:

zone "0.2.168.192.in-addr.arpa" {
type master;
file "192.168.2.0.rev";
notify no;
};

This is incorrect! It should read like such:
zone "2.168.192.in-addr.arpa" {
type master;
file "192.168.2.0.rev";
notify no;
};

notice the 0 left off of the zone statement?