Hi,
I seem to have some weird DNS problem where my DNS server will not or is
unable to respond to some requests (Latest 7.1 updates)
The problem appears to be the actual name being requested since most
names always work (but I'm not sure how others work when this one doesn't)
At the bottom are 2 tcpdumps based on the ether address of the machine
requesting the DNS resolution (I used ether in case there was something
else happening not IP based that I couldn't see or the address was being
changed by something)
The first example with a.b.c.15 just shows the problem
The second example with a.b.c.9 show www.google.com working (I don't even
understand this one since there appears to be no reply) and then the same
failure as the first example
If the request works, I only see the same as the first line of the dump,
but when it fails (always on whirlpool.net.au) I see all the rest of the
dump shown.
I also have a Netgear 5 port switch which filters out a lot of the
network traffic (I'll consider bypassing that next in case it has something
to do with it - but I doubt it is the cause)

a.b.c.1 is the DNS server
a.b.c.9 is the Win98SE machine
a.b.c.9 has ether address 00:a0:cc:52:c6:7a
a.b.c.15 is the Linux machine
a.b.c.15 has the ether address 00:e0:18:1c:d2:52

nslookup works so I guess the problem could be my DNS setup ... but I
don't understand the output I am getting ...
the .15 computer is also running a cacheing nameserver (but the problem
also occurs on the Win98SE machine not running any DNS services)
The .15 and the Win98SE computers get masqueraded on the way out to the
internet but I can't see that being related to the problem since all the
computers are on the same local network (with just switches and hubs
between them)

Obviously the last part of each dump is the problem but I don't understand
how that message could be generated.
i.e. how does it know the UDP port is unreachable?
(it may well be correct but I don't know how it knows since this is should
be a full dump of all the data sent over the wire)
The problem is not firewall related

I've obviously misunderstood something, but I'm at a loss.

I'll try reinstalling the DNS and the .15 computer on the weekend with
7.2 just to be sure it isn't a 7.1 problem, but until then ...

============================================================================

# tcpdump -l -n ether host 00:e0:18:1c:d2:52
User level filter, protocol ALL, TURBO mode (575 frames), datagram packet
sockettcpdump: listening on all devices
12:17:18.090000 eth0 < a.b.c.15.32772 > a.b.c.1.domain: 20682+ A?
whirlpool.net.au. (34) (DF)
12:17:23.080000 eth0 < arp who-has a.b.c.1 tell a.b.c.15
12:17:23.090000 eth0 < a.b.c.15.32772 > a.b.c.1.domain: 20682+ A?
whirlpool.net.au. (34) (DF)
12:17:28.100000 eth0 < a.b.c.15.32772 > a.b.c.1.domain: 20683+ A?
whirlpool.net.au.k1k2.com. (43) (DF)
12:17:28.100000 eth0 < a.b.c.15.32772 > a.b.c.1.domain: 20684+ A?
whirlpool.net.au. (34) (DF)
12:17:33.100000 eth0 < arp reply a.b.c.15 is-at 0:e0:18:1c:d2:52
(0:0:e8:de:a3:6f)
12:17:33.110000 eth0 < a.b.c.15.32772 > a.b.c.1.domain: 20684+ A?
whirlpool.net.au. (34) (DF)
12:17:38.120000 eth0 < a.b.c.15.32772 > a.b.c.1.domain: 20685+ A?
whirlpool.net.au.k1k2.com. (43) (DF)
12:17:47.080000 eth0 < a.b.c.15 > a.b.c.1: icmp: a.b.c.15 udp port 32772
unreachable [tos 0xc0]
12:17:47.080000 eth0 < a.b.c.15 > a.b.c.1: icmp: a.b.c.15 udp
port 32772 unreachable [tos 0xc0]
12:17:47.080000 eth0 < a.b.c.15 > a.b.c.1: icmp: a.b.c.15 udp
port 32772 unreachable [tos 0xc0]
12:17:47.080000 eth0 < a.b.c.15 > a.b.c.1: icmp: a.b.c.15 udp
port 32772 unreachable [tos 0xc0]
12:17:47.080000 eth0 < a.b.c.15 > a.b.c.1: icmp: a.b.c.15 udp
port 32772 unreachable [tos 0xc0]
12:17:47.080000 eth0 < a.b.c.15 > a.b.c.1: icmp: a.b.c.15 udp
port 32772 unreachable [tos 0xc0]

133 packets received by filter

============================================================================


# tcpdump -l -n ether host 00:a0:cc:52:c6:7a
User level filter, protocol ALL, TURBO mode (575 frames), datagram packet
sockettcpdump: listening on all devices
17:06:55.710000 eth0 < a.b.c.9.1032 > a.b.c.1.domain: 2+ A? www.google.com.
(32)

---------------------------------------

17:08:23.770000 eth0 < a.b.c.9.1035 > a.b.c.1.domain: 3+ A?
whirlpool.net.au. (34)
17:08:25.280000 eth0 < a.b.c.9.1035 > a.b.c.1.domain: 3+ A?
whirlpool.net.au. (34)
17:08:27.780000 eth0 < a.b.c.9.1035 > a.b.c.1.domain: 3+ A?
whirlpool.net.au. (34)
17:08:32.270000 eth0 < a.b.c.9.1035 > a.b.c.1.domain: 3+ A?
whirlpool.net.au. (34)
17:08:40.770000 eth0 < a.b.c.9.1036 > a.b.c.1.domain: 4+ A?
whirlpool.net.au.k1k2.com. (43)
17:08:45.780000 eth0 < arp reply a.b.c.9 is-at 0:a0:cc:52:c6:7a
(0:0:e8:de:a3:6f)
17:09:53.800000 eth0 < a.b.c.9 > a.b.c.1: icmp: a.b.c.9 udp port 1035
unreachable
17:09:53.800000 eth0 < a.b.c.9 > a.b.c.1: icmp: a.b.c.9 udp port 1035
unreachable
17:09:53.810000 eth0 < a.b.c.9 > a.b.c.1: icmp: a.b.c.9 udp port 1035
unreachable
17:09:53.820000 eth0 < a.b.c.9 > a.b.c.1: icmp: a.b.c.9 udp port 1035
unreachable
17:09:58.800000 eth0 < arp reply a.b.c.9 is-at 0:a0:cc:52:c6:7a
(0:0:e8:de:a3:6f)
17:16:01.390000 eth0 B arp who-has a.b.c.15 tell a.b.c.9
 
668 packets received by filter

============================================================================

-- 
-Thanks for any help
-Andrew

MS ... if only he hadn't been hang gliding!



_______________________________________________
Seawolf-list mailing list
[EMAIL PROTECTED]
https://listman.redhat.com/mailman/listinfo/seawolf-list

Reply via email to