I installed my RH6.0 as a Router with two ether cards(one is 100M
3c509B, another is 10M D-Link 220), I have my internal segment with all
win98 machines(192.168.0.0, subnet mask 255.255.255.0) connected to
RH6.0 eth1(192.168.0.1), and set all the internal machines default
gateway to 192.168.0.1. My Linux eth0 have a legal ip address
202.109.66.69(subnet mask 255.255.255.224),
and it connected to a Cisco2600 and through it to the INTERNET. Here is
the ASCII graphic:

            |-----|                                       |-----|
INTERNET----|     |---------------------------------------|
|----------192.168.0.100...200
            |-----|<--202.109.66.65
202.109.66.69--->|-----|<----192.16.0.1
           Cisco2600                                       RH6.0
The RH6.0 eth0 and Cisco2600 belong to the same segment(202.109.66.64,
subnet mask 255.255.255.
224). I set up WWW, DNS, E-Mail server on the Linux box to server the
internal 192 segment and
202.109.66.64 segment.

Here goes the problem: when the Cisco2600 was down(not connected to the
out world), then
at the linux box I use the arp command, and the ARP table show up very
very slowly, and in the
192 segment, I found sending and receiving mail was very very slowly
though it still worked,
but WWW was good. When these happened, on the 202.109.66.64 segment all
was good include WWW and
E-Mail.

When 2600 was up(could connect to the out world), the problem segment
came back to normal.
Someone said it was the DNS problem, but I don't think so, cause on the
192 segment ONLY mail
was so slowly, but WWW was OK, and it also used DNS. Another reason is
the arp command does not
use DNS, is that right?

I really wonder about it, Could anyone give me an explanation, what was
wrong?


-- 
To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe"
as the Subject.

Reply via email to