Re: switched ethernet interfaces

Normally, when you set up an ethernet interface, you've got two numbers,
an address and a netmask.   By combining these, you're able to specify a
"network address" that identifies the subnet.

For example, with an address of 172.16.1.10 and a netmask of
255.255.255.0, the subnet is 172.16.1 and the host is 10.   That's because
the 255.255.255.0 is a "bit mask" (each number in the address is an 8-bit
byte, so 255 is x'FF' or all bits on in binary, and 0 is all bits off...
consequently, 255.255.255.0 shows all bits on in the first 3 numbers, and
off in the last... that's why the first 3 numbers of the address are the
subnet, and the last is not, because of this mask)  Hopefully that makes
sense.

If you have two ethernet adapters on the same subnet, with the same
netmask, which one does the system use?   I have no idea.   Which
interface it uses is "undefined" because you're not supposed to do that :)

Maybe the netmask for 172.16.1.13 should be set to *HOST?  In that case,
it shouldn't be used for outbound connections, just for incoming requests
aimed at that address.

On Tue, 3 Jun 2003, Jerome Draper wrote:
>
> We have three ethernet adapters in this AS400 all running TCP/IP.  The default
> gateway (172.16.1.1) is in the same subnet as two of them (172.16.1.10 and
> 172.16.1.13).  The first one, 172.16.1.10, up until now, had been used for
> outbound network connections.  Now or all of the sudden the AS400 has started
> using the other adapter 172.16.1.13.  Why?  How could this be?
>
> The fully qualified domain name is (now?) tied to the 172.16.1.13 adapter.
>