Very Computer

by **Huang Xing Y** » Tue, 17 Jun 1997 04:00:00

The subject says it all...
This is happening on a Ultra 1 (Solaris 2.5.1) with 2 hme cards

My question is - is this a correct behaviour?
I thought each card had a unique address?

I have verified that both interfaces are working correctly
by physically unplugging the network connection one at a time
and pinging to other hosts.

puzzled...

by **Raf LaPiet** » Tue, 17 Jun 1997 04:00:00

Quote:>The subject says it all...
>This is happening on a Ultra 1 (Solaris 2.5.1) with 2 hme cards

>My question is - is this a correct behaviour?
>I thought each card had a unique address?

>I have verified that both interfaces are working correctly
>by physically unplugging the network connection one at a time
>and pinging to other hosts.

>puzzled...

That is standard behavior for sun equiptment. The ether address can be
configured though:

ifconfig hme1 ether aa:1:2:3:4:5

man ifconfig for more info.

by **Casper H.S. Dik - Network Security Engine** » Mon, 30 Jun 1997 04:00:00

Quote:>> ifconfig hme1 ether aa:1:2:3:4:5
>it better you do a
>ifconfig hme1 ether 8:0:20:x:y:z ......

No, it isn't better. 8:0:20 addresses are reserved for Sun.

Addresses with the 2nd bit set in the first octet are guaranteed
not to conmflict with vendor issued addresses.

(bit 2 is 02:; bits are put on the net lsb first)

Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.

by **0000-Admin(00** » Mon, 30 Jun 1997 04:00:00

>>> ifconfig hme1 ether aa:1:2:3:4:5

>>it's better you do a
>>ifconfig hme1 ether 8:0:20:x:y:z ......

>No, it isn't better. 8:0:20 addresses are reserved for Sun.

it is a sun - so you have it easier working with a sniffer

Quote:>Addresses with the 2nd bit set in the first octet are guaranteed
>not to conmflict with vendor issued addresses.

o.k. - but the address wasn't uniq befor, as all interfaces
of a sun have originally the same mac.

by **Casper H.S. Dik - Network Security Engine** » Tue, 01 Jul 1997 04:00:00

Quote:>it is a sun - so you have it easier working with a sniffer

Yes, but teh chance of conflicts becomes much bigger.

Quote:>>Addresses with the 2nd bit set in the first octet are guaranteed
>>not to conmflict with vendor issued addresses.
>o.k. - but the address wasn't uniq befor, as all interfaces
>of a sun have originally the same mac.

Which is normally OK, if you pick a Sun reserved address as new ethernet
address for a second ethernet port, you may find yourself in trouble when
a new Sun arrives that has the ethernet # you picked (ok, that's only
a 1:2^24 chance, but still. I prefer encoding the IP address in the last
four octets and using a local 2 octet prefix with 02 set in the first byte)

Casper
--
Expressed in this posting are my opinions. They are in no way related
to opinions held by my employer, Sun Microsystems.
Statements on Sun products included here are not gospel and may
be fiction rather than truth.

by **Tony Walto** » Tue, 01 Jul 1997 04:00:00

> >Addresses with the 2nd bit set in the first octet are guaranteed
> >not to conmflict with vendor issued addresses.

> o.k. - but the address wasn't uniq befor, as all interfaces
> of a sun have originally the same mac.

IMO Casper is right - the address was unique. The ethernet MAC address
is a *station* address, not a *port* address - as such the requirement
is for each *station* (eg workstation) to have a unique address - this
requirement is satisfied even if all Ethernet cards or, in fact, ports
on the same workstation have the same MAC address.

Regards

--
Tony

The posting above should not be taken as support from Sun Microsystems.
My opinions may differ from those of Sun, and may be dead wrong.

by **mayer han** » Thu, 03 Jul 1997 04:00:00

>> >Addresses with the 2nd bit set in the first octet are guaranteed
>> >not to conmflict with vendor issued addresses.

>> o.k. - but the address wasn't uniq befor, as all interfaces
>> of a sun have originally the same mac.

>IMO Casper is right - the address was unique. The ethernet MAC address
>is a *station* address, not a *port* address - as such the requirement
>is for each *station* (eg workstation) to have a unique address - this
>requirement is satisfied even if all Ethernet cards or, in fact, ports
>on the same workstation have the same MAC address.

i can't agree. each - lets say device - and if a station has 2
interfaces, then there are 2 such devices, must be uniq. then
you get a uniq entry in the arp cache for the relationship
mac and ip-address.
on the other way an other station could address with the same
layer 2 two different layer 3 addresses.

by **Elias Martenso** » Fri, 04 Jul 1997 04:00:00

> i can't agree. each - lets say device - and if a station has 2
> interfaces, then there are 2 such devices, must be uniq. then
> you get a uniq entry in the arp cache for the relationship
> mac and ip-address.
> on the other way an other station could address with the same
> layer 2 two different layer 3 addresses.

No, because the different interfaces are never on the same segment. The
arp cache not only knows what IP address the MAC has, but also on what
interface it is.

Regards / Elias

Very Computer

Q: "ifconfig -a" shows hme0 and hme1 with same ethernet addr?

Q: "ifconfig -a" shows hme0 and hme1 with same ethernet addr?

Q: "ifconfig -a" shows hme0 and hme1 with same ethernet addr?

Q: "ifconfig -a" shows hme0 and hme1 with same ethernet addr?

Q: "ifconfig -a" shows hme0 and hme1 with same ethernet addr?

Q: "ifconfig -a" shows hme0 and hme1 with same ethernet addr?

Q: "ifconfig -a" shows hme0 and hme1 with same ethernet addr?

Q: "ifconfig -a" shows hme0 and hme1 with same ethernet addr?

Q: "ifconfig -a" shows hme0 and hme1 with same ethernet addr?