SEO

I have a 4-port 501-2062 SQEC/S card that I plugged into an Ultra-1
200E running Solaris 8 10/01 patched to Generic_117350-04. Did a
reconfiguration reboot and did a manual ifconfig for qe0 with plumb
up, and tried to ping the connection. The ping responds through the
hme0 port, not the qe0 port. I checked cabling, etc.---if I
disconnect the hme0 cable, things hang and there appears to be no
activity on my hub lights on the qe0 cable, even reversing the cables
at the Ultra.

Checked to see if I have a driver loaded, and if I could see anything
funny anywhere. What I get are:

# ifconfig -a
lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index
1
inet 127.0.0.1 netmask ff000000
hme0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500
index 2
inet 192.168.1.1 netmask ffffff00 broadcast 192.168.1.255
ether 8:0:20:a0:aa:24
qe0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index
3
inet 192.168.1.9 netmask ffffff00 broadcast 192.168.1.255
ether 8:0:20:a0:aa:24
# modinfo |grep ether
# modinfo |grep Ether
71 102d6783 f18a 7 1 hme (10/100Mb Ethernet Driver v1.149)
104 102abfe2 6512 104 1 qe (Quad-MACE Ethernet Driver v1.80)
# netstat -nr

Routing Table: IPv4
Destination Gateway Flags Ref Use Interface
-------------------- -------------------- ----- ----- ------ ---------
192.168.1.0 192.168.1.1 U 1 190 hme0
192.168.1.0 192.168.1.9 U 1 0 qe0
224.0.0.0 192.168.1.1 U 1 0 hme0
127.0.0.1 127.0.0.1 UH 2 28 lo0
#

What do I need to do to get this thing to talk out of the qu ports?

Hank

On Thu, 22 Jul 2004 20:31:24 -0600, Henry van Cleef wrote:

> hme0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500
> index 2
> inet 192.168.1.1 netmask ffffff00 broadcast 192.168.1.255
> ether 8:0:20:a0:aa:24
> qe0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index
> 3
> inet 192.168.1.9 netmask ffffff00 broadcast 192.168.1.255
> ether 8:0:20:a0:aa:24

You cannot have 2 ethernet interfaces on a subnet with the same ethernet
address.

Henry van Cleef wrote:
> I have a 4-port 501-2062 SQEC/S card that I plugged into an Ultra-1
> 200E running Solaris 8 10/01 patched to Generic_117350-04. Did a
> reconfiguration reboot and did a manual ifconfig for qe0 with plumb
> up, and tried to ping the connection. The ping responds through the
> hme0 port, not the qe0 port. I checked cabling, etc.---if I
> disconnect the hme0 cable, things hang and there appears to be no
> activity on my hub lights on the qe0 cable, even reversing the cables
> at the Ultra.
>
> Checked to see if I have a driver loaded, and if I could see anything
> funny anywhere. What I get are:
>
> # ifconfig -a
> lo0: flags=1000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4> mtu 8232 index
> 1
> inet 127.0.0.1 netmask ff000000
> hme0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500
> index 2
> inet 192.168.1.1 netmask ffffff00 broadcast 192.168.1.255
> ether 8:0:20:a0:aa:24
> qe0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index
> 3
> inet 192.168.1.9 netmask ffffff00 broadcast 192.168.1.255
> ether 8:0:20:a0:aa:24
> # modinfo |grep ether
> # modinfo |grep Ether
> 71 102d6783 f18a 7 1 hme (10/100Mb Ethernet Driver v1.149)
> 104 102abfe2 6512 104 1 qe (Quad-MACE Ethernet Driver v1.80)
> # netstat -nr
>
> Routing Table: IPv4
> Destination Gateway Flags Ref Use Interface
> -------------------- -------------------- ----- ----- ------ ---------
> 192.168.1.0 192.168.1.1 U 1 190 hme0
> 192.168.1.0 192.168.1.9 U 1 0 qe0
> 224.0.0.0 192.168.1.1 U 1 0 hme0
> 127.0.0.1 127.0.0.1 UH 2 28 lo0
> #
>
> What do I need to do to get this thing to talk out of the qu ports?
>
> Hank

Dave pointed out that "you cannot have 2 ethernet interfaces on a subnet
with the same ethernet address". That's not strictly true, but it's not
very useful. As you have found out, output data will only go out over
one ethernet but the system will be able to receive data on either.
If the output ethernet goes down, the system will not switch over to
the other one.

Failover is possible, if that's what you are trying to do, but
this isn't the way to do it. There was a very recient thread on
this. It's called IP Multipathing and here are a link on it:

http://www.eng.auburn.edu/~doug/howt...tipathing.html

As for getting the "ge" interface talking, you should have two file
in "/etc" named "hostname.hme0" and "hostname.ge0". These should
contain host names in them. I assume that they both contain the
same name. Temporarily, change the name of "hostname.hme0" to
"hostname.hme0.ori" and reboot. This will then bring up only "ge0"
as your network interface.

If you want to bring up the other interface, add a new entry to the
host file and use a different name and a different subnet.

> cat /etc/hosts
127.0.0.1 localhost
192.168.1.1 mainhost loghost
192.168.10.1 sub_host

> cat /etc/hostname.ge0
mainhost

> cat /etc/hostname.hme0
sub_host

With this setup, a reboot should bring up both interfaces with the
"ge" ingerface being your main connection. I did a similar setup
just this week with a "eri" and a "frcgei" ethernet interface. The
"frcgei" interface didn't respond until they were defined as
hosts on separate sub-nets.

I'm sure that other will give further input on possible router
issues and such.

--
Martin E. Meserve
k7mem@myrealbox.com
http://www.k7mem.150m.com

On Fri, 23 Jul 2004 06:54:13 -0700, K7MEM wrote:

> Dave pointed out that "you cannot have 2 ethernet interfaces on a subnet
> with the same ethernet address". That's not strictly true, but it's not
> very useful.

It is true on a switched network. And I left all the rest unsaid because
I gave the OP a chance to figure out for himself how to resolve the
problem.

In article <pan.2004.07.23.16.13.59.904592@yahoo.com>,
Dave Uhring <daveuhring@yahoo.com> wrote:
>On Fri, 23 Jul 2004 06:54:13 -0700, K7MEM wrote:
>
>> Dave pointed out that "you cannot have 2 ethernet interfaces on a subnet
>> with the same ethernet address". That's not strictly true, but it's not
>> very useful.
>
>It is true on a switched network. And I left all the rest unsaid because
>I gave the OP a chance to figure out for himself how to resolve the
>problem.
>
Thanks for your answer, Dave. Basically, I want to configure the
multiport box as a router, but didn't have a third box up and running
to look like a different subnet. And it turns out that taking an
unused box off my stack of U1E's isn't plug'n play---shuffled too many
hard disks around before making a 400 mile move, so I've got some
reinstalls to do.

Sun's docs seem clear that a box is a router if it's got two IP's
configured and no /etc/notrouter file. If that's not the case, then I
need more calibration.

Hank

On Fri, 23 Jul 2004 22:12:29 -0600, Henry van Cleef wrote:

> Thanks for your answer, Dave. Basically, I want to configure the
> multiport box as a router, but didn't have a third box up and running
> to look like a different subnet. And it turns out that taking an
> unused box off my stack of U1E's isn't plug'n play---shuffled too many
> hard disks around before making a 400 mile move, so I've got some
> reinstalls to do.

If you have a switched network and place two NICs on that network with the
same ethernet address and different IP addresses you confuse the switch.

You can change the ethernet addresses of your qe0 by using ifconfig ether
or use setenv in OBP but I don't remember the exact parameter.

> Sun's docs seem clear that a box is a router if it's got two IP's
> configured and no /etc/notrouter file. If that's not the case, then I
> need more calibration.

Yes that is true, but you were trying to route to the same subnet.
Configure qe0 to 192.168.0.9/24, connect it to the 192.168.0.0/24 network
and you should not have the reported problems.

<snip>
> You can change the ethernet addresses of your qe0 by using ifconfig ether
> or use setenv in OBP but I don't remember the exact parameter.
</snip>

Either from Solaris;

$ eeprom local-mac-address?=true

or OBP;

ok setenv local-mac-address? true

AT

In article <pan.2004.07.24.05.43.48.451225@yahoo.com>,
Dave Uhring <daveuhring@yahoo.com> wrote:
>On Fri, 23 Jul 2004 22:12:29 -0600, Henry van Cleef wrote:
>
>If you have a switched network and place two NICs on that network with the
>same ethernet address and different IP addresses you confuse the switch.
>
>You can change the ethernet addresses of your qe0 by using ifconfig ether
>or use setenv in OBP but I don't remember the exact parameter.
>
Once again, thanks for flagging the fact that the qe's get configured
with the box ethernet address. I can think of about eight ways to
make that problem go away, but solving it is on the back burner, as I
am routing through two different hubs.

>> Sun's docs seem clear that a box is a router if it's got two IP's
>> configured and no /etc/notrouter file. If that's not the case, then I
>> need more calibration.
>
>Yes that is true, but you were trying to route to the same subnet.
>Configure qe0 to 192.168.0.9/24, connect it to the 192.168.0.0/24 network
>and you should not have the reported problems.
>
Understood. I've gotten the box with the SQEC card up and running as
a router, and have tested it. As I mentioned, my objective was to get
a router between a 100BaseT hub, and a 10BaseT hub, and now have this
up and working.

I have a hunch that putting in an /etc/ethers file may work with the
startup scripts---if I can't find a "Sun way," I can always add a
script to change the ethernet addresses with ifconfig.

Hank

On Sun, 25 Jul 2004 11:27:04 -0600, Henry van Cleef wrote:

> Understood. I've gotten the box with the SQEC card up and running as
> a router, and have tested it. As I mentioned, my objective was to get
> a router between a 100BaseT hub, and a 10BaseT hub, and now have this
> up and working.

Glad to hear it.

> I have a hunch that putting in an /etc/ethers file may work with the
> startup scripts---if I can't find a "Sun way," I can always add a
> script to change the ethernet addresses with ifconfig.

See the article posted by Andrew Tyson in this thread. He provided the
needed parameters for modifying the OBP settings, which should solve the
identical MACADDR problem.

ethers(4) does not seem to deal with the problem of the same ethernet
addresses on 2 or more ethernet interfaces. Best to modify the OBP
local-mac-address? setting or as a second choice to use ifconfig qe0
ether.

"Andrew Tyson" <asptysonATyahooD0TcomD0Tau> writes:
><snip>
>> You can change the ethernet addresses of your qe0 by using ifconfig ether
>> or use setenv in OBP but I don't remember the exact parameter.
></snip>
>
>Either from Solaris;
>
>$ eeprom local-mac-address?=true
>

It's wisest to prevent the question mark character
from being interpreted as a shell globbing wildcard:

$ eeprom local-mac-address\?=true
or
$ eeprom 'local-mac-address?=true'


-Greg
--
Do NOT reply via e-mail.
Reply in the newsgroup.