[geeks] Solaris routing problem ....

[Michael C. Vergallen]

Jonathan C. Patschke wrote:
> On Wed, 12 Mar 2008, Michael C. Vergallen wrote:
>
>> how comes I can connect from all the workstations to 192.168.0.4 witch
>> is a zone's ip but when i try to connect from 192.168.0.2 with is the
>> global zone on the server the thing doesn4t work at all however  it
>> does work when the zone is on the same interface as  the global zone
>> ... what am i doing wrong here ?
>
> My WAG is that the server is replying with the MAC address of the second
> port, since the kernel routing table has two entries for 192.168.0.0/24,
> and the 192.168.0.4 interface got priority somewhy (either it was
> brought up first or last--I forget which way Solaris resolves).  This
> would confuse the networking stack on the other end, since it doesn't
> match what the workstation asked for with ARP; its response is to drop
> the frame.
>
> A good check for this would be to run netstat on the server and see if
> a connection attempt results in a connection that doesn't fully
> initialize on the server end.
>
> If you want both interfaces on the same broadcast domain, consider
> trunking them and binding all the IP addresses for that node to the
> trunk.  You'll tend to get more effective use of your available
> bandwidth this way.
>

Dammit I finally got the stuff working properly .... I simply forgot
that I should have assigned the IP numbers for the physical interfaces
to be within different subnets
and now it works with the zones being in the same subnet as the server.
e.i. physical interface qfe0 had to be in the 192.168.1  subnet and then
it works without a problem
 to have zone1 use qfe0:1 with the 192.168.0.3 IP address.
Thanks for the info trough  ...it helped me a bit to figure out what I
did wrong.

[demime 1.01d removed an attachment of type application/pgp-signature which had a name of signature.asc]