nils toedtmann wrote:
On Wed, Mar 02, 2005 at 06:18:34PM -0800, Matt Simonsen wrote:
I am using a bridge for 2 uml clients. It works beautifully with only one client, then after I startup the second the network randomly cuts out for each system. I've noticed I can "kill" one by pinging the other one.
For example if UML#1 network dies and I ping it, suddenly UML#2 is unpingable (from an outside machine). Then I ping UML#2 from the host, no pings lost it responds, but UML#1 dies.
Sounds as if they use the same MAC. Did you check that?
/nils.
Of course! this exact same thing recently happened to me because i had several UMLs using the same root_fs -- so the ip in the file system for eth0 was the same -- so the 'default' MAC ended up being the same (changing the ip after the fact will not change the MAC).
what you should try, that might not be too clear in the bridge section here:
http://user-mode-linux.sourceforge.net/networking.html
is passing a locally administerd MAC with your eth0=tuntap arg like so:
eth0=tuntap,<tapdevice>,<mac>,<hostip>
where the <mac> looks like 02:de:af:00:ba:d0 for the first device and 02:de:af:00:ba:d1 for the second.
-- db
signature.asc
Description: OpenPGP digital signature
