[Netkit.users] Another strange behavior: bug?

[Giovambattista Ianni]

Dear all,

I found some misbehavior when launching a VM with an alternative terminal, like
vstart fw --eth0=A --xterm=gnome

What happens is that the started machine cannot activate the eth0 link on the hub A
(excerpt from dmesg below)
----
Computing a digest to use as ethernet address for device eth0
Netdevice 0 (72:e8:99:2b:97:37) : daemon backend (uml_switch version 3) - 
unix:/home/ianni/.netkit/hubs/vhub_ianni_A.cnct
daemon_open : control connect failed, errno = 2
-----

I did some tracing and found that the corresponding uml_switch process is 
started but dies as soon as the VM kernel is started, for no apparent reason.

The problem could have something to deal with start timings but I'm not sure: for

vstart fw -v --eth0=A --eth1=B --eth2=C --eth3=D --xterm=gnome

I get that, randomly, eth3 and eth2 have their hub survive and get connected 
correctly.

Is somebody else able to reproduce the same situation?
(I'm using gnome-terminal 2.28.1 on a Ubuntu 9.10 as host machine, fresh netkit 
2.7 installation)


regards, gb