On 23/05/15 09:50, Andrew Stuart wrote:
So for some reason the rump kernel and xen aren’t agreeing on how to get the
network started or at least the DHCP part of it.
Correction: So for some reason the rump kernel and EC2 aren’t agreeing on how
to get the network started.
Well, if the network isn't working, the first thing you should check for
is how the network is configured. .. it is configured, right? ;)
I can't quite figure out why we don't see any warnings from
_rumprun_config(), but at any rate, try hardcoding the following at the
top of rumprun_config():
rump_pub_netconfig_dhcp_ipv4_oneshot("xenif0")
If that doesn't help, might be that the hacked up dhcp client is such a
bomb that we've found a dhcp server it doesn't work against. But at
least you should see some sort of error message.
(Use of xenif0 of course assumes that EC2 provides the normal Xen PV
network, otherwise it definitely will not work that way.)
