Ho hum, hi folks...
This is excruciating. I seem to be taking one step forward and two
back at the moment trying to get a network setup.
It seems I''ve had to abandon bridging - the IPs my hosting provider
give me must be bound to the Dom0 eth0 MAC address or so I believe
anyway.
They''ve given me a bundle more so I now have 206.251.255.112 to
206.251.255.117 so I can use the network-route mode for my five DomUs
and Dom0 (206.251.255.113). The gateway for the Dom0 machine is
206.251.255.1 and I can get in and out from Dom0 no problem.
So, I setup the routing mode in xend-config.sxp
In the DomU configuration file for the first DomU I have:
vif = [ ''ip = 206.251.255.115'' ]
ip = "206.251.255.115"
NETMASK="255.255.255.0"
GATEWAY="206.251.255.113"
and in the interfaces on the DomU I have:
auto eth0
iface eth0 inet static
address 206.251.255.115
gateway 206.251.255.113 (Dom0''s IP address)
netmask 255.255.555.0
dns-nameservers 209.216.206.167 206.251.225.22
dns-search jcsolutions.co.uk
DomU starts fine and from it I can ping 206.251.255.113 (Dom0''s eth0/
vif1.0) but no further. Similarly from outside, from my other machine
on 206.251.255.110, I can ping 113 but not 115 (DomU).
DomU''s route -n command gives the following:
Destination Gateway Genmask Flags Metric Ref
Use Iface
206.251.255.0 0.0.0.0 255.255.255.0 U 0 0
0 eth0
0.0.0.0 206.251.255.113 0.0.0.0 UG 0 0
0 eth0
and on Dom0 route -n gives the following:
Destination Gateway Genmask Flags Metric Ref
Use Iface
206.251.255.115 0.0.0.0 255.255.255.255 UH 0 0
0 vif1.0
206.251.255.0 0.0.0.0 255.255.255.0 U 0 0
0 eth0
0.0.0.0 206.251.255.1 0.0.0.0 UG 0 0
0 eth0
(One thing of note possibly here is that if I do it without the -n
flag the 206.251.255.115 in the first route is resolved as
"vhost1.jcsolutions.co.uk" and not
"zimbra.jcsolutions.co.uk" which is
the name in DNS for 206.251.255.115 and I cannot work out where it
might be picking that up from - I''ve checked the local hosts file and
so on and if I ping "zimbra.jcsolutions.co.uk" from Dom0 it does
correctly resolve to 206.251.255.115)
I''ve used Wireshark on Dom0 to watch both vif1.0 for incoming pings
from DomU to the outside world and eth0 on Dom0 to watch for incoming
pings for DomU from the outside world. I see ARP requests asking Dom0
"who has 206.251.255" when trying to ping it from outside but no reply
and no ICMP packets (though I can also ping Dom0 fine from outside).
And I see arp requests asking Dom0 on vif1.0 "Who has
206.251.255.110?" abd then ping requests that get no reply and
don''t
show up on the eth0 trace on Dom0.
(See - you can tell I''ve read the "Using Xen" book chapter on
all this!)
I''m guessing I''m missing a route somewhere. One further thing
of note
perhaps is that in the network-route script I have not put the line:
echo 1 > /proc/sys/net/ipv4/eth0/proxy_arp
in as instructed in the book and elsewhere on this mailing list
because it already has the lines:
echo 1 > /proc/sys/net/ipv4/ip_forward
echo 1 > /proc/sys/net/ipv4/${netdev}/proxy_arp
...in my version already.
So, can anyone shed any light on this for me? I''m absolutely stuck at
the moment and it has taken all week so far with three nights pretty
well working right through because I need to get software setup and
running on these VMs before I go back to work on Monday, so I''m really
really desperate. But I''m pretty sure I must just be missing
something (like a comma or something!) somewhere.
Thanks in advance,
Jock
--
Jock Coats
Warden''s Flat 1e, J Block Morrell Hall, OXFORD, OX3 0FF
local rate: 084 JOCKOXFD (56256933) skype:jock.coats?call
jock.coats@jcsolutions.co.uk http://jockcoats.org.uk
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users