I''ve just moved a drive which contains a Dom0 and multiple DomU''s to a new motherboard. The Dom0 is FC8, kernel 2.6.21-2952, with Xen 3.0-x86_64. The new motherboard is completely different, and I had to make one change to get the domU''s (various Windows and Linux) to boot: in the vif line in the config file, I changed the bridge from xenbr0 to virbr0 (no idea why, but it works). However, I have got one problem. With the old setup, the DomU''s got their IP addresses via a DHCP request to a hardware router, and the router gave each DomU a fixed IP address, according to its MAC address. In the new setup, the router isn''t seeing a DHCP request from the DomU''s. Instead, the Dom0 is serving up addresses (in its own subnet; 192.168.122, instead of 192.168.1). Xen is presumably handling this direct, since there''s no DHCP server installed on the Dom0. The DomU''s can still see the network, but the rest of the network can''t see them. Any ideas on how I can fix this? Thanks - Evan _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
> > In the new setup, the router isn''t seeing a DHCP request from the > DomU''s. Instead, the Dom0 is serving up addresses (in its own subnet; > 192.168.122, instead of 192.168.1). Xen is presumably handling this > direct, since there''s no DHCP server installed on the Dom0. > > The DomU''s can still see the network, but the rest of the networkcan''t> see them. > > Any ideas on how I can fix this? >Are you running the 192.168.122 and 192.168.1 networks across the same ''wire''? If so you probably need to isolate them using vlans. James _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
I think what I''m seeing is what I''d expect on switching from xenbr0 to virbr0. The libvirt/dnsmasq defaults seems to be to create a local 192.168.122 subnet, with no access in from outside, and (external) DHCP disabled. I don''t want to mess with libvirt, since I don''t use the XML config stuff and I don''t want to move away from my known good setup prior to my motherboard change, so I''m concentrating on finding out why xenbr0 no longer works. If I boot up a DomU with xenbr0 in the vif line I get: Error: Device 0 (vif) could not be connected. Could not find bridge device xenbr0. What I think has happened is that the existing fc8-xen on my hard drive is confused about the network hardware on the new motherboard (a Dell R200 server). When the new board originally booted up, networking was disabled, but I fixed this by aliasing eth0 to eth3. The (old) driver seemed to have identified the new NICs at eth3 and eth4, rather than eth0/1. Anyway, that fixed the network on Dom0, but it seems to be giving Xen a headache when it boots any DomU''s. The DomU''s boot up Ok with virbr0; the problem is just xenbr0. I know that FC8 had bugs in the network-bridge script related to aliasing, but network-bridge runs without producing errors, so I''m not convinced that that''s the problem here. :( -Evan _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Evan Lavelle wrote:>What I think has happened is that the existing fc8-xen on my hard >drive is confused about the network hardware on the new motherboard >(a Dell R200 server). When the new board originally booted up, >networking was disabled, but I fixed this by aliasing eth0 to eth3. >The (old) driver seemed to have identified the new NICs at eth3 and >eth4, rather than eth0/1.How about changing the udev rules to reset your interfaces to eth0 and eth1 ? Typically /etc/udev/rules.d/z25_persistent-net.rules contains lines like : SUBSYSTEM=="net", DRIVERS=="?*", ATTRS{address}=="aa:bb:cc:dd:ee:ff", NAME="eth0" When you change hardware, udev will see the new MAC address and create rules for the next available interfaces. Just change eth3 to eth0 (or whatever the correct arrangement is) and after reboot you''ll have the same interface names you originally had. -- Simon Hobson Visit http://www.magpiesnestpublishing.co.uk/ for books by acclaimed author Gladys Hobson. Novels - poetry - short stories - ideal as Christmas stocking fillers. Some available as e-books. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
On Mon, 2009-07-13 at 15:21 +0100, Evan Lavelle wrote:> Error: Device 0 (vif) could not be connected. Could not find bridge > device xenbr0.Perhaps the xen network script fails to set up the bridge properly. Most likely eth0 mac address has changed, so the init stuff automatically renames it to something different. I think Simon Hobson has the right answer for you (he replied a few minutes ago), although I usually have a different approach. Editing /etc/sysconfig/network-scripts/ifcfg-eth0 and changing HWADDR to the new board mac address will also do the trick. A "service network restart" should then properly set up eth0, but you may still need to restart xen stuff in order to get xenbr0 up. Cheers, Radu _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Simon Hobson wrote:> How about changing the udev rules to reset your interfaces to eth0 and > eth1 ?Fantastic! That''s fixed everything. Radu: the HWADDR was already correct; I think there''s a specific aliasing problem going on somewhere. Thanks - Evan _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users