I'm running Open vSwitch Version 1.0.1 on a system using Xen 4.0 hypervisor with the 2.6.31.13 pvops kernel. Generally things work well but I am seeing something mysterious and I'm hoping someone on the list can enlighten me a bit.
When I start up OVS I create a bridge named br0 with a single physical port/interface 'eth3'. So far so good. When I start a guest I use the following command in my HVM create script to assign a emulated interface to the guest: vif = [ 'type=ioemu, mac=00:16:3e:f8:f8:01, bridge=br0' ] OVS outputs the following: Jun 29 13:42:17|00232|reconnect|INFO|unix:/tmp/stream-unix.18791.0: connecting... Jun 29 13:42:17|00233|reconnect|INFO|unix:/tmp/stream-unix.18791.0: connected Jun 29 13:42:17|00815|bridge|INFO|created port tap8.0 on bridge br0 Jun 29 13:42:17|00234|jsonrpc|INFO|unix:/tmp/stream-unix.18791.0: connection closed Jun 29 13:42:17|00235|reconnect|INFO|unix:/tmp/stream-unix.18791.0: connection dropped Jun 29 13:42:17|00236|reconnect|INFO|unix:/tmp/stream-unix.18878.0: connecting... Jun 29 13:42:17|00237|reconnect|INFO|unix:/tmp/stream-unix.18878.0: connected Jun 29 13:42:17|00816|bridge|INFO|created port vif8.0 on bridge br0 Jun 29 13:42:17|00238|jsonrpc|INFO|unix:/tmp/stream-unix.18878.0: connection closed Jun 29 13:42:17|00239|reconnect|INFO|unix:/tmp/stream-unix.18878.0: connection dropped All is well from what I can see. It's tap8.0 and vif8.0 because the domain ID is 8 in this case. The RHEL 5.4 guest starts up normally and proceeds to boot. After a few moments though I see the following output: Jun 29 13:42:55|00817|netdev_linux|WARN|ethtool command ETHTOOL_GSET on network device vif8.0 failed: No such device Jun 29 13:42:55|00818|timeval|WARN|26 ms poll interval (0 ms user, 7 ms system) is over 41 times the weighted mean interval 1 ms (131581 samples) Jun 29 13:42:55|00819|timeval|WARN|context switches: 1 voluntary, 0 involuntary Jun 29 13:42:55|00820|coverage|INFO|Skipping details of duplicate event coverage for hash=586cea98 in epoch 131581 Jun 29 13:42:55|00821|bridge|WARN|bridge br0: interface vif8.0 deleted unexpectedly Jun 29 13:42:55|00822|bridge|WARN|bridge br0: port vif8.0 has no interfaces, dropping After that port / interface vif8.0 disappears but here's the strange thing; even though the vif interface is gone the virtual interface still appears in the RHEL 5.4 HVM guest, I can assign an IP to it and it works fine. In Dom0 I still see the tap8.0 port / interface. Any insight as to what's going on here? Since everything appears to work fine anyway I'm inclined not to worry about it too much but thought I'd throw it out on the list to see if I can figure it out. I'm afraid that it's a sign of some larger problem that may come back to bite me later. - Greg Rose LAN Access Division Intel Corp.
_______________________________________________ discuss mailing list discuss@openvswitch.org http://openvswitch.org/mailman/listinfo/discuss_openvswitch.org
