On 30/01/2012 17:22, Camaleón wrote: > On Mon, 23 Jan 2012 11:29:07 +0000, Shaun wrote: >> >> output from cat /etc/network/interfaces > > (...) > > I have the same setup as yours in my lenny servers but with a difference: > both cards are connected to the same physical switch in the same VLAN. > Mmm... have you tried the other way round? Disconnect eth1 and see if it > works.
I've tired that by setting the active interface from the command line with ifenslave -c. Didn't make a difference the other way around. > Another thing I would test is with no bonding setup at all, configure > both ethernet cards separately and try to ping with both of them, i.e.: Yes I tried this just to eliminate cable/NIC/switch. They work fine when configured independently with different IPs. Also when I do arp -a (via remote console access via iDRAC like technology) I see the arp table is empty following failover. It has entries prior to failover. I have to login via the remote-console access and restart networking to get it up again. I can only think it's something to do with the switch but not sure how to prove this. I can also get my hosting company to connect both NICS to the same switch and I can try testing again? But I would lose the switch redundancy this way :( Is there anything further I can do to debug or eliminate things? Sadly I don't have any access to any of the following or proceeding hops to run tcpdump or such stuff. Shaun -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/4f291c24.4010...@gmsl.co.uk
