On Mon, 29 Feb 2016 06:30:14 PM Russell Coker via luv-main wrote:
xenbr0: received packet on mb0 with own address as source address
What's the best way of tracking down the cause of kernel messages like the above?
As an aside it's a pity that the kernel doesn't log the source address in question and whether it's IPv4 or IPv6.
It turned out that there was a loop in the network caused by dumb switches (not running STP) being connected together. Unplugging mb0 fixed it. So next time you see this the first thing to try is to just unplug the interface in question and see what happens, if everything still works then you just fixed a network loop. The network in question had periods of about a minute of packet loss between 2 Linux systems that were both running STP. After unplugging mb0 that problem went away. I'm now wondering if it would have been better if I hadn't enabled STP, then whoever put in the bad cable would have immediately brought the network down and it would have been obvious that there was a problem. Instead we had less obvious problems for a few days. -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/