Here LEGATO gives adjacent node listener receive timeouts too.
Probably not surprising - all HECnet Ethernet traffic from LEGATO goes
thru Johnny's bridge and psilo, so nobody on HECnet would see any better
results than MIM.
I don't exactly know how to go about figuring out where the bottleneck is.
"ping Psilo.Update.UU.SE" and traceroute give trip times of about 190ms,
which doesn't seem outrageous. The hello timer on circuit QNA-1 on LEGATO
(which is what I assume to be the relevant parameter here) is set to 15
seconds. I don't know what the corresponding timeout is on the MIM side,
but maybe Johnny will tell us. In any case, changing the QNA-1 hello timer
to 5 seconds just for fun didn't seem to make much difference, so the
problem is likely that packets are being lost somewhere.
Bob