> I accidentally came across this problem when I was trying to access my 
> realm at 10.0.0.1. Instead of doing 'ls 10.0.0.1', I did 'ls 10.0.01'.
> 10.0.01 is interpreted as 10.0.0.1 just fine, so after retrying with
> the correct IP, I ended up with two directories in my coda root for 
> the same realm.
> 
> At the next restart, venus segfaults when in vsgent::KillMgrpMember,
> m->KillMember() is invoked with an invalid value for m. I think the
> problem is that the volume replica transient members are reset twice, 
> because it segfaults the second time. 
> 
> Is this a known problem?

Not as far as I know. I have accessed realms through different names
many times, the only problem that I know of is that the servers only
send back callbacks to one of the realm instances, so the other mounts
end up with stale data.

I'll try to reproduce your problem.

Jan