On Thu, Mar 16, 2000 at 11:27:14AM -0000, fsg.student4_at_bt.com wrote:
> Hi,
> 
> I've just moved my Coda servers and clients to an independent
> subnet (IP 10.x.x.x) and reinstalled coda. The root volume
> with only one server works fine. I experience however 
> problems as soon as I create a second server - even before 
> creating a replicated volume. 
> 
> When I include a second server in the files in /vice/db
> (like described in the HowTo and like it worked before)
> the client doesn't start any more. After "Venus starting"
> I get a "[1]+ Killed    venus".

That is because you have not unmounted /coda, venus forks off a child
during startup which tries to mount, but it fails and the child sends a
SIGKILL to venus.

> Both servers show these messages in SrvLog:
> 16:05:43 GetVolObj: VGetVolume(7f000000) error 103
> 16:05:43 GrabFsObj, GetVolObj error Volume not online

Volume not online can indicate several things, are there errors in the
server logs during startup when the volume are salvaged and attached? Is
the timestamp on the VRDB believable (7fxxx should be found on any
server in the VRDB) and consequently is the VRList on the SCM good?
Maybe there is a line in the / vice/db/servers file that parses wrong.
If there is an empty line at the end remove it.
> 
> My subnet has no DNS server. Does Coda need a DNS server,
> for example for reverse lookup, or is the file /etc/hosts
> on all clients and servers enough?

That should work. It's just pulling the ipaddresses from under the names
that has messed with the server's mind a bit.

Jan