On Tue, Feb 08, 2000 at 08:57:06PM -0800, Quinn Weaver wrote:
> 20:45:35 Probably another Venus is running! open(/dev/cfs0) failed (19),
> exiting

If anything has an open reference to /dev/cfs0 this error is returned
(as it is most likely another Venus process). In the early linux kernel
modules there was some problem with the refcnt not being dropped in all
cases, but that has been fixed since version 2.2.9 or something.

> On a Windows client, I get an "unable to open TCP socket" error box.

Ehh, haven't seen that one, maybe the tcp/ip isn't enabled/configured
for your network interfaces?

> If I issue a createvol_rep command on the SCM, I get something like this:
> 
> root_at_boc:~# createvol_rep coda_root E0000100 /vicepa
> Servers are (boc )
> HexGroupId is 7F000003
> creating volume coda_root.0 on boc
> V_BindToServer: binding to host boc.private.vovida.com
> RPC2_NewBinding to server boc.private.vovida.com failed with
> RPC2_NOBINDING (F)
> Failure to make volume on boc!. Abort..
> 
> Needless to say, since I can't start venus, I can't log on as the Coda
> administrator remotely and run createvol_rep from another box, as I
> should. :P

Createvol_rep is run on the server-side, at least I do. It is the
"cfs mkm" that is done on the client.

Could you check in the server logs (/vice/srv/SrvErr and
/vice/srv/SrvLog) where it keels over.

btw. others have had problems with mapping > 270 MB RVM data segments in
the server. It uses more/different libraries compared to rdsinit, so
rdsinit might have been able to map it, but the server cannot load the
memory afterwards. If the log shows something like a failure while
loading rvm data, try with a starting address of 0x50000000.

Jan