On Fri, Jul 13, 2001 at 02:55:08PM +0200, Wim Godden wrote:
> I've got 2 servers up and running, called servA and servB. servA is the
> SCM.
> 
> On both servA and servB, /vice/db/VSGDB contains :
> E0000100 servA servB
> 
> but when I run :
> createvol_rep test E0000100 /vicepa
> I get :
> /usr/local/sbin/createvol_rep: [: too many arguments
^^^
I wonder which test triggered this error.

However, the actual volume creation looks like it went fine.

> So it all seems to work fine. When I start venus now from either of the
> 2 servers, /coda is mounted correctly, but when I do an ls /coda, I get
> connection timeout. I'm not in disconnected state (as mentioned in the
> Manual under Client Problems), so what could be wrong ? All processes
> that should be running are running. But as before, /vice/srv/SrvLog
> remains empty...

The fact that /vice/srv/SrvLog remains empty puzzles me to no end. It is
always written to, often there is actually too much data. Is there some
different path for the server logs specified in /etc/coda/server.conf?

The volumes must exist and at least the attributes for the root object
have been found, so the RVM state is fine.

The ETIMEOUT can only be a result of a failed directory data fetch,
maybe /vicepa isn't accessible to the servers. Strange, they normally
run as root processes. This could also be the reason that SrvLog isn't
written too.

Jan