On Wed, Sep 12, 2007 at 06:07:44PM +0200, Aroua Chamseddine wrote:
> In the slave (ns3m2) I saw the following message in /vice/srv/SrvLog:
>
> *17:28:57 GrabFsObj, GetVolObj error Volume not online
> 17:28:57 RS_LockAndFetch: Couldnt translate Vid for 7f000007.1.1
> 17:28:57 RS_UnlockVol: Couldnt XlateVid 7f000007
> 17:28:57 RS_LockAndFetch: Couldnt translate Vid for 7f000007.1.1
> 17:28:57 RS_UnlockVol: Couldnt XlateVid 7f000007
> 17:28:58 GetVolObj: VGetVolume(7f000007) error 603
> 17:28:58 GrabFsObj, GetVolObj error Volume not online*
>
> (7f000007 est le Vid du volume rep)
>
> What surprised me is that  before this, I mounted 3 volumes and they are 
> working fine.

I think the VRDB file (and possibly the VLDB) files in /vice/db on the
servers have not been propagated correctly. Check if the updatesrv
daemon is runnin on your SCM machine and restart the updateclnt daemon
on the second server.

btw. only the files in /vice/db are master-slave replicated. The Coda
server itself doesn't work that way. All replicas are considered equals
and the servers don't actually don't even talk to each other except
after they have been told by a client that they hold different or
conflicting versions of the same object.

> _*Another problem:
>
> *_ I created a new volume "my_vol" with createvol_rep vol and then I 
> removed it with purgevol_rep vol.  When I try create a new volume with the 
> same name "my_vol".I obtained in the scm(ns3m1) this message:
> *my_vol already exists as a replica on ns3m1*

Is that 'purgevol_rep vol' a typo? I would expect 'purgevol_rep my_vol'?
You can use 'volutil -h <server> getvolumelist' to get a detailed list
of all the volume replicas stored on a specific server. But it can be
that the fact that the updateclnt/updatesrv propagation of the files in
/vice/db is problematic is causing such errors.

Jan