Coda File System

Re: Found no partitions for server node1.

From: Jan Harkes <jaharkes_at_cs.cmu.edu>
Date: Thu, 20 Jul 2000 11:49:18 -0400
On Mon, Jul 17, 2000 at 01:22:08PM +0200, System V wrote:
> Hello 
> 
> Sorry to bother U folks, but I can't get the hang of volume creation.
> I wan't to create a replicated volume.
> So I made this /vice/db/VSGDB
> [root_at_scm /root]# cat /vice/db/VSGDB 
> E0000100 scm
> E0000101 node1
> E0000102 scm node1
> 
> I added node1 to /vice/db/servers
> [root_at_scm /root]# cat /vice/db/servers 
> scm		1
> node1	2
> 
> According to the info I've got, Now I should be able to make a
> replicated volume by doing:
> [root_at_scm /root]# createvol_rep coda.rep E0000102 /vicepa
> Servers are (scm node1 )
> Found no partitions for server node1.
> 
> What is the problem here. I don't get it.

If node1 was recently set up, remove /vice/vol/BigVolumeList before
running createvol_rep. We check that file to see if the partition is
available on the server, but forget to regenerate it when a new Coda
server is added. A fix for this is already in the CVS tree.

> I have some more questions:
> 1. Is it posable to create a replicated root volume

Yes.

> 2. If it is better to make a r/o replicated root volume, how do I do
> that

No, r/o replication is deprecated and you can make a r/w replicated
volume `virtually read-only' by simply setting all ACL's to `rl'.

> 3. Can I put more than one volume in one directory (eg coda.root and
> coda.user on /vicepa)

Yes, volumes are logical collections of files, multiple volumes can
share the same physical partition on the server.

Jan
Received on 2000-07-20 11:50:06