[OpenAFS] mountpoints are always RW...

[Tino Schwarze]

On Tue, Feb 11, 2003 at 04:43:29PM +0000, Hamish Marson wrote:

> >Are all the volumes in the path from 'cellname' down to 'test' 
> >replicated as well?  Once the client is forced down a RW path (because 
> >one or more of the intervening volumes is RW-only) it will continue 
> >down the RW path.
> 
> I thought this was it... root.afs hadn't been replicated... Except now 
> even with root.afs replicated it is still accessing the RW copy...  And 
> I've been though all levels & checked, and ALL volumes are replicated... 
> Is it as simple as the caching clients need restarting once the replicas 
> have been made (Reboot still is it?) Or a timeout needs waiting on?

There might be a timeout involved...

> What puzzles me is that I can't do an lsmount on /afs itself... It says 
> it doesn't exist... e.g.

Because /afs is not in AFS. BTW: Make sure that all replicas are
released before experimenting further.

> Is it possible to explicitly mount the .readonly copies o fthe volumes 
> under /afs/cellname/mumble and the read/write copies under 
> /afs/.cellname/mumble and force the behaviour I want (i.e. no update on 
> the clients till I release), or is that too gross a hack to consider?

You can, of course, use an "exclusive ro" path with the .readonly
volumes mounted on and another path with the RWs for updating, like:

/afs/cell/mumble/mnt1 -> somevol.readonly
/afs/cell/foobar/mnt1 -> somevol

HTH! Tino.

-- 
             * LINUX - Where do you want to be tomorrow? *
                  http://www.tu-chemnitz.de/linux/tag/