[OpenAFS] Re: RW replication

[Esther Filderman]

You can re-make RW volumes from RO volumes, but it is not as simple as
"pushing a button."

AFS's "redundancy point", as you call it, is about RO replications and
the high availability of RO [that  is, static] data across your cell. 
 The basic idea is that static data can be copied to various servers;
in the event of the failure of one, the client will move along to
another copy located on another server.

As I said before, there is no RW replication integral in AFS.  You may
be able to mimic such a thing with hardware and scripts but I'm not
sure it's either easy to do or totally able to be automated.

On Sun, 6 Mar 2005 20:14:14 +0100, Jordi Haarman
<j.haarman@synantics.com> wrote:
> Ok..... I overlooked that. ;-)
> 
> If my RW server burns out. Is it possible to 'upgrade' the RO to RW with a
> click of a button? (otherwise I'm missing the redundancy point here :-) ).
> 
> Jordi
> 
> On Sat, 5 Mar 2005 20:53:32 +0100, Jordi Haarman
> <j.haarman@synantics.com> wrote:
> > 1. What will happen (if it is possible) if  I have a replicated read/write
> > volume on a second server and 2 clients edit the same file on different
> > servers while networkconnectivity is lost between the 2 sections.
> 
> There is no RW replication in AFS.  There is RO replication; you can
> have multiple RO copies of a RW volume.
> 
>