2003e9402f26f%(f04)c
.10f02ec
+ ext Brian White f0c
:

> 
> > >Let's say I have a system with 100 servers, each serving up files from a
> > >large not-heavily-used archive to a local cluster of machines.  I have
> > >these servers in place already for other purposes; using CODA to make
> > >the archive available is just one more benefit.
> > 
> > CODA does not serve data from other file systems.  Data in a CODA
> > server must be stored in CODA via coda.  NFS can export an existing
> > file system.  CODA can not do this.
> 
> I know what you're saying and I understand.  I was thinking of a "volume"
> as a filesystem that is "served" by the various servers to all of the
> clients.  The volume I was referring to was for this example the "large
> not-heavily-used archive".

But I think, if I understand what you want to do, you can make each of
those "servers" be Coda *clients*, and have them serve out to their
local clients via NFS or HTTP.  Would that meet your needs?

> 
> 
> > >I'd like to be able to say that the files in this archive must exist on
> > >at least 3 (for example) servers at any one time, but can be held on more
> > >if it is convenient.  Such a setting would allow me to make use of the disk
> > >space available across all servers (much like a raid) without each server
> > >having to replicate everything.  Frequently used files would obviously
> > >get replicated across many servers (for speed) while infrequently used
> > >files would just reside in the minimum number of places.
> > 
> > Server groups are not dynamic.  When you create a volume, it is assigned
> > to a server group and that group of servers stores that volume and
> > all files in the volume.  The number of servers storing that volume
> > does not change.
> 
> Understood.  I guess the next question would be if there are any thoughts
> about providing this feature?  It seems to me that this would provide the
> benefits of a replicating file system along with the benefits of a raid.
> The total volume size would scale automatically as more servers were added
> to it.

There have been, of course, a number of research projects on similar
things (see Darrell Long's Swift, TicketTAIP, and CHEOPS, to name just
three).  AFAIK, none of that has been integrated with Coda.  Would make
an interesting project.

		--Rod