On Fri, Feb 16, 2001 at 10:41:02AM +0100, Christian Haul wrote:
> Hi.
>
> After reading the manuals I'm not sure if I could tackle my problem
> with coda. Perhaps you could give me some comments on my planned
> scenario.
>
> There are two networks, only occasionally connected through a modem
> line.
...
> My first idea was to have a SCM on the laptop and replication servers
> on W and H. Since I'll be the only one editing files I don't expect
> many server-server conflicts.
That server layout is a recipe for disaster. How are H, L and W going to
resolve when they get out of sync? Don't forget the client cache, and
the fact that a Coda client can work while disconnected from the server.
One server at work, and a hoard profile that pulls yor home directory
into the client caches on the laptop and at home is probably better.
> Will I be able to export a NFS share with coda, i.e. have
> server:/home/foo as /vicepa?
No, we have a lot of additional metadata such as version vectors and
ACL's which are kept track of in RVM by the Coda servers. /vicepa is
just a store for the bulk data (filecontents).
> Will replication & synchronization work as expected (i.e. seamlessly)?
>
> Are there other, better suited tools for this you know of?
This is definitely becoming a FAQ, possibly Intermezzo or unison and
there is always rsync.
Jan
