(Illustration by Gaich Muramatsu)
Hi, I have an installation where I see a potential use for Coda. We have a few servers and an increasing number of Linux clients running some home-brewed GUI software on various releases of Red Hat. This was fine for a while, but as we move from eight to nearer fifty clients, it turns into a bit of a maintainability disaster, as you may well imagine, so I'm looking at options which might let us centralise some of this. My favoured solution would be to have a small local (or nfs-) root partition on each client, a local /tmp and /var and the rest (basically /home and /usr) shared over the network. Where possible, I'd like to avoid nfs (wouldn't we all? :) so Coda seemed like the obvious choice here. Caching is rather important here, as we don't want to swamp the network for basically read-only data (/home isn't too vital - not much goes there) but we'd like a more comprehensive and reliable solution than a few shell scripts and rsync. I'm confident that we can hack the distribution without serious pain, but I'd like to hear someone on this list tell me that: 1. This sounds sensible 2. Coda on Linux is reliable enough to deal with this sort of thing in a production environment 3. There's a neater way than a symlink to do /usr over Coda Any offers? :) Cheers, Matthew.Received on 1999-09-22 12:17:04