[Gluster-users] Using GlusterFS in a distributed office

[Francesco Pedron]

Hi everybody,
I want to start saying that I'm totally new to GlusterFS. I've played 
with it just in the last week.

Here in Italy, in a small university's project, we are trying to make a 
"distributed office". It's a part of a specific linux distribution that 
we are developing.
We would like to have file distributed and replicated on a number of the 
office's computers without using a central file server.
So we are thinking about making a number of office's computers at the 
same time a server and a client of the GlusterFS. In this way data 
should be replicated on this computers and if a computer fails we can 
change it. The others would be just clients.

I try to esplain some other of our ideas in detail:

- we would make servers only with a fraction of office's computers
- office's computers will never be turned off; they only go to sleep and 
could be waked up with Wake up On Lan
- we would manage sleep of the office's computer in a distributed way 
with a costum app, and servers would go to sleep all togheter only when 
no other "client" is logged in to his desktop
- when the whole office is in sleep and a client is waked up it will 
wake up all the servers with a Wake up On Lan
- if a server crashes our distributed app will take care of that and we 
would elige a new server.

Me and another couple of guys are the ones that should take care of that 
"distributed office" and, before we start to seriously play with 
GlusterFS and to do some scripting, we would like to know if, in your 
opinion, what we would like to do is possible with GlusterFS.

Thanks,
Francesco