[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: slightly-OT: centralized user management

On Fri, Jul 29, 2005 at 08:01:28PM -0400, Roberto C. Sanchez wrote:
> Greetings,
> I currently have a small home network (1 server, 1 workstation, 1
> laptop) with only two users.  What I would like to do is to setup some
> sort of centralized user authentication mechanism (NIS, LDAP, whatever)
> with home directories mounted from the server.  This is primarily since
> I will be adding more machines and more users in the near future.
> I would like recommendations/experiences from the list on what is a good
> approach and maybe some resources.  The server already runs NFS, and I
> have experience with a combined NIS/NFS setup for a computer lab I
> formerly admined.
> One thing that is an absolute necessity is a sort of "roaming profile"
> support similar to that which is available with certain Redmond-based
> legacy operating systems.  Specifically, logging into machine A will
> cause a "copy" of my $HOME to be cached on the machine.  Machine A is a
> laptop and I unplug it from the network.  I would like to be able to
> login to the machine, make changes to my files/whatever and then have
> them automagically sync up with the server next time machine A
> reconnects to the network at home.  This may obviate the need for NFS.
> Additionally, it would be necessary for the login credentials to be
> cached so that disconnected login would actually work.  I am pretty sure
> that this is possible, but I am not really sure what the best approach
> is.  Ideas and recommendations are welcome.

Two things come to mind, neither of which I've used (or used enough)
to judge their viability.

(1) rsync
  which I'm told is a way of keeping two file systems in sync with one another

and if you'd like a revision control system, 

(2) monotone
  which I've tried just enough to have trouble with it and still be
enthusiastic about its potential, even though it is still in active
development (current version is 0.21) and there may be compatibility
problems between versions.  Its big plus is that it is totally
distributed.  It can be used to maintain a database of variants of,
say, source files even though it has *no* central server.
It is specifically designed for the case that parts of a team of
developers may work in isolation, say, on detached laptops, using
their own copies of the repositor(y/ies), which are synchronised
when it happens to be convenient.

-- hendrik

Reply to: