[GRASSLIST:2568] Re: grass5 and network connections

[giorgio19291 at inwind.it]

Yes please tell me whats changes I have to do.

 I have to solve This problem.
My machine "A" (Pentium III and 256 MB Ram and  little Hard-Disk) is more
efficient than machines "B" (Pentium MMX with 64 MB and  Big Hard Disk to
store data) so I hope to solve more fast my problems of visualization and
calculations if  Grass5 is on "A" and data to read are on "B"? I hope this
is good solution. Can you give me your opinion?
Thanks A lot
Giorgio.



----- Original Message -----
From: "Eric G. Miller" <egm2 at jps.net>
To: <grasslist at baylor.edu>
Sent: Sunday, October 07, 2001 7:59 PM
Subject: [GRASSLIST:2561] Re: grass5 and network connections


> On Sun,  7 Oct 2001 18:52:06 +0200,
"giorgio19291 at inwind.it"<giorgio19291 at inwind.it> wrote:
>
> > Hi All,
> >  I have a little private network of two linux machines and want to know
> > if is possible use grass5 on a computer  with data and location stored
> > on the other computer of my network? (I have tried to mount filesystem
> > where is my location with smbfs or nfs but I can not open the monitors
> > or set region in this way).
> > Any Ideas
>
> I think it would work if you install grass5 on the other machine and
> used ssh while forwarding X to the machine your working from.  Not the
> nicest solution, but GRASS uses lockfiles and unix sockets which don't
> really work over a network.
>
> I have a patch that changes where the unix sockets are placed
> (~/.grass5/com/, rather than <mapset>/.tmp/).  It might help with the
> unix sockets thing (since the sockets would stay local even if the data
> is remote), but this wouldn't solve the lockfile issue (though, that
> may not be an issue if you have write permissions to the NFS share).
>
> It might be sufficient to do the job, but I don't have an NFS setup to
> test it.  If you're interested, I can tell you what to modify in the
> source build tree to enable that (the code is already there, just not
> built by default).
>
>
> --
> Eric G. Miller <egm2 at jps.net>