multi-g.manuals; concurrency locking

[Ken Sibley]

> 
> BTW, does anybody know for sure what happens if you simply disable the
> concurrency locking mechanism in GIS.sh? I actually wanted to set up a lab
> sometime soon with a large server and multiple PC's running X11 emulation
> software over TCP/IP to the server, and run GRASS from each PC. I guess
> that might be hard to do with the concurrency locks in place.
> 
Not much happens actually.  The problems arise when two sessions are
running in the same mapset.  If you change your region in this mapset
while a process is running in another session the results are 
unpredictable.  This happens mostly when you run long processes; we have
some r.surf.contour jobs that run for 4-10 days at a time.

What you are planning to do with PC's running X emulation is not a problem.
GRASS only checks to see if the same user is running more than once.
Give each user a seperate login and a seperate mapset and as long
as the server can handle the load the individual users will be happy.
We have concurrent users running on machines daily here.  Let me know
if you have more questions about this and yes UNIX is great.

Ken Sibley
USDA-SCS
ksibley at ncg.scs.ag.gov