[GRASSLIST:1017] Parallel use of GRASS

[Jan Hartmann]

I have been working this way on a Beowulf cluster with a shared 
filesystem. The results were excellent: submitting many GRASS
commands in parallel from different mapsets in the same database sped up 
  working time  considerably. IMO this is one of the great advantages of 
OS source systems like GRASS and Beowulf compared to commercial systems.

I wonder if the locking mechanism could be permanently adapted this way. 
As far as I can see, GRASS only writes to the current mapset, so only 
there a lock has to be provided. If lockfiles were put and read in the 
current mapset, concurrent use of GRASS would be possible in most cases. 
   Would this be somethig for a feature request?

Jan


Glynn Clements wrote:
> Luis Goncalves Seco wrote:
> 
> 
>>I whould like to know, if it is possible to have more then 
>>one session of GRASS running at the same time?
> 
> 
> You would need to remove the locking code from $GISBASE/etc/Init.sh:
> 
> 	# Check for concurrent use
> 	"$ETC/lock" "$lockfile" $$
> 	case $? in
> 		...
> 
> If you remove this, you would need to ensure that:
> 
> a) you don't run multiple sessions using the same mapset, and
> 
> b) that you don't use a monitor name (x0, x1, PNG etc) multiple times;
> monitor names are per-user, not per-session or per-mapset.
>