[GRASS-user] ignoring .gislock and consequences?

Hamish hamish_b at yahoo.com
Mon May 10 06:09:17 EDT 2010


Jonathan Greenberg wrote:
> I recall some time ago there was a way to "hack" grass (is this a more
> formal boot-time option now?) and have it ignore .gislock to run
> concurrent processes in the same mapset.  I'm curious if this is still
> possible, and, if so, if it would be safe to do this with multiple
> r.mapcalc runs which use different input files and different output
> names?  Thanks!

all in all it is much simpler+cleaner to just background the job by putting
an "&" at the end of the unix shell line.

for simple use on a multi-core CPU see the "poor-man's multi-processing
trick":
  http://grass.osgeo.org/wiki/R.sun#Automation


if there is a chance processes may interact, use multiple mapsets with
@othermapset and g.mapset, g.mapsets, and temporarily creating new mapsets
+ GRASS_BATCH_FILE:
 cd $LOCATION/
 grass64 -c temp_mapset_$$/

see Markus N's cluster & batch job pages/scripts on the wiki for more
details.



Hamish



      


More information about the grass-user mailing list