From EGEE-see WIki
Dear GIMS,
As you can see, the CMS VO is aggressively installing CMS application software at
SEE-GRID/EGEE clusters.
If you see:
GlueHostApplicationSoftwareRunTimeEnvironment:
.........
VO-cms-CMKIN_4_2_0_dar VO-cms-OSCAR_3_6_5_SLC3_dar
VO-cms-ORCA_8_7_1_SLC3_dar
at gstat monitoring, this means the installation has been succesful.
The CMS jobs may require more than 50 hours of CPU time to complete,
even on a 2.8 GHz P4 CPU.
Unfortunately, the default limit during installation of LCG2_4_0
is 48 hours of CPU time for the cms queue.
Therefore it is necessary to increase the CPU and walltime
limits for the cms queue.
This is done on the CE node by:
qmgr
Max open servers: 4
Qmgr: set queue cms resources_max.cput=200:00:00
Qmgr: set queue cms resources_max.walltime=200:00:00
to set these limits to 200 hours (which is a bit optimistic - it means more than one week).
Unfortunately, when you run configure_node on CE, these changes may be lost!
Cheers,
Emanouil Atanassov
emanouil@parallel.bas.bg
