Hi everyone,
I think GridICE is fantastic. I hope Ganglia/Nagios/RRDtool/R-GMA can
all be combined into a single all singing all dancing Grid Monitoring
package. I have just been pointed to the RAL GridICE site:
http://grid-ice.esc.rl.ac.uk/gridice/site/site.php
I don't quite understand what I am looking at in the VO view:
http://grid-ice.esc.rl.ac.uk/gridice/vo/vo.php?VOname=lhcb#
specifically the VO Jobs Graph. There are specifically three things
which confuse me:
1. What users from LHCb have submitted 500 running jobs, and where have
the 1000 waiting jobs come from?
2. Why is it that all the other VOs seem to have *exactly* the same
number of jobs in running and waiting states?
3. When there are so many free resources, why are so many jobs (2/3rds)
in a waiting state? Since there are supposedly 2000+ CPUs on line now
for LCG, I would have thought LCG would not have *any* waiting jobs,
unless they had very extreme/strange requirements.
Cheers,
Ian.
--
Ian Stokes-Rees [log in to unmask]
Particle Physics, Oxford http://www-pnp.physics.ox.ac.uk/~stokes
|