On 9 Oct 2007, at 12:01, Santanu Das wrote:
> Graeme Stewart wrote:
>> For biomed you won't expect to see problems - they don't use prd
>> accounts. But for ATLAS the YAIM setup of lcg-info-dynamic-
>> scheduler.conf is just wrong.
> Okay, this is the latest result, for ATLAS and the VOView appears
> to be fine again:
Could be that jobs are running as ordinary atlas pool accounts?
The duplicate lines are weird. What happens if you run the plugin by
hand, i.e.,
/opt/lcg/libexec/lcg-info-dynamic-scheduler -c /opt/lcg/etc/lcg-info-
dynamic-scheduler.conf
Do you get double values output?
If you run qstat (or condor equivalent of), how many atlas jobs are
counted? Which pool accounts are they running as? What primary group
name do those accounts have?
> On 9 Oct 2007, at 12:09, Santanu Das wrote:
>>> You mean "lcg-info-dynamic-scheduler.conf"? I actually, already
>>> did in my original replay. Here it is again:
>>>
Yeah, sorry - missed that.
The important mapping is one from any atlas production accounts'
primary groups (LHS) to atlas (RHS).
Cheers
Graeme
--
Dr Graeme Stewart - http://wiki.gridpp.ac.uk/wiki/User:Graeme_stewart
ScotGrid - http://www.scotgrid.ac.uk/ http://scotgrid.blogspot.com/
|