On Tue, 3 Jun 2003, Ian Stokes-Rees wrote:
> What causes this sort of behaviour?
I don't know for sure, but one possibility is that the broker queries
the GRISs at every matching site before submitting, and they can get
into a state where they take a very long time to respond - although
several days is rather worse than I've seen!
> 1) Find out how many jobs the RB is trying to schedule; and,
As an ordinary user, I don't think you can.
> 2) Find out what the state of the queues are at the various sites the RB
> serves.
You can get some information from the info system, for example:
ldapsearch -x -h gppinfo06.gridpp.rl.ac.uk -p 2135 -b \
"mds-vo-name=local,o=grid" "objectclass=computingelement" | egrep -i \
"(ceid:)|(idlejobs|runningjobs)"
CEId: epcf36.ph.bham.ac.uk:2119/jobmanager-pbs-S
RunningJobs: 0
IdleJobs: 0
MaxRunningJobs: 99999
CEId: epcf36.ph.bham.ac.uk:2119/jobmanager-pbs-M
RunningJobs: 0
IdleJobs: 0
MaxRunningJobs: 1
CEId: epcf36.ph.bham.ac.uk:2119/jobmanager-pbs-L
RunningJobs: 0
IdleJobs: 0
MaxRunningJobs: 6
CEId: ce0-gla.scotgrid.ac.uk:2119/jobmanager-pbs-gridqs
RunningJobs: 0
IdleJobs: 0
MaxRunningJobs: 96
...
Stephen
|