Hello Enrico,
LHC Computer Grid - Rollout wrote:
> Unfortunately for INFN-PADOVA we have queues with different
> numbers of WN running on the same cluster. We are committed
> to provide different service levels to different VOs, for
> example we have a contract with a VO to dedicate always 4
> CPUs (2 WN) and in case of need we can change this value.
> So the numbers reported for our site are wrong, the right
> value should be the one of "biggest" queue, I hope this will not
> complicated too much the (already) complex situation...
We have a similar situation, but this is not reflected in the information system. We are running the torque/maui combination (from the outside similar to openpbs) and have limits on the number of CPUs some VOs may use.
How have you changed the information system to get different numbers in there? I have seen that edg-ce-all uses pbsnodes -a to obtain the number of CPUs, which is of course the same for all queues.
Regards,
Fokke
--------
Fokke Dijkstra
High Performance Computing
SARA - Reken- en Netwerkdiensten http://www.sara.nl
Tel. +31 20 592 8004 Fax. +31 20 668 3167
|