Hi ...
>> IMHO, I think that the gLite Cluster staff should present several use
>> cases and how to implement those different scenarios using gLite
>> Cluster. This would make the transition easier. I understand that
>> gLite-Cluster is really useful, but from my first attempt trying to
>> check on the service, it seems that I have to go through an adiabatic
>> transition from my current topology to the new one.
> What kind of thing do you find difficult? In terms of deciding how many clusters you need I think it's quite hard to give general guidance because sites are all different, and the needs of their users may also be different, e.g. support for jobs with large memory requirements.
Maybe it is just my problem of not understanding enough the Glue Schema
concepts (I guess I'm not the only one) but I think that examples of use
cases should be given.
Some time ago the following document was provided to WLCG site admin in
order to publish correct information using several CEs connected to
several subclusters:
http://www.italiangrid.org/sites/default/files/GridMap_CPU_counting.pdf
A good/general starting point would be how to implement the scenario
explained there using gLite-Cluster.
Cheers
Goncalo
|