On Apr 7, 2008, at 3:01 PM, Adam Bazinet wrote:
Thanks, missed that somehow.
You know, I always knew the RFT protocol allowed deletion, but I think
that's the first time I've caught on that there was a commandline
interface too.
I'm a little worried about the schema. I found this page:
http://viewcvs.globus.org/viewcvs.cgi/ws-mds/usefulrp/schema/schema/mds/usefulrp/batchproviders.xsd?view=markup&content-type=text%2Fvnd.viewcvs-markup&revision=1.6
Is that the latest version of the schema? It seems to match up with
the output I'm used to seeing from various scheduler providers. In
any case, there isn't really a field for disk space, unless I just
missed it. So, is it possible to extend this schema somehow? We
already customize all of our scheduler providers anyway. I suppose
I could go back to your original idea of publishing something
separate to the Index service, but it would be nice if this
information were integrated the way I had in mind.
Part of the catch is going to be that xsd:any element. You can jam
anything in there you want. I think there might be a tiny bit of
trouble, because I think GRAM focuses on the Compute Element side of
the GLUE schema. There is already a schema for the Storage Elements,
though, which you can find some description of at http://www-numi.fnal.gov/offline_software/srt_public_context/GridTools/docs/glue_schema.html
I don't think it would be too heinous to include a Storage Element
definition in the xsd:any section of the Compute Element definition
published by the scheduler provider. But thinking things like that is
probably why they don't have me writing schema in the first place.
I'll ask someone who does more with our info services whether we
already publish SE information anywhere, and how one is meant to
correlate the CE and SE to each other.
Charles
