 |
 |
On 30 July 2012 14:01, J Nebrensky <[log in to unmask]> wrote: > On Mon, 30 Jul 2012, Sam Skipsey wrote: >> On 30 July 2012 11:22, Stephen Jones <[log in to unmask]> wrote: > ... >>> Anyway, the point of this message is this: while finding >>> a way to keep the "GridPP Approved VOs" synchronized to >>> the CIC Portal, it came out that several pieces of odd VO data >>> were not disseminated via the CIC Portal. >>> >>> The fields like this that I know about are: >>> >>> VO_VONAME_SW_DIR=blah blah blah >>> VO_VONAME_DEFAULT_SE=blah blah blah >> >> This almost always points at the site's SE (given that almost no site >> has two production SEs). > > Hello, > > OTOH I can imagine that small VOs with limited manpower might want a > single default SE Gride-wide - not sure how they'd disseminate that > though. > Well, they could, in principle. In practice, it's always been assumed that DEFAULT_SE will be a site-local SE, since the data model implied it. I think it would confuse sites if a VO tried to specify a single common SE, just because no-one does it. >>> VO_VONAME_STORAGE_DIR=blah blah blah >> >> This is SE type dependant, but is easily derived once you know the >> type of the SE (for dpm sites, for example, it should always be >> /dpm/$sitedomainnameprefix/home/$voname ). > > I'm pretty sure STORAGE_DIR was for the long-gone Classic SE and is now > listed as obsolete - if needed that information should be generated from > the relevant variables by YAIM itself rather than a duplicate entry. Quite right, it is obsolete. I seem to recall YAIM still configuring it for sites for a while after, though, although a check shows that it is, indeed, not present any longer. Sam > > Thanks > > Henry > > -- > Dr. Henry Nebrensky [log in to unmask] > http://people.brunel.ac.uk/~eesrjjn > "The opossum is a very sophisticated animal. > It doesn't even get up until 5 or 6 p.m."