[glue-wg] VOTE by 8/1: GLUE 2.1 for describing cloud resources?

[stephen.burke at stfc.ac.uk]

Maria Alandes Pradillo [mailto:Maria.Alandes.Pradillo at cern.ch] said:
> The only problem I see if we go for a 2.1 version is that we need to
> understand how to deploy all this in the hierarchical BDII model.

We've done this before: basically we deploy the new schema first, before anything publishes. You have to make sure that at least all the top BDIIs are upgraded before going further (although the consequence on a non-upgraded instance is only that the information is missing). You would also need site BDIIs to upgrade, but only at sites which are going to publish the new objects, and so far that's a fairly small number of sites and they could be told to do that at the same time as upgrading their cloud services to the new publisher.

> I´m not sure what happens when you want to publish something in LDAP that is not declared in the LDAP schema file.

If you publish something not in the schema it just gets rejected. In particular, if the schema has a mandatory attribute which is *not* published the whole object gets rejected as invalid, and then everything in the tree below it, which is why we should not add new mandatory attributes to existing objects.

> I guess this is the reason why we had two different LDAP trees for GLUE 1.3
> and GLUE 2.0, right?

Not really, that was done to keep the two sets of information apart and make sure that existing queries never saw GLUE 2 objects.

Stephen

-- 
Scanned by iCritical.