LHC Computer Grid - Rollout
> [mailto:[log in to unmask]] On Behalf Of Jeff Templon said:
> I still don't understand this part of your original message though:
>
> > The new pbs dynamic plug-in for the generic information
> provider fixes
> > the assumption that the pbs server is on the same node as the CE.
I'm not sure if this will make things better or worse, but there are two
separate issues:
1) The pbs info provider had a lot of bugs/features in it, many reported by
me last week (look at the most recent bugs in savannah assigned to lfield).
One of those was that it assumed that the pbs server, i.e. the thing where
you do "qstat xxx@server" to get the info, was the same machine as the
gatekeeper, which is true at most sites but not all. Many of those bugs are
now supposed to be fixed, although I haven't yet looked at the new code
myself.
2) There is an independent issue with the YAIM tools which set up the
configuration for the info provider. It seems that they were written on the
assumption that there is one queue per VO, and it was at least not obvious
to people with a different setup how to adapt things to get the right
configuration for their queues. Laurence has said that this is possible and
the documentation on how to do it will be improved.
> > If you think the assumption that the queue name is the same
> as the VO
> > name is incorrect, then please could you submit a bug.
>
> Why would we need to submit a bug if the code isn't making the
> assumption?
The queue per VO is at least the *default* for YAIM setups even if it can be
changed, but it isn't obvious (to me at least) that it should be the
default. The bug is against YAIM (and assigned to okeeble), not against the
info provider.
Stephen
|