> -----Original Message-----
> From: Testbed Support for GridPP member institutes [mailto:TB-
>
> On Thu, Aug 6, 2009 at 01:53, Ewan
> MacMahon<[log in to unmask]> wrote:
> >> -----Original Message-----
> > - Thirdly, I suspect you're scaring off other jobs.
>
> This one is not our problem.
>
I entirely agree in principle, users, even really big ones, shouldn't
have to think about this sort of thing, but from a site point of view,
it's something we have to deal with one way or another. Preferably in
a way that doesn't affect the VOs at all, but it's still worth
mentioning. It'd be interesting to know if anyone else has seen
anything similar.
> I thought that the WMS used the VO view which defended against this
> brokering error?
>
I don't know. I'm also not absolutely sure that my diagnosis is correct;
it's definitely safe to say that we had a large backlog of pilots, the
general expected response time went very high, and we didn't get matched
any WMS work, and that when we didn't have a large backlog any more the
expected response time came down and WMS work started arriving again. It
is possible that it was just a coincidence.
> Or was it ATLAS user jobs via WMS which you got?
>
Not mostly; it was back to our usual mix of odds and sods - quite a bit
of pheno with smatterings of individual users from various other VOs.
Ewan
|