Hi Ben
Thanks for the update. Future meetings can be seen here if it helps:
http://indico.cern.ch/categoryDisplay.py?categId=338. The next is on
18th October. Unfortunately there are always a few clashes.
Regarding Steve's tests the answer is "yes". More fully this is what he
mentioned a few days ago to me:
"The atlas test page won't display for some reason. I am in the process
of setting a new atlas test system while simultaneously moving to
SL4/ATLAS release 13 etc. ... The old tests are still running and the
history will be there (eventually)."
ATLAS production jobs have stopped at the moment while they upgrade
their distribution software.
Cheers,
Jeremy
> -----Original Message-----
> From: Testbed Support for GridPP member institutes [mailto:TB-
> [log in to unmask]] On Behalf Of Ben Waugh
> Sent: 27 September 2007 00:05
> To: [log in to unmask]
> Subject: Re: UKI monthly meeting - Thursday 27th September at 10:30 in
EVO
>
> Dear Jeremy,
>
> Sorry, my memory evidently isn't that good. I'll be teaching,
Gianfranco
> is away, and I doubt anyone from UCL-CENTRAL can attend at such short
> notice, especially as they are quite busy installing Legion, the new
UCL
> computing facility. So here is a brief status report.
>
> UCL-HEP has huge numbers of queued ATLAS jobs, which I guess might tie
> in with the fact that the CE (pc90.hep.ucl.ac.uk) is on the list under
> AOB of sites publishing wrong VOView information. Solving this
properly
> will require some more expert input, so I have closed the ATLAS queue
> for now. Am I right in thinking that closing one queue (for a
particular
> VO) does not require a broadcast message?
>
> UCL-CENTRAL seems to be healthy. The site was getting hardly any jobs
> until we changed the way of counting waiting jobs. By including local
> jobs in the count of waiting jobs we seemed to be preventing any LCG
> jobs from arriving. There were often thousands of queued local jobs,
but
> it seems reasonable not to count these since LCG has a dedicated
> fair-share and thus does not wait behind all the other jobs.
>
> ATLAS production jobs were still not arriving because of some missing
> tags (VO-atlas-cloud-UK and VO-atlas-tier-T2) which have now been
> re-added by Alessandro de Salvo. Now of course ATLAS have wider
problems
> so there are still no production jobs arriving.
>
> And finally just a comment that Steve Lloyd's ATLAS tests were very
> useful in the days when they worked rather than timing out or running
> out of memory. Can they be resurrected in any way?
>
> Cheers,
> Ben
>
> Coles, J (Jeremy) wrote:
> > Dear All
> >
> > Those with a good memory will recall that the next UKI deployment &
> > operations meeting was scheduled for tomorrow, Thursday 27th
September
> > at 10:30 in EVO. This is still the case and the agenda is being
> > constructed here:
> > http://indico.cern.ch/conferenceDisplay.py?confId=19902.
> >
> > This will be a first attempt to use EVO for the UKI meeting. At the
last
> > meeting most of those present indicated that they were already using
EVO
> > or wanted to switch. Please test your connection early to ensure the
> > meeting is not delayed. For those who have yet to use EVO you will
need
> > to setup your account via the following link:
http://evo.caltech.edu/.
> >
> > Please let me know if there is anything specific that you would like
> > covered at the meeting - the agenda is (intentionally) not full
> >
> > Many thanks,
> > Jeremy
>
> --
> Dr Ben Waugh Tel. +44 (0)20 7679
7223
> Dept of Physics and Astronomy Internal: 37223
> University College London
> London WC1E 6BT
|