Dear Julio,
This problem with ATLAS - an habitual offender
with regard to cleaning up after jobs - had been
flagged earlier in the week and it was some time
before anybody from ATLAS responded to our queries.
In point of fact, everything still present on
our SE was copied over to a new disk last night
(Sunday evening) and will be available again
when our site is back on-line.
We have now allocated a quota of 150Gb to ATLAS
locally so that, prior to our 3Tb disks arriving,
space can be allocated on a more equitable basis.
Best wishes,
Barry.
On Mon, 17 Jan 2005, Julio Lozano Bahilo wrote:
> Date: Mon, 17 Jan 2005 10:29:49 +0100
> From: Julio Lozano Bahilo <[log in to unmask]>
> Reply-To: LHC Computer Grid - Rollout <[log in to unmask]>
> To: [log in to unmask]
> Subject: Re: [LCG-ROLLOUT] Final Notice ... Re: [LCG-ROLLOUT] IMPORTANT:
> comunications site -> experiment
>
> Hello David,
> first of all I have to say that I coudn't read your last message.
> We had been working hard to understand the implications of dele-
> ting the files on your site. The ATLAS production system hasn't
> been as reliable as we wanted and we've had problems concerning
> the copy and registration of output files.
> We managed, anyway, to clean by about 2/3 the disk space on your
> machine at some point and decided that we had to look carefully
> into the rest of the files to see which should be replicated and
> how it should be done to keep the consistency with the catalogue
> system and our production database. We thought that it was more
> than sufficient since we didn't foresee that it was so urgent for
> you to clean TOTALLY your disk space.
> Sending an ultimatum at the last minute on a friday evening, when
> you still had more than 200 GB free out of your 340 GB total, is
> not a very kind move with respect to the ATLAS collaboration, I
> have to say. If some sites are not capable of providing a good
> quality of service ( for whatever reasons ) they should be exclu-
> ded from the list of LCG production sites.
> Cheers,
> Julio.
>
> Dr D J Colling wrote:
>
> > Hi Julio (or any other appropriate atlas user)
> >
> > We Still have:
> >
> > 119G /stage/lcg2-data/atlas/datafiles
> > 591M /stage/lcg2-data/atlas/generated
> > 110M /stage/lcg2-data/atlas/logfiles
> >
> > on gw38.hep.ph.ic.ac.uk. I suspect that this is from jobs that were still
> > finishing and already running when we shut the queue. There is currently
> > one job still running (all the others have finished ... although I did
> > kill about 5 that had clearly stalled).
> >
> > Please could you remove this data if you want to keep it. Any not removed
> > by 12-noon tomorrow is likely to be removed the hard way...
> >
> > All the best,
> > david
> >
> > On Thu, 13 Jan 2005, Dr D J Colling wrote:
> >
> >> Hi Julio,
> >>
> >> Thanks for this.
> >>
> >> The storage element is gw38.hep.ph.ic.ac.uk and the atlas area is
> >> /stage/lcg2-data/atlas
> >>
> >> so if you do something like:
> >>
> >> edg-gridftp-ls gsiftp://gw38.hep.ph.ic.ac.uk/stage/lcg2-data/atlas
> >>
> >> you see:
> >>
> >> generated
> >> datafiles
> >> logfiles
> >>
> >>
> >> which to me looks like some sort of standard production set up and so
> >> you
> >> are probably better placed to know which files need to be copied where
> >> than I am. What I can tell you is that if I log on to the SE directly
> >> and
> >> do a du -hs on the directory I get
> >>
> >> 267G /stage/lcg2-data/atlas/datafiles
> >> 13G /stage/lcg2-data/atlas/generated
> >> 98M /stage/lcg2-data/atlas/logfiles
> >>
> >>
> >> The rest of the disk is taken up by the experiment software directories.
> >> When you have cleared the data areas we will split these on to seperate
> >> disks and so increase the space available (slightly) but also making
> >> sure
> >> that data produced by running jobs cannot block code installation etc.
> >>
> >> All the best,
> >> david
> >>
> >>
> >>
> >> On Thu, 13 Jan 2005, Julio Lozano Bahilo wrote:
> >>
> >>> Hello David,
> >>> please, tell me which are the files to be transfered .. I'm going
> >>> to try to copy those files to a different storage place.
> >>> Cheers,
> >>> Julio Lozano.
> >>>
> >>> Dr D J Colling wrote:
> >>>
> >>>> However still no reply from an Atlas person. I really don't want to do
> >>>> this but if there is no movement of this data tomorrow morning we are
> >>>> going to have to delete these files with all the problems that this
> >>>> will
> >>>> cause.
> >>>>
> >>>> All the best,
> >>>> david
> >>>>
> >>>> On Thu, 13 Jan 2005, Burke, S (Stephen) wrote:
> >>>>
> >>>>> LHC Computer Grid - Rollout
> >>>>>
> >>>>>> [mailto:[log in to unmask]] On Behalf Of Dr D J Colling
> >>>>>> said:
> >>>>>> mail to [log in to unmask] bounced (probably because I am
> >>>>>> member of CMS)> proving it not to be a very reliable way of
> >>>>>> communicating
> >>>>>> with the Atlas community.
> >>>>>
> >>>>>
> >>>>>
> >>>>> Actually I forwarded it to atlas-lcg anyway, but I would rather not
> >>>>> have the
> >>>>> system relying on that :)
> >>>>>
> >>>>> Stephen
> >>>>>
> >>>
> >>
>
---------------------------------------------------------
Dr Barry MacEvoy
High Energy Physics Group
Imperial College London
Blackett Laboratory
Prince Consort Road Tel: 020 7594 7802
LONDON SW7 2BZ Mobile: 07767 323871
England Fax: 020 7823 8830
---------------------------------------------------------
|