Hi
I will not comment on accounting issue until we have full figures. Most of the CREAMCE's are still failing to run Nagios jobs in 330 min. Jobs are staying in the queue and then killed by Nagios after 330 mins. Only exception seems to be Oxford and Imperial.
I would suggest sites to have a look in logs and if there are many ops jobs in queue then please kill it. It will give chance to new ops jobs to run. Please let me know if you see anything unusual.
Thanks
Kashif
> -----Original Message-----
> From: Testbed Support for GridPP member institutes [mailto:TB-
> [log in to unmask]] On Behalf Of Jeremy Coles
> Sent: 14 July 2014 09:49
> To: [log in to unmask]
> Subject: Re: "Upstream" problem for UKI?
>
> Hi Winnie,
>
> We can review what is happening at the ops meeting tomorrow and log the
> issue for reference. With regards to WLCG availability & reliability, I circulate
> figures monthly upon which site admins can comment and request a re-
> computation directly with WLCG if there is a significant issue. If by accounting
> you mean the GridPP assessment for funding shares, then we can check for
> the next round at the time data is used, but I think the time periods are so
> long that this current problem (if it is translating into an accounting issue - I've
> yet to confirm that) is unlikely to have much impact.
>
> Jeremy
>
>
>
>
> On 14 Jul 2014, at 08:12, Winnie Lacesso <[log in to unmask]>
> wrote:
>
> > Bonjour!
> >
> > The problem is continuing right to now.
> > Well my hope/expectation is that Grid Admin won't penalize sites in
> > ANY WAY for this CREAM-CE outage (NOT!) since the problem is upstream.
> >
> > Can someone guarantee that please?
> >
> > Yours in Grid reliability/availability/accounting concern....
> >
> > Winnie Lacesso / Bristol University Particle Physics Computing Systems
> > HH Wills Physics Laboratory, Tyndall Avenue, Bristol, BS8 1TL, UK
|