Begin forwarded message:
> From: Alastair Dewhurst <[log in to unmask]>
> Date: 13 February 2014 12:04:10 GMT
> To: "atlas-uk-comp-operations (Discussion list for Atlas-UK computing operations)" <[log in to unmask]>, "atlas-support-cloud-uk (ATLAS support contact for UK cloud)" <[log in to unmask]>
> Subject: Validating T2D and ABCD status
>
> Hi Site Admins
>
> When it comes to deciding how ATLAS distribute data to Tier 2s, ATLAS uses a metric which sorts sites depending on their availability and network performance. These categories, ABCD then define how much new data a site will get and are updated once a month. Up until now this has been a mostly manual process but development work has been ongoing to make this automated. At Tuesday's ADC Weekly meeting a presentation was given about the new automated system:
> https://indico.cern.ch/event/299400/contribution/9/material/slides/0.pdf
>
> The new metrics can be found at:
> For T2D status:
> http://dashb-atlas-ssb.cern.ch/dashboard/request.py/siteviewhistorywithstatistics?columnid=10097#time=720&start_date=&end_date=&use_downtimes=false&merge_colors=false&sites=all&clouds=UK
> if you are wondering why your site has a specific status more information on the T2D test results can be found here:
> http://gnegri.web.cern.ch/gnegri/T2D/t2dStats.html
>
> For ABCD status:
> http://dashb-atlas-ssb.cern.ch/dashboard/request.py/siteviewhistorywithstatistics?columnid=10102#time=720&start_date=&end_date=&use_downtimes=false&merge_colors=false&sites=all&clouds=UK
> The ABCD status is based on the Availability of the analysis queue as measured here:
> http://dashb-atlas-ssb.cern.ch/dashboard/request.py/siteviewhistorywithstatistics?columnid=564#time=720&start_date=&end_date=&use_downtimes=false&merge_colors=false&sites=all&clouds=UK
> You may remember last year Cloud support would message sites if there availability fell below 90% to ask for a reason. If you want to know if your site's analysis queue is put into test mode because of failing jobs, we can setup an automatic email notification.
>
> A few things to note:
> - The metrics have only been working for a couple of weeks.
> - There are some display fixes that need to be done (such as removing T1 and T3s from the T2D metric).
> - The new metrics should take into account Scheduled Downtimes (i.e. not hold that against you). If you have had a SD please check that this is working.
>
> These links are probably useful for site admins to Bookmark for future reference. If you have any comments/queries please email UK cloud support and we will forward them on to the developers. We can also speak to the developers in person during computing and software week at CERN 24th - 28th Feb. The hope is to use the automated system starting in March.
>
> Thanks
>
> Alastair
__________________________________________________
Dr Elena Korolkova
Email: [log in to unmask]
Tel.: +44 (0)114 2223553
Fax: +44 (0)114 2223555
Department of Physics and Astronomy
University of Sheffield
Sheffield, S3 7RH, United Kingdom
|