Yes, we kill on order a few hundred jobs a month, usually from single
non-HEP users. Only get a fraction of these (few).
2009/1/30 Coles, J (Jeremy) <[log in to unmask]>:
> John/Peter
>
> Does your eyeballing of the pages lead you to often take action - that
> is to actually kill the inefficient jobs? If so could you estimate
> whether you catch few/many/all?
>
> Thanks,
> Jeremy
>
> -----Original Message-----
> From: Testbed Support for GridPP member institutes
> [mailto:[log in to unmask]] On Behalf Of John Bland
> Sent: 30 January 2009 10:28
> To: [log in to unmask]
> Subject: Re: Does your site actively remove inefficient/stalled jobs?
>
> Liverpool use the same monami/eyeball combo.
>
> John
>
> Peter Love wrote:
>> @ lancs we use Paul's monami stuff to monitor this with eyeballs. eg.
>>
> http://fal-pygrid-17.lancs.ac.uk:8123/ganglia/?r=day&sg=no&c=LCG-Service
> Nodes&h=fal-pygrid-18.lancs.ac.uk
>>
>> 2009/1/30 Stephen Childs <[log in to unmask]>:
>>> Brew, CAJ (Chris) wrote:
>>>> Hi Jeremy,
>>>>
>>>> At RALPP we don't actively look for them. If one of us notices that
> the
>>>> far is full but not doing much work we might investigate then go on
> to
>>>> kill jobs but unless there are a lot of them we probably wouldn't
>>>> notice.
>>> Same at TCD. Periodically I run a local command qeffic which displays
> a list
>>> of jobs ordered by efficiency. Actually it's just an alias for this:
>>>
>>> showq -r|sort -n -k 4|sed -e 's/^[ \t]*//' -e '/^$/d'
>>>
>>> Stephen
>>> --
>>> Dr. Stephen Childs,
>>> Research Fellow, EGEE Project, phone:
> +353-1-8961797
>>> Computer Architecture Group, email: Stephen.Childs @
> cs.tcd.ie
>>> Trinity College Dublin, Ireland web:
> http://www.cs.tcd.ie/Stephen.Childs
>>>
>
>
> --
> Dr John Bland, Systems Administrator
> Room 220, Oliver Lodge
> Particle Physics Group, University of Liverpool
> Mail: [log in to unmask]
> Tel : 0151 794 2911
> "I canna change the laws of physics, Captain!"
> --
> Scanned by iCritical.
>
|