Hi,
I don't know whether this is a related problem.
We are seeing load on our CE go quite high (and still increasing).
We have a number of biomed jobs running and every minute a new
'globus-job-manager' process is started by the biomed uid on the CE.
Currently there are around 500 'globus-job-manager' processes on the
CE with 19 running biomed jobs.
Thanks
krishan
On Fri, 10 Aug 2007, Coles, J (Jeremy) wrote:
> Dear All
>
>
>
> It would seem we have an important issue to raise at the EGEE ops
> meeting on Monday. Are there any other sites which have been having
> similar problems (with biomed or any other VO)?
>
>
>
> Many thanks,
>
> Jeremy
>
>
>
> -----Original Message-----
> From: Testbed Support for GridPP member institutes
> [mailto:[log in to unmask]] On Behalf Of Santanu Das
> Sent: 10 August 2007 18:37
> To: [log in to unmask]
> Subject: Re: Biomed jobs killing our CE
>
>
>
> And it's not letting our WNs to breath - left ~57GB in the biomeds' pool
> directory.
>
> cheers,
> Santanu
>
> Kostas Georgiou wrote:
>
> On Mon, Aug 06, 2007 at 09:59:25PM +0100, Peter Gronbech wrote:
>
>
>
> This is exactly what happened at oxford they sent about 1000 jobs. I
> have set a que limit of 100 now for biomed.
>
>
>
> What worries me is that the jobs involved are just pilot jobs that
> download jobs from an external site, as far as I can tell at the moment
> there is no authentication worth mentioning involved in the download of
> the jobs which is clearly not acceptable.
>
> Cheers,
> Kostas
>
>
>
>
>
|