Hi Alastair
To add to Stephen's comment about machine features taken from the VM world, you may also find the GDB summary on implementation useful;
"Implementation availability
• Torque 2.5/PBS 2.3: maintained by NIKHEF, probably not implementing the last specification
• To be checked by Jeff
• PBS Professional and Grid Engine: KIT, up-to-date
• LSF: Ulrich, up-to-date, in production since last Monday
• Several hacks to work-around the fact that LSF cannot be queried directly: may require some tweaking at other sites"
Simone made a comment that there are things in this that could be used very quickly by ATLAS. Perhaps RAL could get involved in further site testing/initial rollout.
Jeremy
On 17 Oct 2012, at 17:23, Stephen Burke wrote:
> Testbed Support for GridPP member institutes [mailto:TB-
>> [log in to unmask]] On Behalf Of Alastair Dewhurst said:
>> I could get the values from the information system but I have no idea if they
>> are correct.
>
> They are read dynamically from the batch system configuration so they should be correct.
>
>> Also I know that at least one site, normalises wall time
>> depending on the speed of the machine. In this case a single value in
>> SchedConfig is not sufficient. Instead we can set an environment variable
>> per node:
>> $ATLAS_QUEUEHOURS
>> The pilot will pick up this environment variable and use this instead.
>
> LCG has a general proposal for this kind of information, it would be better for atlas to use that rather than inventing its own! E.g. see the talk at the last GDB:
>
> http://indico.cern.ch/getFile.py/access?contribId=14&sessionId=0&resId=0&materialId=slides&confId=155073
>
> Stephen
>
> --
> Scanned by iCritical.
|