Hi
perhaps it is time to have all sites apply the "max two concurrent
processes" on SGM accounts.
On the other hand, we have seen instances where the user got SGM because
the sites LCAS / LCMAPS / VOMS subsystem was not correctly configured.
I suggest a polite mail to the relevant person asking why they are
running so much as SGM.
This same person was running SGM here, because the VOMS subsystem failed
to give him a normal ATLAS account, because we had run out of normal
ATLAS pool accounts. So LCMAPS fell back from VOMS to legacy gridmap
support, and in the legacy gridmap this person was listed as SGM. it
was our fault, not his, that he acquired an SGM account.
JT
Kyriakos Ginis wrote:
> On Thu, Oct 26, 2006 at 10:23:30AM +0200, Jeff Templon wrote:
>> Hi,
>>
>> My take: we need to figure out why torque doesn't catch the CPU time,
>> and we need to account wall time, otherwise I think these jobs are fine.
>>
>
> Hello,
>
> Regarding the 'condor' issue these jobs might be fine, but I wonder why
> the SGM accounts are still used for things that don't look like software
> installation and configuration :-)
>
>
>>> atlassgm 21442 20845 92 Oct25 ? 22:09:41 ./2Qgen
>
>
>
>
|